mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-07-27 04:24:45 +00:00
Code Issues Packages Projects Releases Wiki Activity

chore(data_scanned): Simplify the OCR test images. Add GT for easyocr, tesserocr, tesseract

Browse Source 
Signed-off-by: Nikos Livathinos <nli@zurich.ibm.com>
This commit is contained in:
Nikos Livathinos 2024-10-03 16:39:16 +02:00
parent 1d4517ffb4
commit 7ab3b62c18
48 changed files with 289 additions and 1607 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

158
tests/data_scanned/scanned_01.easyocr.doctags.txt
View File

@ -1,142 +1,20 @@
<document> <document>
<paragraph><location><page_1><loc_6><loc_82><loc_19><loc_89></location>Campuslines is provided as a service 1o the TCU community Announcements 01 @vents public meetings and other general campus Information should be brought by the Skitf otfice, Moudy 2915 or sent to TCU Box 32929. The Skitf reserves the rightto edit lor style and taste</paragraph> <subtitle-level-1><location><page_1><loc_21><loc_84><loc_76><loc_87></location>TableFormer: Table Structure Understanding with Transformers</subtitle-level-1>
<paragraph><location><page_1><loc_6><loc_75><loc_19><loc_81></location>DEPARTMENT OF MuSIC MASTER CLASS SERIES AnD JaZZ AND PERCUSSION DEPARTMENTS present Bob Brelthaupt In drumnsel and jazz master class Irom M toanoon today An open rehearsal with the TCU Jazz Ensemble Il beneldIrom noon t0 pmn</paragraph> <subtitle-level-1><location><page_1><loc_8><loc_78><loc_29><loc_80></location>1. Details on the datasets</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_7><loc_73><loc_19><loc_74></location>Tcu INTERNATIONAL</subtitle-level-1> <subtitle-level-1><location><page_1><loc_8><loc_76><loc_25><loc_78></location>1.1. Data preparation</subtitle-level-1>
<paragraph><location><page_1><loc_6><loc_71><loc_19><loc_73></location>ASSOCIATION will meet at 4 Pm loday in tne Bass Living Room Call Darlene Roth 9229448</paragraph> <paragraph><location><page_1><loc_8><loc_51><loc_47><loc_75></location>As a first step of our data preparation process; we have calculated statistics over the datasets across the following dimensions: (1) table size measured in the number of rows and columns, (2) complexity of the table, (3) strictness of the provided HTML structure and (4) completeness (i.e. no omitted bounding boxes) A table is considered to be simple if it does not contain row spans or column spans. Addition ally, a table has a strict HTML structure if every row has the same number of columns after taking into account any row Or column spans. Therefore a strict HTML structure looks always rectangular: However; HTML is a lenient encoding format, i.e. tables with rows of different sizes might still be regarded as correct due to implicit display rules. These implicit rules leave room for ambiguity; which we want lo avoid. As such, we prefer to have strict" tables, i.e. tables where every row has exactly the same length.</paragraph>
<subtitle-level-1><location><page_1><loc_7><loc_67><loc_17><loc_68></location>WEATHER</subtitle-level-1> <paragraph><location><page_1><loc_8><loc_20><loc_47><loc_51></location>We have developed technique that tries to derive missing bounding box out of its neighbors. As a first step; we use the annotation data to generate the most fine'grained that covers the table structure. In case of strict HTML tables. all squares are associated with some table cell and in the presence of table spans a cell extends across mul tiple grid squares. When enough bounding boxes are known for a rectangular table, it is possible to compute the geometrical border lines between the grid rows and columns. Eventually this information is used to generate the missing bounding boxes. Additionally; the existence of unused grid squares indicates that the table rows have unequal number of columns and the overall structure is non-strict. The generation of missing bounding boxes for non-strict HTML ta bles is ambiguous and therefore quite challenging. Thus, we have decided to simply discard those tables. In case of PubTabNet we have computed missing bounding boxes for 489 of the simple and 699 of the complex tables. RegardFinTabNet, 689 of the simple and 98% of the complex tables require the generation of bounding boxes grid grid ing</paragraph>
<subtitle-level-1><location><page_1><loc_25><loc_67><loc_44><loc_68></location>CAMPUS CALENDAR</subtitle-level-1> <paragraph><location><page_1><loc_8><loc_18><loc_47><loc_21></location>Figure 7 illustrates the distribution of the tables across different dimensions per dataset.</paragraph>
<paragraph><location><page_1><loc_6><loc_64><loc_17><loc_66></location>Today's weather will be partly cloudy with highs in the 80s .</paragraph> <subtitle-level-1><location><page_1><loc_8><loc_15><loc_25><loc_17></location>1.2. Synthetic datasets</subtitle-level-1>
<paragraph><location><page_1><loc_6><loc_59><loc_17><loc_64></location>On Thursday; look for highs in the 80s and lows in the 60s with a chance of thunderstorms</paragraph> <paragraph><location><page_1><loc_8><loc_10><loc_47><loc_14></location>Aiming t0 train and evaluate our models in a broader spectrum of table data we have synthesized four types of datasets_ Each one contains tables with different appear -</paragraph>
<subtitle-level-1><location><page_1><loc_5><loc_56><loc_18><loc_57></location>CORRECTION</subtitle-level-1> <paragraph><location><page_1><loc_36><loc_82><loc_62><loc_85></location>Supplementary Material</paragraph>
<paragraph><location><page_1><loc_6><loc_52><loc_18><loc_56></location>Charlsie Mays was incorrectly identified as 0 junior in yesterday $ Skift . Mays is a senior.</paragraph> <paragraph><location><page_1><loc_50><loc_74><loc_89><loc_80></location>ances in regard to their size; structure, and content. synthetic dataset contains 150k examples, summing up to 60Ok synthetic examples. All datasets are divided into Train; Test and Val splits (8O%, 1O% , 109) . style Every</paragraph>
<figure> <paragraph><location><page_1><loc_50><loc_70><loc_89><loc_74></location>The process of generating a synthetic dataset can be decomposed into the following steps:</paragraph>
<location><page_1><loc_5><loc_47><loc_19><loc_50></location> <paragraph><location><page_1><loc_50><loc_60><loc_89><loc_71></location>1 Prepare styling and content templates: The styling templates have been manually designed and organized into groups of scope specific appearances financial data, marketing data; etc.) Additionally; we have prepared curated collections of content templates by extracting the most frequently used terms out of non-synthetic datasets PubTabNet, FinTabNet, etc.). (e.g (e.g</paragraph>
</figure> <paragraph><location><page_1><loc_50><loc_43><loc_89><loc_60></location>2 Generate table structures: The structure of each synthetic dataset assumes a horizontal table header which potentially spans ovCr multiple rows and table body that may contain a combination of row spans and column spans. However, spans are not allowed to cross the header body boundary. The table structure is described by the parameters: Total number of table rows and columns, number of header rows, type of spans (header only spans, row only spans, column only spans, both row and column spans) maximum span size and the ratio of the table area covered by spans</paragraph>
<paragraph><location><page_1><loc_21><loc_64><loc_34><loc_66></location>Last Day t0 withdraw trom classes for Fall</paragraph> <paragraph><location><page_1><loc_50><loc_37><loc_89><loc_43></location>3 Generate content: Based on the dataset theme. a set of suitable content templates is chosen first. Then; this content can be combined with purely random text to produce the synthetic content.</paragraph>
<subtitle-level-1><location><page_1><loc_21><loc_66><loc_24><loc_66></location>Today</subtitle-level-1> <paragraph><location><page_1><loc_50><loc_31><loc_89><loc_37></location>4 Apply styling templates: Depending on the domain of the synthetic dataset; a set of styling templates is first manually selected Then, style is randomly selected to format the appearance of the synthesized table.</paragraph>
<paragraph><location><page_1><loc_21><loc_62><loc_34><loc_64></location>2 pm t0 6 pm , PC Visual ts Display: Elizabeth Leal Student Center Lounge</paragraph> <paragraph><location><page_1><loc_50><loc_23><loc_89><loc_31></location>5 Render the complete tables: The synthetic table is finally rendered by a web browser engine to generate the bounding boxes for each table cell. A batching technique is utilized to optimize the runtime overhead of the rendering process.</paragraph>
<paragraph><location><page_1><loc_21><loc_59><loc_34><loc_60></location>Chapel, Noon University Rober Carr Chapel</paragraph> <subtitle-level-1><location><page_1><loc_50><loc_18><loc_89><loc_22></location>2. Prediction post-processing for PDF documents</subtitle-level-1>
<paragraph><location><page_1><loc_21><loc_60><loc_34><loc_62></location>8 p Thealer TCU Tne Madwoman of Chaillot University Thealre</paragraph> <paragraph><location><page_1><loc_50><loc_9><loc_89><loc_17></location>Although TableFormer can predict the table structure and the bounding boxes for tables recognized inside PDF docu ments, this is not enough when a full reconstruction of the original table is required. This happens mainly due the folrcasons: lowing7</paragraph>
<paragraph><location><page_1><loc_21><loc_57><loc_34><loc_59></location>amnto Pm The Artist andine Baseball Card . Moudy Bldg ernioil space</paragraph>
<subtitle-level-1><location><page_1><loc_21><loc_56><loc_26><loc_56></location>Thursday</subtitle-level-1>
<paragraph><location><page_1><loc_21><loc_51><loc_34><loc_53></location>8 pm. Theatre TCU Ine Madwvoman of Cnaillot , Unis versity Thealre</paragraph>
<paragraph><location><page_1><loc_21><loc_53><loc_34><loc_56></location>Lucy Harris Linn Institute Noon 10 pn PC Visual Arts: Arts Festival, Frog Fountain</paragraph>
<paragraph><location><page_1><loc_22><loc_50><loc_34><loc_51></location>Mamnslo 4p Mn.e The Artisi</paragraph>
<figure>
<location><page_1><loc_5><loc_25><loc_34><loc_47></location>
</figure>
<paragraph><location><page_1><loc_35><loc_65><loc_48><loc_66></location>and the Baseball Card * Moudy Bldg exhibit space</paragraph>
<paragraph><location><page_1><loc_35><loc_63><loc_48><loc_65></location>Noon Honzon Awards luncheon Student Center Ballroom</paragraph>
<paragraph><location><page_1><loc_35><loc_61><loc_48><loc_63></location>5 pm International Students Association meeting Student Center</paragraph>
<subtitle-level-1><location><page_1><loc_34><loc_60><loc_38><loc_60></location>Friday</subtitle-level-1>
<paragraph><location><page_1><loc_35><loc_57><loc_44><loc_58></location>PC TGIF: Karaoke</paragraph>
<paragraph><location><page_1><loc_35><loc_58><loc_48><loc_60></location>8 pm Theatre TCU ~The Madwoman of Chaillot; University Theatre</paragraph>
<paragraph><location><page_1><loc_35><loc_56><loc_48><loc_57></location>University Leadership Retreal</paragraph>
<paragraph><location><page_1><loc_35><loc_54><loc_48><loc_55></location>Student Account Paymenis Due</paragraph>
<paragraph><location><page_1><loc_36><loc_55><loc_46><loc_56></location>PC Film: "Jurassic Park"</paragraph>
<paragraph><location><page_1><loc_35><loc_52><loc_48><loc_54></location>1am1o pmn The Arist andine Baseball Card Moudy Bldg exnibil space</paragraph>
<paragraph><location><page_1><loc_35><loc_50><loc_47><loc_52></location>Noon (o pm , TCU Jazz Ensemble performance</paragraph>
<figure>
<location><page_1><loc_35><loc_25><loc_49><loc_47></location>
</figure>
<figure>
<location><page_1><loc_5><loc_17><loc_48><loc_24></location>
</figure>
<subtitle-level-1><location><page_1><loc_6><loc_15><loc_15><loc_16></location>EMPLOYMENT |</subtitle-level-1>
<paragraph><location><page_1><loc_5><loc_10><loc_15><loc_14></location>Chlldren'* Clothing part-uno posilion ovak able Flenblo hours Store Mon -Sat open 10 00 10 500 Call The Mudpuppy 731-2581</paragraph>
<paragraph><location><page_1><loc_5><loc_7><loc_15><loc_9></location>[Now hiring part-lime holp [lor snow ski ropair and renlals, rcquirod, 2312 Monloomery . Exp</paragraph>
<paragraph><location><page_1><loc_16><loc_12><loc_26><loc_17></location>Fort Worth Ladies Specialty Boubque local cdin the Slockyard Stabon neods experonccd sales person Weckonds non-smnoker , 626-8900</paragraph>
<paragraph><location><page_1><loc_27><loc_12><loc_37><loc_17></location>TERM PaPERS TyPED FASTI Laser prinled. Rush oders and cedit cards accepled, Across trom TCU al 3023 South Univorslty Drive ACCURACY +PLUS 826 4969</paragraph>
<paragraph><location><page_1><loc_38><loc_12><loc_48><loc_17></location>Iromn 3 dillerent lundraise(s lasling elther 30 daysNo investment Earn SSS lor your group plus personal cash bonuses lor yoursell . Cal 1-800-932-0528, Exl65</paragraph>
<paragraph><location><page_1><loc_16><loc_8><loc_26><loc_11></location>Great Money-Grval Funl Become Flash photographer, must have 35mm camnera, will train FLASH 924-2626,</paragraph>
<paragraph><location><page_1><loc_18><loc_6><loc_24><loc_8></location>TYPING</paragraph>
<paragraph><location><page_1><loc_27><loc_9><loc_37><loc_11></location>Letters, lermn papers, theses, Or anything else you mnay nccd typed, 214-641-1021 call;</paragraph>
<subtitle-level-1><location><page_1><loc_27><loc_7><loc_38><loc_8></location>[FUNDRAISINC]</subtitle-level-1>
<paragraph><location><page_1><loc_27><loc_6><loc_37><loc_7></location>FUNDRAISINGChoose</paragraph>
<subtitle-level-1><location><page_1><loc_39><loc_9><loc_47><loc_11></location>FOR SALE</subtitle-level-1>
<paragraph><location><page_1><loc_38><loc_7><loc_48><loc_9></location>Soulhwost and ruslic lurniture cuslomn madeCall lorInloralon 922-9563</paragraph>
<subtitle-level-1><location><page_1><loc_49><loc_89><loc_68><loc_91></location>The Adventuros of</subtitle-level-1>
<paragraph><location><page_1><loc_61><loc_89><loc_68><loc_91></location>Sperfrog</paragraph>
<figure>
<location><page_1><loc_49><loc_80><loc_91><loc_89></location>
</figure>
<subtitle-level-1><location><page_1><loc_49><loc_79><loc_67><loc_80></location>Calvin and Hobbes</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_80><loc_79><loc_93><loc_80></location>by Bill Watterson</subtitle-level-1>
<figure>
<location><page_1><loc_49><loc_70><loc_60><loc_78></location>
</figure>
<figure>
<location><page_1><loc_60><loc_69><loc_71><loc_78></location>
</figure>
<figure>
<location><page_1><loc_71><loc_69><loc_82><loc_78></location>
</figure>
<subtitle-level-1><location><page_1><loc_62><loc_65><loc_80><loc_68></location>TCU DAILY SKIFF Since 1902</subtitle-level-1>
<paragraph><location><page_1><loc_50><loc_61><loc_64><loc_65></location>Ine ICUDauy SkfisVocLcedbynie sludents 01 Teras Chnstan Universty sponsoredoytne journalism department 4a0 published Tuescoy throug) Fnday Cuning Inea Ong Spnnq Semesters encept dunngfnns heesand cunnghol days</paragraph>
<paragraph><location><page_1><loc_50><loc_60><loc_64><loc_61></location>Ine Sh C stbutec Iree 0n Camous silcevs lacully slatanc ns103</paragraph>
<paragraph><location><page_1><loc_50><loc_58><loc_64><loc_60></location>CiRculaTion Ooo Tuesoay through Frday</paragraph>
<paragraph><location><page_1><loc_50><loc_56><loc_64><loc_58></location>SUBSCRIPTIONS Togot suusohblion byma call 921-7000 and asklo enen 801 16274 Subscrphion rales are $20per mester</paragraph>
<paragraph><location><page_1><loc_50><loc_53><loc_64><loc_55></location>PHOTOGRAPHS Sial pholovaphs are avlabeo9 purcuse Dy teaders Quine ShifFor pniong nes conlactIne pholo desk Quce</paragraph>
<paragraph><location><page_1><loc_50><loc_50><loc_64><loc_52></location>EDITORIAL POLICY Uns gned edilonoae oonon Dage feoterenline Meaoethe Sheqtona DoardKncn composeoormne coior manag nqcoior assistant mnanaqing coior, ne4s edior</paragraph>
<paragraph><location><page_1><loc_64><loc_63><loc_78><loc_65></location>Camns edio conoed I0an crector and spons edito Signed leiters an0 columns representte Op nionolihe Vnii0</paragraph>
<paragraph><location><page_1><loc_79><loc_63><loc_92><loc_65></location>SKIFFTELEPHONE DIRECTORY Tne four-dioi ertens ons (6000 senes) Can numtoCis Can b (OOchco Dy dialing 921-7722 Idowed Dy the erens on</paragraph>
<paragraph><location><page_1><loc_64><loc_60><loc_78><loc_62></location>LETTERS To IHE EDITOR Tne Skff elcones lefters ione editorlor publica ton Letters'must b0tyoco doublesoaced Siq7e0 010 Fco Soowords</paragraph>
<paragraph><location><page_1><loc_64><loc_56><loc_78><loc_61></location>soaced Siq7e0 010 Soowords Lerers Shoabe suomnited teast Io cays Deloe publicalion 10 Ine Sk nensoon Moudy 2915 to TCU Bor 22929 0 to Yax 921-7133 must IncudeIne ollhor $ classihcalonmnajor and phone nunocr Ine Sa iesenes ne nqn lo edt € (Ejeci any lenters lo stye andlasie They</paragraph>
<paragraph><location><page_1><loc_64><loc_53><loc_78><loc_56></location>Leters to inc edtor are ulso accaoico vocoma Dorn anc Inrouqh [1 TCU conoer syslern Jolave Vokeener Cial921-7683 Toleave e ma seno to tne Suis TCUvaracoress Iistecbelo:</paragraph>
<paragraph><location><page_1><loc_64><loc_51><loc_78><loc_53></location>Skivy 1 member 01 Ine Assocaled Press The</paragraph>
<paragraph><location><page_1><loc_64><loc_50><loc_78><loc_51></location>MaILING ADDRESS Po 32929 Fodworh Teras 76129 80'</paragraph>
<figure>
<location><page_1><loc_49><loc_34><loc_78><loc_47></location>
</figure>
<paragraph><location><page_1><loc_57><loc_29><loc_71><loc_33></location>fpposios CaNTInA</paragraph>
<paragraph><location><page_1><loc_57><loc_28><loc_71><loc_29></location>Full/Part-time positions</paragraph>
<paragraph><location><page_1><loc_57><loc_27><loc_68><loc_28></location>Training provided</paragraph>
<paragraph><location><page_1><loc_57><loc_26><loc_72><loc_27></location>No experience necessary</paragraph>
<paragraph><location><page_1><loc_54><loc_24><loc_74><loc_26></location>APPLY IN PERsoN Tuesday & Wednesday 3:00 p.m. 5:00 p.m.</paragraph>
<paragraph><location><page_1><loc_53><loc_23><loc_71><loc_24></location>10433 LoMBARDY LN. (Now OPEN)</paragraph>
<paragraph><location><page_1><loc_54><loc_22><loc_75><loc_23></location>723 S. CentrAL ExpresswaY (RicharDsON)</paragraph>
<paragraph><location><page_1><loc_59><loc_21><loc_70><loc_22></location>Equal Opportunity Employer</paragraph>
<subtitle-level-1><location><page_1><loc_53><loc_16><loc_76><loc_20></location>PAPPADEAUX SEAFOOD KITCHEN</subtitle-level-1>
<paragraph><location><page_1><loc_57><loc_14><loc_72><loc_15></location>No experience necessary</paragraph>
<paragraph><location><page_1><loc_58><loc_15><loc_71><loc_17></location>Full/Part-time positions Training provided</paragraph>
<subtitle-level-1><location><page_1><loc_59><loc_13><loc_70><loc_14></location>Apply in person:</subtitle-level-1>
<paragraph><location><page_1><loc_52><loc_12><loc_76><loc_13></location>Tuesday & Wednesday 3.00 p.m: 5.00 pm.</paragraph>
<paragraph><location><page_1><loc_55><loc_8><loc_73><loc_12></location>Dallas: 3520 Oaklawn al Lemmon Irving: 10428 Lombardy Ln, (open soonl) Richardson: 725 $ Cenlral Expressway Arlinglon: 1304 Copeland Rd. at Collins Equal Opporlwuty Ernployer</paragraph>
<paragraph><location><page_1><loc_51><loc_5><loc_76><loc_7></location>Inquiring minds read the Skiff.</paragraph>
<table>
<location><page_1><loc_79><loc_58><loc_92><loc_62></location>
<row_0><col_0><body>Main nmber Ecioe s Uine</col_0><col_1><body>921-7426 6266</col_1></row_0>
<row_1><col_0><body>Camnpus Dusk</col_0><col_1><body>6267</col_1></row_1>
<row_2><col_0><body>News Desk</col_0><col_1><body>6269</col_1></row_2>
<row_3><col_0><body>Sponts Desk</col_0><col_1><body>6266</col_1></row_3>
<row_4><col_0><body>Photo Des:</col_0><col_1><body>6546</col_1></row_4>
<row_5><col_0><body>Bac"shop Far</col_0><col_1><body>6553 821-7133</col_1></row_5>
</table>
<paragraph><location><page_1><loc_79><loc_51><loc_92><loc_58></location>Aner Holrs (Mianiqx '0 am) 921-7683 Sound OA Lineivo ce Mal 921-7683 Advertising 921-7426 Clossifiec 921-7426 Business Manager 6274 Subscnolions 6274 Stucent Plo catons Drecior 6558 Journalism Oflice 921-7425 Imnage Kagazine 821-7429 TCU Var (addruss 921-7161 ProduclionManger 821-7427 Ski)</paragraph>
<paragraph><location><page_1><loc_81><loc_44><loc_90><loc_46></location>What's news? Find out in the</paragraph>
<paragraph><location><page_1><loc_80><loc_43><loc_92><loc_44></location>TCU Daily Skiff</paragraph>
<paragraph><location><page_1><loc_80><loc_40><loc_92><loc_41></location>TRAFFIC TICKETS bulonlyin Fort</paragraph>
<paragraph><location><page_1><loc_79><loc_33><loc_92><loc_37></location>JAMEs R. MALLoRY JOosnesaï Law Ave Fort WorthTX 76107-1793 924-3236</paragraph>
<paragraph><location><page_1><loc_80><loc_37><loc_92><loc_40></location>defended orth Arlington; and clsewhere In arrant County No pornises aso esults Any nine andany count cosis arc hotinclodcd on fce lor legul rcprescntation</paragraph>
<paragraph><location><page_1><loc_80><loc_25><loc_92><loc_29></location>London $459</paragraph>
<paragraph><location><page_1><loc_80><loc_14><loc_93><loc_24></location>Paris 5479' Brussels 8549 [Fraakifurt 5579 [Milan 5659* [Stockholm 5699*| [Los Angeles $250' [Mexico City $198' Wuvobunl aaa A n uln</paragraph>
<figure>
<location><page_1><loc_82><loc_10><loc_90><loc_14></location>
</figure>
<paragraph><location><page_1><loc_80><loc_5><loc_93><loc_10></location>6715 Hillcrest Dallas , IX 75205 214-363-9941 Eurailpasses issued on-the-spotl</paragraph>
<paragraph><location><page_1><loc_82><loc_76><loc_93><loc_78></location>Shed GOIEN USed ioit, Ke Figured She Slould BEME To Oxe</paragraph>
<paragraph><location><page_1><loc_89><loc_73><loc_93><loc_75></location>HEY?</paragraph>
<subtitle-level-1><location><page_1><loc_73><loc_89><loc_92><loc_91></location>by Ba Ranan & Adamn Wright</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_20><loc_89><loc_34><loc_90></location>CAMPUSLINES</subtitle-level-1>
<paragraph><location><page_1><loc_20><loc_84><loc_33><loc_88></location>AlcohOL And DRUG EDUCATIon RECOVERY SUPPORT GROUP meets at 5 pamn Wednesdays 1o Rickel 107 . Call 921-7100.</paragraph>
<paragraph><location><page_1><loc_20><loc_81><loc_33><loc_84></location>PSI CHI psychology honor sociely will havea general meetong al 6.15 pm today Winton-Scott 217 All interested slucenis are Invited to altend</paragraph>
<paragraph><location><page_1><loc_20><loc_77><loc_33><loc_80></location>FRENCH CLUB will meet al 6 30p m today in Student Center 218 1o discuss upcoming evenis</paragraph>
<subtitle-level-1><location><page_1><loc_20><loc_76><loc_33><loc_77></location>FELLOWSHIP OF CHRISTIAN</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_35><loc_76><loc_47><loc_77></location>RAPEISEXUAL ASSAULT</subtitle-level-1>
<paragraph><location><page_1><loc_20><loc_73><loc_33><loc_76></location>ATHLETES welcomes Ken Horton of McKinney Bible Church at 9 pm tonight in the Club Roomn ol thc Coliseum. Everyone is welcome,</paragraph>
<paragraph><location><page_1><loc_34><loc_72><loc_48><loc_76></location>SURVIVORS GROUP meeting from 2 t0 3.30 pm on Fridays , is forming at the TCU Counseling Center Call Dorothy M Barra at 92 -7863 {0 set up an initial screening appointment</paragraph>
<paragraph><location><page_1><loc_20><loc_71><loc_33><loc_72></location>BAPTIST STUDENT MINISTRY (previously Baptist Stu-</paragraph>
<paragraph><location><page_1><loc_34><loc_84><loc_47><loc_89></location>dent Union) meels at 12.30 on Thursdays lor Noondays bible study) in Sludent Center 216 and at pm on Mondays for Jumpstart worship tme,in Sludent Center 205-206.</paragraph>
<subtitle-level-1><location><page_1><loc_35><loc_83><loc_47><loc_84></location>TcU CathoLiC CoMMU-</subtitle-level-1>
<paragraph><location><page_1><loc_34><loc_77><loc_47><loc_83></location>NITY presents supper program a( 5-30 0 very Thursday in the Faculty Center of Reed HallRoom 214 All are welcome The organization is having retreat during MidSemester Break (Oct. 13-16). Call University Ministries tor more information.</paragraph>
</document> </document>

2
tests/data_scanned/scanned_01.easyocr.json
View File

File diff suppressed because one or more lines are too long

241
tests/data_scanned/scanned_01.easyocr.md
View File

@ -1,240 +1,35 @@
Campuslines is provided as a service 1o the TCU community Announcements 01 @vents public meetings and other general campus Information should be brought by the Skitf otfice, Moudy 2915 or sent to TCU Box 32929. The Skitf reserves the rightto edit lor style and taste ## TableFormer: Table Structure Understanding with Transformers
DEPARTMENT OF MuSIC MASTER CLASS SERIES AnD JaZZ AND PERCUSSION DEPARTMENTS present Bob Brelthaupt In drumnsel and jazz master class Irom M toanoon today An open rehearsal with the TCU Jazz Ensemble Il beneldIrom noon t0 pmn ## 1. Details on the datasets
## Tcu INTERNATIONAL ## 1.1. Data preparation
ASSOCIATION will meet at 4 Pm loday in tne Bass Living Room Call Darlene Roth 9229448 As a first step of our data preparation process; we have calculated statistics over the datasets across the following dimensions: (1) table size measured in the number of rows and columns, (2) complexity of the table, (3) strictness of the provided HTML structure and (4) completeness (i.e. no omitted bounding boxes) A table is considered to be simple if it does not contain row spans or column spans. Addition ally, a table has a strict HTML structure if every row has the same number of columns after taking into account any row Or column spans. Therefore a strict HTML structure looks always rectangular: However; HTML is a lenient encoding format, i.e. tables with rows of different sizes might still be regarded as correct due to implicit display rules. These implicit rules leave room for ambiguity; which we want lo avoid. As such, we prefer to have strict" tables, i.e. tables where every row has exactly the same length.
## WEATHER We have developed technique that tries to derive missing bounding box out of its neighbors. As a first step; we use the annotation data to generate the most fine'grained that covers the table structure. In case of strict HTML tables. all squares are associated with some table cell and in the presence of table spans a cell extends across mul tiple grid squares. When enough bounding boxes are known for a rectangular table, it is possible to compute the geometrical border lines between the grid rows and columns. Eventually this information is used to generate the missing bounding boxes. Additionally; the existence of unused grid squares indicates that the table rows have unequal number of columns and the overall structure is non-strict. The generation of missing bounding boxes for non-strict HTML ta bles is ambiguous and therefore quite challenging. Thus, we have decided to simply discard those tables. In case of PubTabNet we have computed missing bounding boxes for 489 of the simple and 699 of the complex tables. RegardFinTabNet, 689 of the simple and 98% of the complex tables require the generation of bounding boxes grid grid ing
## CAMPUS CALENDAR Figure 7 illustrates the distribution of the tables across different dimensions per dataset.
Today's weather will be partly cloudy with highs in the 80s . ## 1.2. Synthetic datasets
On Thursday; look for highs in the 80s and lows in the 60s with a chance of thunderstorms Aiming t0 train and evaluate our models in a broader spectrum of table data we have synthesized four types of datasets_ Each one contains tables with different appear -
## CORRECTION Supplementary Material
Charlsie Mays was incorrectly identified as 0 junior in yesterday $ Skift . Mays is a senior. ances in regard to their size; structure, and content. synthetic dataset contains 150k examples, summing up to 60Ok synthetic examples. All datasets are divided into Train; Test and Val splits (8O%, 1O% , 109) . style Every
The process of generating a synthetic dataset can be decomposed into the following steps:
<!-- image --> 1 Prepare styling and content templates: The styling templates have been manually designed and organized into groups of scope specific appearances financial data, marketing data; etc.) Additionally; we have prepared curated collections of content templates by extracting the most frequently used terms out of non-synthetic datasets PubTabNet, FinTabNet, etc.). (e.g (e.g
Last Day t0 withdraw trom classes for Fall 2 Generate table structures: The structure of each synthetic dataset assumes a horizontal table header which potentially spans ovCr multiple rows and table body that may contain a combination of row spans and column spans. However, spans are not allowed to cross the header body boundary. The table structure is described by the parameters: Total number of table rows and columns, number of header rows, type of spans (header only spans, row only spans, column only spans, both row and column spans) maximum span size and the ratio of the table area covered by spans
## Today 3 Generate content: Based on the dataset theme. a set of suitable content templates is chosen first. Then; this content can be combined with purely random text to produce the synthetic content.
2 pm t0 6 pm , PC Visual ts Display: Elizabeth Leal Student Center Lounge 4 Apply styling templates: Depending on the domain of the synthetic dataset; a set of styling templates is first manually selected Then, style is randomly selected to format the appearance of the synthesized table.
Chapel, Noon University Rober Carr Chapel 5 Render the complete tables: The synthetic table is finally rendered by a web browser engine to generate the bounding boxes for each table cell. A batching technique is utilized to optimize the runtime overhead of the rendering process.
8 p Thealer TCU Tne Madwoman of Chaillot University Thealre ## 2. Prediction post-processing for PDF documents
amnto Pm The Artist andine Baseball Card . Moudy Bldg ernioil space Although TableFormer can predict the table structure and the bounding boxes for tables recognized inside PDF docu ments, this is not enough when a full reconstruction of the original table is required. This happens mainly due the folrcasons: lowing7
## Thursday
8 pm. Theatre TCU Ine Madwvoman of Cnaillot , Unis versity Thealre
Lucy Harris Linn Institute Noon 10 pn PC Visual Arts: Arts Festival, Frog Fountain
Mamnslo 4p Mn.e The Artisi
<!-- image -->
and the Baseball Card * Moudy Bldg exhibit space
Noon Honzon Awards luncheon Student Center Ballroom
5 pm International Students Association meeting Student Center
## Friday
PC TGIF: Karaoke
8 pm Theatre TCU ~The Madwoman of Chaillot; University Theatre
University Leadership Retreal
Student Account Paymenis Due
PC Film: "Jurassic Park"
1am1o pmn The Arist andine Baseball Card Moudy Bldg exnibil space
Noon (o pm , TCU Jazz Ensemble performance
<!-- image -->
<!-- image -->
## EMPLOYMENT |
Chlldren'* Clothing part-uno posilion ovak able Flenblo hours Store Mon -Sat open 10 00 10 500 Call The Mudpuppy 731-2581
[Now hiring part-lime holp [lor snow ski ropair and renlals, rcquirod, 2312 Monloomery . Exp
Fort Worth Ladies Specialty Boubque local cdin the Slockyard Stabon neods experonccd sales person Weckonds non-smnoker , 626-8900
TERM PaPERS TyPED FASTI Laser prinled. Rush oders and cedit cards accepled, Across trom TCU al 3023 South Univorslty Drive ACCURACY +PLUS 826 4969
Iromn 3 dillerent lundraise(s lasling elther 30 daysNo investment Earn SSS lor your group plus personal cash bonuses lor yoursell . Cal 1-800-932-0528, Exl65
Great Money-Grval Funl Become Flash photographer, must have 35mm camnera, will train FLASH 924-2626,
TYPING
Letters, lermn papers, theses, Or anything else you mnay nccd typed, 214-641-1021 call;
## [FUNDRAISINC]
FUNDRAISINGChoose
## FOR SALE
Soulhwost and ruslic lurniture cuslomn madeCall lorInloralon 922-9563
## The Adventuros of
Sperfrog
<!-- image -->
## Calvin and Hobbes
## by Bill Watterson
<!-- image -->
<!-- image -->
<!-- image -->
## TCU DAILY SKIFF Since 1902
Ine ICUDauy SkfisVocLcedbynie sludents 01 Teras Chnstan Universty sponsoredoytne journalism department 4a0 published Tuescoy throug) Fnday Cuning Inea Ong Spnnq Semesters encept dunngfnns heesand cunnghol days
Ine Sh C stbutec Iree 0n Camous silcevs lacully slatanc ns103
CiRculaTion Ooo Tuesoay through Frday
SUBSCRIPTIONS Togot suusohblion byma call 921-7000 and asklo enen 801 16274 Subscrphion rales are $20per mester
PHOTOGRAPHS Sial pholovaphs are avlabeo9 purcuse Dy teaders Quine ShifFor pniong nes conlactIne pholo desk Quce
EDITORIAL POLICY Uns gned edilonoae oonon Dage feoterenline Meaoethe Sheqtona DoardKncn composeoormne coior manag nqcoior assistant mnanaqing coior, ne4s edior
Camns edio conoed I0an crector and spons edito Signed leiters an0 columns representte Op nionolihe Vnii0
SKIFFTELEPHONE DIRECTORY Tne four-dioi ertens ons (6000 senes) Can numtoCis Can b (OOchco Dy dialing 921-7722 Idowed Dy the erens on
LETTERS To IHE EDITOR Tne Skff elcones lefters ione editorlor publica ton Letters'must b0tyoco doublesoaced Siq7e0 010 Fco Soowords
soaced Siq7e0 010 Soowords Lerers Shoabe suomnited teast Io cays Deloe publicalion 10 Ine Sk nensoon Moudy 2915 to TCU Bor 22929 0 to Yax 921-7133 must IncudeIne ollhor $ classihcalonmnajor and phone nunocr Ine Sa iesenes ne nqn lo edt € (Ejeci any lenters lo stye andlasie They
Leters to inc edtor are ulso accaoico vocoma Dorn anc Inrouqh [1 TCU conoer syslern Jolave Vokeener Cial921-7683 Toleave e ma seno to tne Suis TCUvaracoress Iistecbelo:
Skivy 1 member 01 Ine Assocaled Press The
MaILING ADDRESS Po 32929 Fodworh Teras 76129 80'
<!-- image -->
fpposios CaNTInA
Full/Part-time positions
Training provided
No experience necessary
APPLY IN PERsoN Tuesday & Wednesday 3:00 p.m. 5:00 p.m.
10433 LoMBARDY LN. (Now OPEN)
723 S. CentrAL ExpresswaY (RicharDsON)
Equal Opportunity Employer
## PAPPADEAUX SEAFOOD KITCHEN
No experience necessary
Full/Part-time positions Training provided
## Apply in person:
Tuesday & Wednesday 3.00 p.m: 5.00 pm.
Dallas: 3520 Oaklawn al Lemmon Irving: 10428 Lombardy Ln, (open soonl) Richardson: 725 $ Cenlral Expressway Arlinglon: 1304 Copeland Rd. at Collins Equal Opporlwuty Ernployer
Inquiring minds read the Skiff.
| Main nmber Ecioe s Uine | 921-7426 6266 |
|---------------------------|-----------------|
| Camnpus Dusk | 6267 |
| News Desk | 6269 |
| Sponts Desk | 6266 |
| Photo Des: | 6546 |
| Bac"shop Far | 6553 821-7133 |
Aner Holrs (Mianiqx '0 am) 921-7683 Sound OA Lineivo ce Mal 921-7683 Advertising 921-7426 Clossifiec 921-7426 Business Manager 6274 Subscnolions 6274 Stucent Plo catons Drecior 6558 Journalism Oflice 921-7425 Imnage Kagazine 821-7429 TCU Var (addruss 921-7161 ProduclionManger 821-7427 Ski)
What's news? Find out in the
TCU Daily Skiff
TRAFFIC TICKETS bulonlyin Fort
JAMEs R. MALLoRY JOosnesaï Law Ave Fort WorthTX 76107-1793 924-3236
defended orth Arlington; and clsewhere In arrant County No pornises aso esults Any nine andany count cosis arc hotinclodcd on fce lor legul rcprescntation
London $459
Paris 5479' Brussels 8549 [Fraakifurt 5579 [Milan 5659* [Stockholm 5699*| [Los Angeles $250' [Mexico City $198' Wuvobunl aaa A n uln
<!-- image -->
6715 Hillcrest Dallas , IX 75205 214-363-9941 Eurailpasses issued on-the-spotl
Shed GOIEN USed ioit, Ke Figured She Slould BEME To Oxe
HEY?
## by Ba Ranan & Adamn Wright
## CAMPUSLINES
AlcohOL And DRUG EDUCATIon RECOVERY SUPPORT GROUP meets at 5 pamn Wednesdays 1o Rickel 107 . Call 921-7100.
PSI CHI psychology honor sociely will havea general meetong al 6.15 pm today Winton-Scott 217 All interested slucenis are Invited to altend
FRENCH CLUB will meet al 6 30p m today in Student Center 218 1o discuss upcoming evenis
## FELLOWSHIP OF CHRISTIAN
## RAPEISEXUAL ASSAULT
ATHLETES welcomes Ken Horton of McKinney Bible Church at 9 pm tonight in the Club Roomn ol thc Coliseum. Everyone is welcome,
SURVIVORS GROUP meeting from 2 t0 3.30 pm on Fridays , is forming at the TCU Counseling Center Call Dorothy M Barra at 92 -7863 {0 set up an initial screening appointment
BAPTIST STUDENT MINISTRY (previously Baptist Stu-
dent Union) meels at 12.30 on Thursdays lor Noondays bible study) in Sludent Center 216 and at pm on Mondays for Jumpstart worship tme,in Sludent Center 205-206.
## TcU CathoLiC CoMMU-
NITY presents supper program a( 5-30 0 very Thursday in the Faculty Center of Reed HallRoom 214 All are welcome The organization is having retreat during MidSemester Break (Oct. 13-16). Call University Ministries tor more information.

2
tests/data_scanned/scanned_01.easyocr.pages.json
View File

File diff suppressed because one or more lines are too long

BIN
tests/data_scanned/scanned_01.pdf
View File

Binary file not shown.

BIN
tests/data_scanned/scanned_01.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 2.3 MiB

After

Width:  |  Height:  |  Size: 496 KiB

20
tests/data_scanned/scanned_01.tesseract.doctags.txt Normal file
View File

@ -0,0 +1,20 @@
<document>
<subtitle-level-1><location><page_1><loc_21><loc_84><loc_76><loc_87></location>TableFormer: Table Structure Understanding with Transformers</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_8><loc_78><loc_29><loc_80></location>1. Details on the datasets</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_8><loc_76><loc_25><loc_78></location>1.1. Data preparation</subtitle-level-1>
<paragraph><location><page_1><loc_8><loc_51><loc_47><loc_75></location>As a first step of our data preparation process; we have calculated statistics over the datasets across the following dimensions: (1) table size measured in the number of rows and columns, (2) complexity of the table, (3) strictness of the provided HTML structure and (4) completeness (i.e. no omitted bounding boxes) A table is considered to be simple if it does not contain row spans or column spans. Addition ally, a table has a strict HTML structure if every row has the same number of columns after taking into account any row Or column spans. Therefore a strict HTML structure looks always rectangular: However; HTML is a lenient encoding format, i.e. tables with rows of different sizes might still be regarded as correct due to implicit display rules. These implicit rules leave room for ambiguity; which we want lo avoid. As such, we prefer to have strict" tables, i.e. tables where every row has exactly the same length.</paragraph>
<paragraph><location><page_1><loc_8><loc_20><loc_47><loc_51></location>We have developed technique that tries to derive missing bounding box out of its neighbors. As a first step; we use the annotation data to generate the most fine'grained that covers the table structure. In case of strict HTML tables. all squares are associated with some table cell and in the presence of table spans a cell extends across mul tiple grid squares. When enough bounding boxes are known for a rectangular table, it is possible to compute the geometrical border lines between the grid rows and columns. Eventually this information is used to generate the missing bounding boxes. Additionally; the existence of unused grid squares indicates that the table rows have unequal number of columns and the overall structure is non-strict. The generation of missing bounding boxes for non-strict HTML ta bles is ambiguous and therefore quite challenging. Thus, we have decided to simply discard those tables. In case of PubTabNet we have computed missing bounding boxes for 489 of the simple and 699 of the complex tables. RegardFinTabNet, 689 of the simple and 98% of the complex tables require the generation of bounding boxes grid grid ing</paragraph>
<paragraph><location><page_1><loc_8><loc_18><loc_47><loc_21></location>Figure 7 illustrates the distribution of the tables across different dimensions per dataset.</paragraph>
<subtitle-level-1><location><page_1><loc_8><loc_15><loc_25><loc_17></location>1.2. Synthetic datasets</subtitle-level-1>
<paragraph><location><page_1><loc_8><loc_10><loc_47><loc_14></location>Aiming t0 train and evaluate our models in a broader spectrum of table data we have synthesized four types of datasets_ Each one contains tables with different appear -</paragraph>
<paragraph><location><page_1><loc_36><loc_82><loc_62><loc_85></location>Supplementary Material</paragraph>
<paragraph><location><page_1><loc_50><loc_74><loc_89><loc_80></location>ances in regard to their size; structure, and content. synthetic dataset contains 150k examples, summing up to 60Ok synthetic examples. All datasets are divided into Train; Test and Val splits (8O%, 1O% , 109) . style Every</paragraph>
<paragraph><location><page_1><loc_50><loc_70><loc_89><loc_74></location>The process of generating a synthetic dataset can be decomposed into the following steps:</paragraph>
<paragraph><location><page_1><loc_50><loc_60><loc_89><loc_71></location>1 Prepare styling and content templates: The styling templates have been manually designed and organized into groups of scope specific appearances financial data, marketing data; etc.) Additionally; we have prepared curated collections of content templates by extracting the most frequently used terms out of non-synthetic datasets PubTabNet, FinTabNet, etc.). (e.g (e.g</paragraph>
<paragraph><location><page_1><loc_50><loc_43><loc_89><loc_60></location>2 Generate table structures: The structure of each synthetic dataset assumes a horizontal table header which potentially spans ovCr multiple rows and table body that may contain a combination of row spans and column spans. However, spans are not allowed to cross the header body boundary. The table structure is described by the parameters: Total number of table rows and columns, number of header rows, type of spans (header only spans, row only spans, column only spans, both row and column spans) maximum span size and the ratio of the table area covered by spans</paragraph>
<paragraph><location><page_1><loc_50><loc_37><loc_89><loc_43></location>3 Generate content: Based on the dataset theme. a set of suitable content templates is chosen first. Then; this content can be combined with purely random text to produce the synthetic content.</paragraph>
<paragraph><location><page_1><loc_50><loc_31><loc_89><loc_37></location>4 Apply styling templates: Depending on the domain of the synthetic dataset; a set of styling templates is first manually selected Then, style is randomly selected to format the appearance of the synthesized table.</paragraph>
<paragraph><location><page_1><loc_50><loc_23><loc_89><loc_31></location>5 Render the complete tables: The synthetic table is finally rendered by a web browser engine to generate the bounding boxes for each table cell. A batching technique is utilized to optimize the runtime overhead of the rendering process.</paragraph>
<subtitle-level-1><location><page_1><loc_50><loc_18><loc_89><loc_22></location>2. Prediction post-processing for PDF documents</subtitle-level-1>
<paragraph><location><page_1><loc_50><loc_9><loc_89><loc_17></location>Although TableFormer can predict the table structure and the bounding boxes for tables recognized inside PDF docu ments, this is not enough when a full reconstruction of the original table is required. This happens mainly due the folrcasons: lowing7</paragraph>
</document>

1
tests/data_scanned/scanned_01.tesseract.json Normal file
View File

File diff suppressed because one or more lines are too long

35
tests/data_scanned/scanned_01.tesseract.md Normal file
View File

@ -0,0 +1,35 @@
## TableFormer: Table Structure Understanding with Transformers
## 1. Details on the datasets
## 1.1. Data preparation
As a first step of our data preparation process; we have calculated statistics over the datasets across the following dimensions: (1) table size measured in the number of rows and columns, (2) complexity of the table, (3) strictness of the provided HTML structure and (4) completeness (i.e. no omitted bounding boxes) A table is considered to be simple if it does not contain row spans or column spans. Addition ally, a table has a strict HTML structure if every row has the same number of columns after taking into account any row Or column spans. Therefore a strict HTML structure looks always rectangular: However; HTML is a lenient encoding format, i.e. tables with rows of different sizes might still be regarded as correct due to implicit display rules. These implicit rules leave room for ambiguity; which we want lo avoid. As such, we prefer to have strict" tables, i.e. tables where every row has exactly the same length.
We have developed technique that tries to derive missing bounding box out of its neighbors. As a first step; we use the annotation data to generate the most fine'grained that covers the table structure. In case of strict HTML tables. all squares are associated with some table cell and in the presence of table spans a cell extends across mul tiple grid squares. When enough bounding boxes are known for a rectangular table, it is possible to compute the geometrical border lines between the grid rows and columns. Eventually this information is used to generate the missing bounding boxes. Additionally; the existence of unused grid squares indicates that the table rows have unequal number of columns and the overall structure is non-strict. The generation of missing bounding boxes for non-strict HTML ta bles is ambiguous and therefore quite challenging. Thus, we have decided to simply discard those tables. In case of PubTabNet we have computed missing bounding boxes for 489 of the simple and 699 of the complex tables. RegardFinTabNet, 689 of the simple and 98% of the complex tables require the generation of bounding boxes grid grid ing
Figure 7 illustrates the distribution of the tables across different dimensions per dataset.
## 1.2. Synthetic datasets
Aiming t0 train and evaluate our models in a broader spectrum of table data we have synthesized four types of datasets_ Each one contains tables with different appear -
Supplementary Material
ances in regard to their size; structure, and content. synthetic dataset contains 150k examples, summing up to 60Ok synthetic examples. All datasets are divided into Train; Test and Val splits (8O%, 1O% , 109) . style Every
The process of generating a synthetic dataset can be decomposed into the following steps:
1 Prepare styling and content templates: The styling templates have been manually designed and organized into groups of scope specific appearances financial data, marketing data; etc.) Additionally; we have prepared curated collections of content templates by extracting the most frequently used terms out of non-synthetic datasets PubTabNet, FinTabNet, etc.). (e.g (e.g
2 Generate table structures: The structure of each synthetic dataset assumes a horizontal table header which potentially spans ovCr multiple rows and table body that may contain a combination of row spans and column spans. However, spans are not allowed to cross the header body boundary. The table structure is described by the parameters: Total number of table rows and columns, number of header rows, type of spans (header only spans, row only spans, column only spans, both row and column spans) maximum span size and the ratio of the table area covered by spans
3 Generate content: Based on the dataset theme. a set of suitable content templates is chosen first. Then; this content can be combined with purely random text to produce the synthetic content.
4 Apply styling templates: Depending on the domain of the synthetic dataset; a set of styling templates is first manually selected Then, style is randomly selected to format the appearance of the synthesized table.
5 Render the complete tables: The synthetic table is finally rendered by a web browser engine to generate the bounding boxes for each table cell. A batching technique is utilized to optimize the runtime overhead of the rendering process.
## 2. Prediction post-processing for PDF documents
Although TableFormer can predict the table structure and the bounding boxes for tables recognized inside PDF docu ments, this is not enough when a full reconstruction of the original table is required. This happens mainly due the folrcasons: lowing7

1
tests/data_scanned/scanned_01.tesseract.pages.json Normal file
View File

File diff suppressed because one or more lines are too long

171
tests/data_scanned/scanned_01.tesserocr.doctags.txt
View File

@ -1,155 +1,20 @@
<document> <document>
<paragraph><location><page_1><loc_6><loc_82><loc_19><loc_88></location>Campusiines is provided as a service to the TCU community. Announcements of events, public meetings and other general campus infor mation should be brought by the Skiff office, Moudy 291S or sent to TCU Box 32929 The Skiff reserves the right to edit for style and taste</paragraph> <subtitle-level-1><location><page_1><loc_22><loc_85><loc_76><loc_86></location>TableFormer: Table Structure Understanding with Transformers</subtitle-level-1>
<paragraph><location><page_1><loc_6><loc_75><loc_19><loc_81></location>DEPARTMENT OF MUSIC MASTER CLASS SERIES AND JAZZ AND PERCUSSION DEPARTMENTS pr</paragraph> <paragraph><location><page_1><loc_36><loc_83><loc_61><loc_85></location>Supplementary Material</paragraph>
<subtitle-level-1><location><page_1><loc_7><loc_73><loc_19><loc_74></location>TCU INTE RNATIONAL</subtitle-level-1> <subtitle-level-1><location><page_1><loc_8><loc_79><loc_29><loc_80></location>1. Details on the datasets</subtitle-level-1>
<paragraph><location><page_1><loc_6><loc_71><loc_19><loc_73></location>ASSOCIATION</paragraph> <subtitle-level-1><location><page_1><loc_8><loc_76><loc_25><loc_77></location>1.1. Data preparation</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_7><loc_67><loc_16><loc_68></location>WEATHER</subtitle-level-1> <paragraph><location><page_1><loc_8><loc_51><loc_47><loc_75></location>As a first step of our data preparation process, we have calculated statistics over the datasets across the following dimensions: (1) table size measured in the number of rows and columns, (2) complexity of the table, (3) strictness of the provided HTML structure and (4) completeness (i.e. no omitted bounding boxes). A table is considered to be simple if it does not contain row spans or column spans. Additionally, a table has a strict HTML structure if every row has the same number of columns after taking into account any row or column spans. Therefore a strict HTML structure looks always rectangular. However, HTML is a lenient encoding format, i.e. tables with rows of different sizes might still be regarded as correct due to implicit display rules. These implicit rules leave room for ambiguity, which we want to avoid. As such, we prefer to have "strict" tables, i.e. tables where every row has exactly the same length.</paragraph>
<subtitle-level-1><location><page_1><loc_6><loc_56><loc_18><loc_57></location>CORRECTION</subtitle-level-1> <paragraph><location><page_1><loc_8><loc_21><loc_47><loc_51></location>We have developed a technique that tries to derive a missing bounding box out of its neighbors. As a first step, we use the annotation data to generate the most fine-grained grid that covers the table structure. In case of strict HTML tables, all grid squares are associated with some table cell and in the presence of table spans a cell extends across multiple grid squares. When enough bounding boxes are known for a rectangular table, it is possible to compute the geometrical border lines between the grid rows and columns. Eventually this information is used to generate the missing bounding boxes. Additionally, the existence of unused grid squares indicates that the table rows have unequal number of columns and the overall structure is non-strict. The generation of missing bounding boxes for non-strict HTML tables is ambiguous and therefore quite challenging. Thus, we have decided to simply discard those tables. In case of PubTabNet we have computed missing bounding boxes for 48% of the simple and 69% of the complex tables. Regarding FinTabNet, 68% of the simple and 98% of the complex tables require the generation of bounding boxes.</paragraph>
<figure> <paragraph><location><page_1><loc_8><loc_18><loc_47><loc_21></location>Figure 7 illustrates the distribution of the tables across different dimensions per dataset.</paragraph>
<location><page_1><loc_5><loc_47><loc_19><loc_50></location> <subtitle-level-1><location><page_1><loc_8><loc_15><loc_25><loc_16></location>1.2. Synthetic datasets</subtitle-level-1>
</figure> <paragraph><location><page_1><loc_8><loc_10><loc_47><loc_14></location>Aiming to train and evaluate our models in a broader spectrum of table data we have synthesized four types of datasets. Each one contains tables with different appear-</paragraph>
<figure> <paragraph><location><page_1><loc_50><loc_74><loc_89><loc_80></location>ances in regard to their size, structure, style and content. Every synthetic dataset contains 150k examples, summing up to 600k synthetic examples. All datasets are divided into Train, Test and Val splits (80%, 10%, 10%).</paragraph>
<location><page_1><loc_5><loc_25><loc_34><loc_47></location> <paragraph><location><page_1><loc_50><loc_71><loc_89><loc_73></location>The process of generating a synthetic dataset can be decomposed into the following steps:</paragraph>
</figure> <paragraph><location><page_1><loc_50><loc_60><loc_89><loc_70></location>1. Prepare styling and content templates: The styling templates have been manually designed and organized into groups of scope specific appearances (e.g. financial data, marketing data, etc.) Additionally, we have prepared curated collections of content templates by extracting the most frequently used terms out of non-synthetic datasets (e.g. PubTabNet, FinTabNet, etc.).</paragraph>
<figure> <paragraph><location><page_1><loc_50><loc_43><loc_89><loc_60></location>2. Generate table structures: The structure of each synthetic dataset assumes a horizontal table header which potentially spans over multiple rows and a table body that may contain a combination of row spans and column spans. However, spans are not allowed to cross the header -body boundary. The table structure is described by the parameters: Total number of table rows and columns, number of header rows, type of spans (header only spans, row only spans, column only spans, both row and column spans), maximum span size and the ratio of the table area covered by spans.</paragraph>
<location><page_1><loc_35><loc_25><loc_49><loc_47></location> <paragraph><location><page_1><loc_50><loc_37><loc_89><loc_43></location>3. Generate content: Based on the dataset theme, a set of suitable content templates is chosen first. Then, this content can be combined with purely random text to produce the synthetic content.</paragraph>
</figure> <paragraph><location><page_1><loc_50><loc_31><loc_89><loc_37></location>4. Apply styling templates: Depending on the domain of the synthetic dataset, a set of styling templates is first manually selected. Then, a style is randomly selected to format the appearance of the synthesized table.</paragraph>
<figure> <paragraph><location><page_1><loc_50><loc_23><loc_89><loc_31></location>5. Render the complete tables: The synthetic table is finally rendered by a web browser engine to generate the bounding boxes for each table cell. A batching technique is utilized to optimize the runtime overhead of the rendering process.</paragraph>
<location><page_1><loc_5><loc_17><loc_48><loc_24></location> <subtitle-level-1><location><page_1><loc_50><loc_19><loc_89><loc_22></location>2. Prediction post-processing for PDF documents</subtitle-level-1>
</figure> <paragraph><location><page_1><loc_50><loc_10><loc_89><loc_17></location>Although TableFormer can predict the table structure and the bounding boxes for tables recognized inside PDF documents, this is not enough when a full reconstruction of the original table is required. This happens mainly due the following reasons:</paragraph>
<subtitle-level-1><location><page_1><loc_6><loc_15><loc_15><loc_16></location>EMPLOYMENT</subtitle-level-1>
<paragraph><location><page_1><loc_5><loc_10><loc_15><loc_14></location>10:00 to 5:00. Call The Mudpuppy 731-2581</paragraph>
<paragraph><location><page_1><loc_5><loc_7><loc_15><loc_9></location>»</paragraph>
<paragraph><location><page_1><loc_16><loc_12><loc_26><loc_17></location>Fort Worth Ladies Specialty Boutique located in the Stockyard Station needs experenced sales person Weekends, non-smoker 626-8900</paragraph>
<paragraph><location><page_1><loc_16><loc_8><loc_26><loc_11></location>Great Money--Great Fun! Become a Flash photographer, must have 35mm camera, will train FLASH, 924-2626</paragraph>
<paragraph><location><page_1><loc_27><loc_9><loc_37><loc_11></location>Letters, term papers, the$@5, oF anything else you may need typed, call 214-641-1023. 1021</paragraph>
<subtitle-level-1><location><page_1><loc_27><loc_7><loc_37><loc_8></location>a</subtitle-level-1>
<paragraph><location><page_1><loc_27><loc_12><loc_37><loc_17></location>TERM PAPERS TYPED FAST! Laser printed Rush orders and credit cards accepted. Across from TCU at 3023 South University Drive. ACCURACY +PLUS, 926-4969</paragraph>
<paragraph><location><page_1><loc_38><loc_12><loc_48><loc_17></location>plus personal cash bonuses for yourself Call 1-800-932-0528, Ext. 65. from 3 different fundraisers lasting either 3 or 7 days. No investment Earn $$$ for your group</paragraph>
<subtitle-level-1><location><page_1><loc_40><loc_9><loc_48><loc_11></location>-</subtitle-level-1>
<paragraph><location><page_1><loc_38><loc_7><loc_48><loc_9></location>Southwest and rustic furfor information, 922-9563.</paragraph>
<paragraph><location><page_1><loc_16><loc_9><loc_16><loc_9></location>|</paragraph>
<subtitle-level-1><location><page_1><loc_20><loc_89><loc_33><loc_90></location>CAMPUSLINES</subtitle-level-1>
<paragraph><location><page_1><loc_20><loc_85><loc_33><loc_88></location>ALCOHOL AND DRUG EDUCATION 1 pedehag en} SUPPORT GROUP at § Nednesdays in Ricke 3</paragraph>
<paragraph><location><page_1><loc_20><loc_81><loc_33><loc_84></location>PSI CHI psyct are</paragraph>
<paragraph><location><page_1><loc_20><loc_77><loc_33><loc_80></location>FRENCH CLUB w fay</paragraph>
<subtitle-level-1><location><page_1><loc_20><loc_76><loc_33><loc_77></location>FELLOWSHIP OF owes te</subtitle-level-1>
<paragraph><location><page_1><loc_20><loc_73><loc_33><loc_77></location>te AT HLETES © we</paragraph>
<paragraph><location><page_1><loc_20><loc_71><loc_33><loc_72></location>BAPTIST STUDENT MINISTRY i tu</paragraph>
<subtitle-level-1><location><page_1><loc_21><loc_66><loc_24><loc_66></location>Today</subtitle-level-1>
<paragraph><location><page_1><loc_27><loc_9><loc_27><loc_10></location>|</paragraph>
<paragraph><location><page_1><loc_27><loc_9><loc_27><loc_9></location>|</paragraph>
<paragraph><location><page_1><loc_34><loc_85><loc_47><loc_88></location>dent Union Thursd dible st meets at 12:30 on ays for Noondays (a udy) in Student Center 216 and at 7 p.m. on Mondays for Jume start, a worship time, in Student Center 20 5-206</paragraph>
<subtitle-level-1><location><page_1><loc_35><loc_83><loc_47><loc_84></location>TCU CATHOLIC COMMU-</subtitle-level-1>
<paragraph><location><page_1><loc_34><loc_77><loc_47><loc_83></location>NITY p gran Supper pro fr every nthe Fa ulty Center 14. Al are rmanizator eat during Mid fr Break et. 13-1 format a</paragraph>
<subtitle-level-1><location><page_1><loc_35><loc_76><loc_47><loc_77></location>RAPE/SEXUAL ASSAULT</subtitle-level-1>
<paragraph><location><page_1><loc_34><loc_72><loc_47><loc_76></location>SURVIVORS GROUP meeting</paragraph>
<figure>
<location><page_1><loc_49><loc_80><loc_91><loc_89></location>
</figure>
<subtitle-level-1><location><page_1><loc_81><loc_79><loc_93><loc_80></location>by Bill Watterson</subtitle-level-1>
<figure>
<location><page_1><loc_49><loc_70><loc_60><loc_78></location>
</figure>
<paragraph><location><page_1><loc_50><loc_66><loc_50><loc_67></location>|</paragraph>
<paragraph><location><page_1><loc_50><loc_58><loc_64><loc_59></location>SIRCULATION 4</paragraph>
<paragraph><location><page_1><loc_50><loc_56><loc_64><loc_58></location>SUBSCRIPTIONS</paragraph>
<paragraph><location><page_1><loc_50><loc_53><loc_64><loc_55></location>PHOTOGRAPHS if t '</paragraph>
<paragraph><location><page_1><loc_50><loc_50><loc_64><loc_52></location>EDITORIAL POLICY edit t Oe t</paragraph>
<figure>
<location><page_1><loc_71><loc_64><loc_91><loc_78></location>
</figure>
<paragraph><location><page_1><loc_65><loc_60><loc_78><loc_62></location>LETTERS TO THE EDITOR e Skiff</paragraph>
<paragraph><location><page_1><loc_65><loc_56><loc_78><loc_60></location>1 ] ! ef blication to the Skiff the ' ficat naior</paragraph>
<paragraph><location><page_1><loc_65><loc_53><loc_78><loc_56></location>e ee wea coejte @ mai form and through the { . 4</paragraph>
<paragraph><location><page_1><loc_65><loc_51><loc_78><loc_53></location>f kiff nember of the sted Pre</paragraph>
<paragraph><location><page_1><loc_64><loc_50><loc_78><loc_51></location>MAILING ADDRESS Ff</paragraph>
<figure>
<location><page_1><loc_49><loc_34><loc_78><loc_47></location>
</figure>
<paragraph><location><page_1><loc_69><loc_32><loc_70><loc_33></location>S</paragraph>
<paragraph><location><page_1><loc_66><loc_30><loc_70><loc_30></location>CANTINA</paragraph>
<paragraph><location><page_1><loc_57><loc_28><loc_71><loc_29></location>* Full/Part-time positions</paragraph>
<paragraph><location><page_1><loc_57><loc_27><loc_68><loc_28></location>* Training provided</paragraph>
<paragraph><location><page_1><loc_57><loc_26><loc_72><loc_27></location>* No experience necessary</paragraph>
<paragraph><location><page_1><loc_58><loc_25><loc_59><loc_26></location>IN</paragraph>
<paragraph><location><page_1><loc_59><loc_25><loc_63><loc_26></location>PERSON</paragraph>
<paragraph><location><page_1><loc_60><loc_24><loc_61><loc_25></location>3:00</paragraph>
<paragraph><location><page_1><loc_62><loc_24><loc_64><loc_25></location>p.m.</paragraph>
<paragraph><location><page_1><loc_64><loc_25><loc_67><loc_26></location>Tuesday</paragraph>
<paragraph><location><page_1><loc_64><loc_24><loc_64><loc_24></location>-</paragraph>
<paragraph><location><page_1><loc_65><loc_24><loc_67><loc_25></location>5:00</paragraph>
<paragraph><location><page_1><loc_68><loc_25><loc_68><loc_26></location>&</paragraph>
<paragraph><location><page_1><loc_67><loc_24><loc_69><loc_25></location>p.m.</paragraph>
<paragraph><location><page_1><loc_53><loc_23><loc_71><loc_24></location>* 10433 Lompanoy Ln. (Now Open)</paragraph>
<paragraph><location><page_1><loc_54><loc_23><loc_54><loc_23></location>*</paragraph>
<paragraph><location><page_1><loc_54><loc_22><loc_56><loc_23></location>723</paragraph>
<paragraph><location><page_1><loc_57><loc_22><loc_58><loc_23></location>S.</paragraph>
<paragraph><location><page_1><loc_58><loc_22><loc_62><loc_23></location>Centrar</paragraph>
<paragraph><location><page_1><loc_62><loc_22><loc_68><loc_23></location>Expressway</paragraph>
<paragraph><location><page_1><loc_69><loc_22><loc_75><loc_23></location>(Richaroson)</paragraph>
<paragraph><location><page_1><loc_59><loc_21><loc_70><loc_22></location>Equa/l Opportunity Employer</paragraph>
<subtitle-level-1><location><page_1><loc_53><loc_16><loc_75><loc_20></location>PAPPADEALLY SEAFOOD KITCHEN</subtitle-level-1>
<paragraph><location><page_1><loc_57><loc_16><loc_57><loc_16></location>*</paragraph>
<paragraph><location><page_1><loc_57><loc_15><loc_57><loc_15></location>*</paragraph>
<paragraph><location><page_1><loc_63><loc_15><loc_71><loc_16></location>positions provided</paragraph>
<paragraph><location><page_1><loc_58><loc_16><loc_65><loc_16></location>Full/Part-time</paragraph>
<paragraph><location><page_1><loc_58><loc_15><loc_63><loc_16></location>Training</paragraph>
<paragraph><location><page_1><loc_57><loc_14><loc_72><loc_15></location>* No experience necessary</paragraph>
<subtitle-level-1><location><page_1><loc_59><loc_13><loc_69><loc_14></location>Apply in person:</subtitle-level-1>
<paragraph><location><page_1><loc_53><loc_12><loc_76><loc_13></location>Tuesday & Wednesday, 3:00 p.m. -5:00 p.m.</paragraph>
<paragraph><location><page_1><loc_55><loc_8><loc_73><loc_12></location>Dallas: 3520 Oaklawn at Lemmon Irving: 10428 Lombardy Ln. (open soon!) Richardson: 725 S. Central Expressway Arlington: 1304 Copeland Rd. at Collins Equal Opportunity Employer</paragraph>
<paragraph><location><page_1><loc_52><loc_6><loc_76><loc_7></location>Inquiring minds read the Skiff.</paragraph>
<paragraph><location><page_1><loc_69><loc_25><loc_74><loc_26></location>Wednesday</paragraph>
<figure>
<location><page_1><loc_60><loc_69><loc_71><loc_78></location>
</figure>
<subtitle-level-1><location><page_1><loc_63><loc_65><loc_80><loc_67></location>TCU DAILY SKIFF Since 1902</subtitle-level-1>
<paragraph><location><page_1><loc_79><loc_63><loc_92><loc_65></location>SKIFF phenyl en OMRSCTORY nm be reactwd by Giahrig</paragraph>
<paragraph><location><page_1><loc_79><loc_58><loc_80><loc_58></location>After</paragraph>
<paragraph><location><page_1><loc_81><loc_58><loc_81><loc_58></location>H</paragraph>
<paragraph><location><page_1><loc_81><loc_55><loc_82><loc_56></location>sife</paragraph>
<paragraph><location><page_1><loc_79><loc_55><loc_81><loc_55></location>Busine</paragraph>
<paragraph><location><page_1><loc_82><loc_55><loc_85><loc_55></location>Manag</paragraph>
<paragraph><location><page_1><loc_80><loc_54><loc_81><loc_54></location>'¢</paragraph>
<paragraph><location><page_1><loc_80><loc_53><loc_82><loc_53></location>uM"airs:</paragraph>
<paragraph><location><page_1><loc_79><loc_52><loc_81><loc_53></location>Image TCU</paragraph>
<paragraph><location><page_1><loc_84><loc_54><loc_84><loc_54></location>at</paragraph>
<paragraph><location><page_1><loc_83><loc_53><loc_84><loc_53></location>Jt</paragraph>
<paragraph><location><page_1><loc_81><loc_53><loc_84><loc_53></location>Magazine</paragraph>
<paragraph><location><page_1><loc_79><loc_51><loc_82><loc_52></location>Vax Product</paragraph>
<paragraph><location><page_1><loc_83><loc_52><loc_85><loc_52></location>(address:</paragraph>
<paragraph><location><page_1><loc_83><loc_51><loc_85><loc_52></location>Manager</paragraph>
<paragraph><location><page_1><loc_82><loc_46><loc_86><loc_46></location>What's</paragraph>
<paragraph><location><page_1><loc_86><loc_46><loc_90><loc_46></location>news?</paragraph>
<paragraph><location><page_1><loc_82><loc_45><loc_84><loc_45></location>Find</paragraph>
<paragraph><location><page_1><loc_85><loc_45><loc_86><loc_45></location>out</paragraph>
<paragraph><location><page_1><loc_80><loc_43><loc_83><loc_44></location>TCU</paragraph>
<paragraph><location><page_1><loc_87><loc_45><loc_88><loc_45></location>in</paragraph>
<paragraph><location><page_1><loc_84><loc_43><loc_88><loc_44></location>Daily</paragraph>
<paragraph><location><page_1><loc_88><loc_43><loc_92><loc_44></location>Skiff</paragraph>
<paragraph><location><page_1><loc_80><loc_40><loc_85><loc_41></location>TRAFFIC</paragraph>
<paragraph><location><page_1><loc_86><loc_40><loc_92><loc_41></location>TICKETS</paragraph>
<paragraph><location><page_1><loc_80><loc_37><loc_92><loc_39></location>|iiigeaeme clsewher re in Tarrant County. ) promuses as to results Any fine and any court costs are not included fee for le zal repres entation</paragraph>
<paragraph><location><page_1><loc_84><loc_36><loc_85><loc_37></location>R.</paragraph>
<paragraph><location><page_1><loc_80><loc_36><loc_84><loc_37></location>James</paragraph>
<paragraph><location><page_1><loc_83><loc_35><loc_86><loc_36></location>Atiorney</paragraph>
<paragraph><location><page_1><loc_81><loc_33><loc_91><loc_35></location>24 Sandage Fort Worth, TX 76109-1793 2203236</paragraph>
<paragraph><location><page_1><loc_86><loc_35><loc_92><loc_37></location>M. ALLORY at La i</paragraph>
<paragraph><location><page_1><loc_80><loc_14><loc_92><loc_18></location>Stockholm $699" Los Angeles $250" Mexico City $198" See oc caret bor 0 Wort Ged on md wed bord toe Lyall wrortivante</paragraph>
<figure>
<location><page_1><loc_82><loc_10><loc_90><loc_14></location>
</figure>
<paragraph><location><page_1><loc_82><loc_9><loc_90><loc_10></location>6715 Hillcrest Dallas, TX 75205</paragraph>
<paragraph><location><page_1><loc_80><loc_7><loc_92><loc_8></location>214-363-9941</paragraph>
<paragraph><location><page_1><loc_82><loc_6><loc_90><loc_7></location>Eurailpasses</paragraph>
<paragraph><location><page_1><loc_88><loc_45><loc_90><loc_45></location>the</paragraph>
<paragraph><location><page_1><loc_82><loc_76><loc_93><loc_78></location>ios ][N © on TT ad or HE SHOULD | me</paragraph>
<paragraph><location><page_1><loc_90><loc_76><loc_93><loc_76></location>ee</paragraph>
<paragraph><location><page_1><loc_89><loc_73><loc_93><loc_75></location>WEY!</paragraph>
<paragraph><location><page_1><loc_88><loc_69><loc_92><loc_71></location>win Po iS</paragraph>
<paragraph><location><page_1><loc_54><loc_25><loc_57><loc_26></location>Appty</paragraph>
<paragraph><location><page_1><loc_79><loc_95><loc_79><loc_96></location>:</paragraph>
<paragraph><location><page_1><loc_86><loc_52><loc_86><loc_52></location>Sk</paragraph>
<paragraph><location><page_1><loc_99><loc_92><loc_100><loc_93></location>g-/</paragraph>
</document> </document>

2
tests/data_scanned/scanned_01.tesserocr.json
View File

File diff suppressed because one or more lines are too long

272
tests/data_scanned/scanned_01.tesserocr.md
View File

@ -1,271 +1,35 @@
Campusiines is provided as a service to the TCU community. Announcements of events, public meetings and other general campus infor mation should be brought by the Skiff office, Moudy 291S or sent to TCU Box 32929 The Skiff reserves the right to edit for style and taste ## TableFormer: Table Structure Understanding with Transformers
DEPARTMENT OF MUSIC MASTER CLASS SERIES AND JAZZ AND PERCUSSION DEPARTMENTS pr Supplementary Material
## TCU INTE RNATIONAL ## 1. Details on the datasets
ASSOCIATION ## 1.1. Data preparation
## WEATHER As a first step of our data preparation process, we have calculated statistics over the datasets across the following dimensions: (1) table size measured in the number of rows and columns, (2) complexity of the table, (3) strictness of the provided HTML structure and (4) completeness (i.e. no omitted bounding boxes). A table is considered to be simple if it does not contain row spans or column spans. Additionally, a table has a strict HTML structure if every row has the same number of columns after taking into account any row or column spans. Therefore a strict HTML structure looks always rectangular. However, HTML is a lenient encoding format, i.e. tables with rows of different sizes might still be regarded as correct due to implicit display rules. These implicit rules leave room for ambiguity, which we want to avoid. As such, we prefer to have "strict" tables, i.e. tables where every row has exactly the same length.
## CORRECTION We have developed a technique that tries to derive a missing bounding box out of its neighbors. As a first step, we use the annotation data to generate the most fine-grained grid that covers the table structure. In case of strict HTML tables, all grid squares are associated with some table cell and in the presence of table spans a cell extends across multiple grid squares. When enough bounding boxes are known for a rectangular table, it is possible to compute the geometrical border lines between the grid rows and columns. Eventually this information is used to generate the missing bounding boxes. Additionally, the existence of unused grid squares indicates that the table rows have unequal number of columns and the overall structure is non-strict. The generation of missing bounding boxes for non-strict HTML tables is ambiguous and therefore quite challenging. Thus, we have decided to simply discard those tables. In case of PubTabNet we have computed missing bounding boxes for 48% of the simple and 69% of the complex tables. Regarding FinTabNet, 68% of the simple and 98% of the complex tables require the generation of bounding boxes.
Figure 7 illustrates the distribution of the tables across different dimensions per dataset.
<!-- image --> ## 1.2. Synthetic datasets
Aiming to train and evaluate our models in a broader spectrum of table data we have synthesized four types of datasets. Each one contains tables with different appear-
<!-- image --> ances in regard to their size, structure, style and content. Every synthetic dataset contains 150k examples, summing up to 600k synthetic examples. All datasets are divided into Train, Test and Val splits (80%, 10%, 10%).
The process of generating a synthetic dataset can be decomposed into the following steps:
<!-- image --> 1. Prepare styling and content templates: The styling templates have been manually designed and organized into groups of scope specific appearances (e.g. financial data, marketing data, etc.) Additionally, we have prepared curated collections of content templates by extracting the most frequently used terms out of non-synthetic datasets (e.g. PubTabNet, FinTabNet, etc.).
2. Generate table structures: The structure of each synthetic dataset assumes a horizontal table header which potentially spans over multiple rows and a table body that may contain a combination of row spans and column spans. However, spans are not allowed to cross the header -body boundary. The table structure is described by the parameters: Total number of table rows and columns, number of header rows, type of spans (header only spans, row only spans, column only spans, both row and column spans), maximum span size and the ratio of the table area covered by spans.
<!-- image --> 3. Generate content: Based on the dataset theme, a set of suitable content templates is chosen first. Then, this content can be combined with purely random text to produce the synthetic content.
## EMPLOYMENT 4. Apply styling templates: Depending on the domain of the synthetic dataset, a set of styling templates is first manually selected. Then, a style is randomly selected to format the appearance of the synthesized table.
10:00 to 5:00. Call The Mudpuppy 731-2581 5. Render the complete tables: The synthetic table is finally rendered by a web browser engine to generate the bounding boxes for each table cell. A batching technique is utilized to optimize the runtime overhead of the rendering process.
» ## 2. Prediction post-processing for PDF documents
Fort Worth Ladies Specialty Boutique located in the Stockyard Station needs experenced sales person Weekends, non-smoker 626-8900 Although TableFormer can predict the table structure and the bounding boxes for tables recognized inside PDF documents, this is not enough when a full reconstruction of the original table is required. This happens mainly due the following reasons:
Great Money--Great Fun! Become a Flash photographer, must have 35mm camera, will train FLASH, 924-2626
Letters, term papers, the$@5, oF anything else you may need typed, call 214-641-1023. 1021
## a
TERM PAPERS TYPED FAST! Laser printed Rush orders and credit cards accepted. Across from TCU at 3023 South University Drive. ACCURACY +PLUS, 926-4969
plus personal cash bonuses for yourself Call 1-800-932-0528, Ext. 65. from 3 different fundraisers lasting either 3 or 7 days. No investment Earn $$$ for your group
## -
Southwest and rustic furfor information, 922-9563.
|
## CAMPUSLINES
ALCOHOL AND DRUG EDUCATION 1 pedehag en} SUPPORT GROUP at § Nednesdays in Ricke 3
PSI CHI psyct are
FRENCH CLUB w fay
## FELLOWSHIP OF owes te
te AT HLETES © we
BAPTIST STUDENT MINISTRY i tu
## Today
|
dent Union Thursd dible st meets at 12:30 on ays for Noondays (a udy) in Student Center 216 and at 7 p.m. on Mondays for Jume start, a worship time, in Student Center 20 5-206
## TCU CATHOLIC COMMU-
NITY p gran Supper pro fr every nthe Fa ulty Center 14. Al are rmanizator eat during Mid fr Break et. 13-1 format a
## RAPE/SEXUAL ASSAULT
SURVIVORS GROUP meeting
<!-- image -->
## by Bill Watterson
<!-- image -->
|
SIRCULATION 4
SUBSCRIPTIONS
PHOTOGRAPHS if t '
EDITORIAL POLICY edit t Oe t
<!-- image -->
LETTERS TO THE EDITOR e Skiff
1 ] ! ef blication to the Skiff the ' ficat naior
e ee wea coejte @ mai form and through the { . 4
f kiff nember of the sted Pre
MAILING ADDRESS Ff
<!-- image -->
S
CANTINA
* Full/Part-time positions
* Training provided
* No experience necessary
IN
PERSON
3:00
p.m.
Tuesday
-
5:00
&
p.m.
* 10433 Lompanoy Ln. (Now Open)
*
723
S.
Centrar
Expressway
(Richaroson)
Equa/l Opportunity Employer
## PAPPADEALLY SEAFOOD KITCHEN
*
positions provided
Full/Part-time
Training
* No experience necessary
## Apply in person:
Tuesday & Wednesday, 3:00 p.m. -5:00 p.m.
Dallas: 3520 Oaklawn at Lemmon Irving: 10428 Lombardy Ln. (open soon!) Richardson: 725 S. Central Expressway Arlington: 1304 Copeland Rd. at Collins Equal Opportunity Employer
Inquiring minds read the Skiff.
Wednesday
<!-- image -->
## TCU DAILY SKIFF Since 1902
SKIFF phenyl en OMRSCTORY nm be reactwd by Giahrig
After
H
sife
Busine
Manag
'¢
uM"airs:
Image TCU
at
Jt
Magazine
Vax Product
(address:
Manager
What's
news?
Find
out
TCU
in
Daily
Skiff
TRAFFIC
TICKETS
|iiigeaeme clsewher re in Tarrant County. ) promuses as to results Any fine and any court costs are not included fee for le zal repres entation
R.
James
Atiorney
24 Sandage Fort Worth, TX 76109-1793 2203236
M. ALLORY at La i
Stockholm $699" Los Angeles $250" Mexico City $198" See oc caret bor 0 Wort Ged on md wed bord toe Lyall wrortivante
<!-- image -->
6715 Hillcrest Dallas, TX 75205
214-363-9941
Eurailpasses
the
ios ][N © on TT ad or HE SHOULD | me
ee
WEY!
win Po iS
Appty
:
Sk
g-/

2
tests/data_scanned/scanned_01.tesserocr.pages.json
View File

File diff suppressed because one or more lines are too long

113
tests/data_scanned/scanned_02.easyocr.doctags.txt
View File

@ -1,96 +1,25 @@
<document> <document>
<paragraph><location><page_1><loc_7><loc_84><loc_48><loc_89></location>You Can Win a Cash Award _ and Scholarship Money for Your in College</paragraph>
<subtitle-level-1><location><page_1><loc_8><loc_73><loc_64><loc_84></location>Reader's Digest</subtitle-level-1>
<paragraph><location><page_1><loc_13><loc_70><loc_63><loc_73></location>Open to All College Students (Faculty, too! )</paragraph>
<paragraph><location><page_1><loc_7><loc_73><loc_65><loc_80></location>$41,000 CONTEST</paragraph>
<paragraph><location><page_1><loc_13><loc_67><loc_56><loc_69></location>Nothing to buy. nothing to write</paragraph>
<paragraph><location><page_1><loc_17><loc_63><loc_57><loc_67></location>and you may find you know more about than think! sou people</paragraph>
<paragraph><location><page_1><loc_7><loc_56><loc_44><loc_62></location>How well do you know human nature? Can you tell what subjects interest people most? Here is a chance to test your judgment ~show how an editor you areand you may win 85,000 for yourself, plus 85,000 in scholarship funds for college. good your</paragraph>
<paragraph><location><page_1><loc_7><loc_45><loc_44><loc_53></location>Why'do far more college graduates read Reader's Digest than any other magazine? What is it that makes the Digest the-most widely read magazine in the world with Il million bought each month in the United Stai plus 9 million abroad? is it read each month by at least 60 million people, in 12 languages Arabic, Danish, English, Finnish, French, German; Italian, Japanese, Norwegian; Portuguese, Spanish and Swedish? copies tes, Why</paragraph>
<paragraph><location><page_1><loc_7><loc_53><loc_44><loc_56></location>Its fun to try _ Maybe you can top other students in colleges across the country and you can match wits with thé editors of Reader's Digest.</paragraph>
<paragraph><location><page_1><loc_7><loc_40><loc_44><loc_45></location>Can you spot in a typical issue of Reader's Digest the universal human values that link scholars, statesmen, scientists, writers, businessmen, housewives? Can you pick out the articles that will be most popular with the average Digest reader?</paragraph>
<subtitle-level-1><location><page_1><loc_7><loc_38><loc_44><loc_40></location>You may find _ you knor more about people than you think!</subtitle-level-1>
<paragraph><location><page_1><loc_7><loc_33><loc_44><loc_38></location>Heres all you do. Study the descriptions (at right) of the'articles in the October Reader's Digest ~or, better still, read complete articles in the issue itself (But you are not required to buy The Reader's Digest to enter the contest ) Then simply list the six articles-in order of preference -that you think readers of the magazine will lilce best. This will be compared with a nationwidesurvey conducted among a cross gection of Digest subscribers the</paragraph>
<paragraph><location><page_1><loc_7><loc_30><loc_45><loc_33></location>Follow the directions given below. Fill in the entry blank, paste it on 0 card, and get it into the mail before the deadline. Additional blanks are obtainable at boolrstore. your post . college</paragraph>
<paragraph><location><page_1><loc_7><loc_28><loc_45><loc_30></location>All entries must be postmarked not later than midniglt, October 25, 1956. Don't delay. In case of ties, the entry with the carliegt poetmark will win.</paragraph>
<figure> <figure>
<location><page_1><loc_7><loc_21><loc_22><loc_27></location> <location><page_1><loc_16><loc_87><loc_82><loc_91></location>
</figure> </figure>
<paragraph><location><page_1><loc_23><loc_22><loc_45><loc_26></location>Just pick in order the six- articles you think most readers of October Reader's Digest will like the best.</paragraph> <subtitle-level-1><location><page_1><loc_11><loc_80><loc_20><loc_82></location>Purpose</subtitle-level-1>
<paragraph><location><page_1><loc_9><loc_13><loc_36><loc_21></location>READER' S DIGEST CONTEST, Box 4, Great Neck, L 1., Now York In the epace oppogite the word "FIRST" write tho number of tho urticlo you thínk will bo tho 'moBt popular of all. Oppogite the word "SECOND' write the numbor of the articlo you think will ronk gecond in populority. Liat in thig way tho numbore 0f tho 8ix top articleb in tho ordor of thoir populority. (Noto:Uøeonly the numbers of orticleb chooge you Do not write the title of any article ) Clip and puste this coupon on @ Government post card .</paragraph> <paragraph><location><page_1><loc_11><loc_69><loc_88><loc_80></location>The Dean's Professional Development Award for Staff is to allow CLASS staff the opportunity to attend conferences and workshops in their field for the sole purpose of professional development. The intent is to defray costs associated with attendance. The maximum amount of the award is $2,000 per staff member. Up to four awards will be made per year, contingent upon the availability of funding. Staff members that are awarded must wait three years from the date of award notification before reapplying again.</paragraph>
<paragraph><location><page_1><loc_37><loc_19><loc_39><loc_20></location>Flrst</paragraph> <subtitle-level-1><location><page_1><loc_12><loc_66><loc_21><loc_68></location>Eligibility</subtitle-level-1>
<paragraph><location><page_1><loc_37><loc_18><loc_40><loc_19></location>Second</paragraph> <paragraph><location><page_1><loc_12><loc_64><loc_51><loc_66></location>All staff currently employed in CLASS are eligible.</paragraph>
<paragraph><location><page_1><loc_37><loc_16><loc_39><loc_18></location>Thlyd Founh</paragraph> <subtitle-level-1><location><page_1><loc_12><loc_61><loc_37><loc_63></location>What the Award Will Fund</subtitle-level-1>
<paragraph><location><page_1><loc_37><loc_15><loc_39><loc_16></location>FIllh</paragraph> <paragraph><location><page_1><loc_12><loc_58><loc_56><loc_61></location>Costs associated with conference/workshop including:</paragraph>
<paragraph><location><page_1><loc_37><loc_14><loc_39><loc_15></location>Slxih</paragraph> <paragraph><location><page_1><loc_15><loc_57><loc_23><loc_58></location>Airfare</paragraph>
<paragraph><location><page_1><loc_9><loc_13><loc_12><loc_14></location>Name</paragraph> <paragraph><location><page_1><loc_15><loc_55><loc_24><loc_56></location>Lodging</paragraph>
<paragraph><location><page_1><loc_26><loc_13><loc_30><loc_13></location>ddreas</paragraph> <paragraph><location><page_1><loc_15><loc_53><loc_23><loc_54></location>Meals</paragraph>
<paragraph><location><page_1><loc_9><loc_11><loc_11><loc_12></location>City_</paragraph> <paragraph><location><page_1><loc_15><loc_51><loc_32><loc_53></location>Registration fees</paragraph>
<paragraph><location><page_1><loc_9><loc_10><loc_16><loc_11></location>Namc of college_</paragraph> <paragraph><location><page_1><loc_15><loc_49><loc_37><loc_51></location>Ground Transportation</paragraph>
<paragraph><location><page_1><loc_46><loc_10><loc_66><loc_11></location>Ite popularity and influcnce arc world-wlde</paragraph> <subtitle-level-1><location><page_1><loc_12><loc_46><loc_41><loc_48></location>What the Award Will Not Fund</subtitle-level-1>
<paragraph><location><page_1><loc_29><loc_11><loc_31><loc_12></location>State</paragraph> <paragraph><location><page_1><loc_12><loc_43><loc_78><loc_46></location>expenses incurred outside of the scope of the proposed development activity. Any</paragraph>
<subtitle-level-1><location><page_1><loc_48><loc_59><loc_64><loc_61></location>You CAN WIM :</subtitle-level-1> <subtitle-level-1><location><page_1><loc_11><loc_40><loc_29><loc_43></location>Granting Schedule</subtitle-level-1>
<paragraph><location><page_1><loc_48><loc_56><loc_63><loc_59></location>s5000 cash 1st prize 85000 plus for the scholarship fund of your college or_</paragraph> <paragraph><location><page_1><loc_12><loc_38><loc_41><loc_40></location>Earliest Submission Date: August 1st</paragraph>
<subtitle-level-1><location><page_1><loc_48><loc_54><loc_64><loc_56></location>2ud s1000 cash prize</subtitle-level-1> <paragraph><location><page_1><loc_11><loc_36><loc_36><loc_38></location>Applications Due: October 1s</paragraph>
<paragraph><location><page_1><loc_68><loc_54><loc_90><loc_56></location>20. What Your sense of humor lells about you. What the jokes you like, the way you laugh reveal about you.</paragraph> <paragraph><location><page_1><loc_11><loc_36><loc_36><loc_38></location>Applications Due: October 1s</paragraph>
<paragraph><location><page_1><loc_48><loc_53><loc_63><loc_54></location>plus 81000 for the scholarship fund of your college or</paragraph> <paragraph><location><page_1><loc_12><loc_34><loc_42><loc_36></location>Notification of Awards: November 1st</paragraph>
<paragraph><location><page_1><loc_68><loc_53><loc_90><loc_54></location>21 The sub ihat wouldn' stay down Stirring saga of the US.Sv Squalus' rescue from depth of 40 fathoms</paragraph> <paragraph><location><page_1><loc_11><loc_28><loc_85><loc_32></location>Please submit applications to CLASSGrt@uh edu by the deadline. Please write "Professional Development-Staff" in the subject line.</paragraph>
<paragraph><location><page_1><loc_48><loc_50><loc_64><loc_52></location>TEN Any of $500 cash prizes plus 8500 for the scholarship fund of your college or</paragraph> <paragraph><location><page_1><loc_12><loc_19><loc_86><loc_27></location>PLEASE NOTE: Please include a supporting letter from your Department Chair or Immediate Supervisor. Incomplete applications not be reviewed. Applications will be considered incomplete until all information has been received, at which time an email confirming receipt will be sent to you. will</paragraph>
<subtitle-level-1><location><page_1><loc_50><loc_48><loc_62><loc_49></location>Any %f 100 $10 prizes</subtitle-level-1>
<paragraph><location><page_1><loc_47><loc_43><loc_65><loc_46></location>Andifyour entry is the best from your college you will receive an extra award aI additional $10 in book credit at your college bookstore.</paragraph>
<paragraph><location><page_1><loc_50><loc_46><loc_62><loc_48></location>in book credit from your local bookstore college</paragraph>
<subtitle-level-1><location><page_1><loc_49><loc_42><loc_63><loc_43></location>FOLLOW THESE EASY RULES</subtitle-level-1>
<paragraph><location><page_1><loc_47><loc_38><loc_65><loc_42></location>1. Read the descriptions in this adver tisement of the articleg that appear in October Reader'ø Digeet_ Or, better, read the complete articleg Then select the 6 that you tbink most readerg will like best.</paragraph>
<paragraph><location><page_1><loc_47><loc_25><loc_65><loc_30></location>3. This conlest is open only to studentg and faculty memberg in thc U, S. excluding employces of The Reader'ø Digeat, itg advertising agenciep and their familieg, It iø gubject to 'all federal, Btate and local lawg and rogulations. college</paragraph>
<paragraph><location><page_1><loc_47><loc_30><loc_65><loc_38></location>2On the entry blank at left, write the nunber of each article you select. Ligt them in what you think will be the order of popularity, from first to sixth place. Your gclectiong will be judged by comparigon with & national gurvey which ranks in order of popdlarity the 6 articleg that reàderg like begt. Fillin and mail the coupon. All entrieg mugt be poatmarked not later than midnight, October 25, 1956.</paragraph>
<paragraph><location><page_1><loc_48><loc_24><loc_61><loc_25></location>4. Only ono entry por person</paragraph>
<paragraph><location><page_1><loc_47><loc_20><loc_65><loc_24></location>5. In case of tles entricg poetmarked eaxlieet will win. Entrieg will bo judged by 0 E McIntyro;_ Inc., whogo decigion will be final. All ontrieg become property of The Reader's Digest; none returned.</paragraph>
<paragraph><location><page_1><loc_47><loc_17><loc_65><loc_20></location>6 All winners notified by' mail. List of if cagh-prizo winnorg mailed you enclogoagelf-addresscd, etampedenvolope</paragraph>
<figure>
<location><page_1><loc_48><loc_10><loc_64><loc_16></location>
</figure>
<paragraph><location><page_1><loc_68><loc_49><loc_90><loc_51></location>23 Doctors should lell palienIs the Iruth. When the doctor operated, exactly what did he do? Why written record of your medical bistory may someday save your life</paragraph>
<paragraph><location><page_1><loc_68><loc_51><loc_90><loc_53></location>22. Madame BuHerflyin bobby sox. How new freedoms have changed life for Japanese women; what the men think.</paragraph>
<paragraph><location><page_1><loc_68><loc_47><loc_90><loc_49></location>24 wonderful you are Here'8 why affection and admiration aren't much good unless expressed; why locked-up emotions eventually wither . "How</paragraph>
<paragraph><location><page_1><loc_68><loc_45><loc_90><loc_47></location>25. Harry_ Holt and hoarlful of children. Story of farmer who singlehandedly finds homes for bundredg of Korean war orphang.</paragraph>
<paragraph><location><page_1><loc_68><loc_43><loc_90><loc_45></location>26 Our lax laws mako usdishonest. How unlair tax lawg are cauging Berious moral deterioration_</paragraph>
<paragraph><location><page_1><loc_68><loc_42><loc_91><loc_43></location>27 _ Voneroal dlsoase now threat Io .Youth. How V.D. i8 epreading among teen-agers and gane advice to victims_</paragraph>
<paragraph><location><page_1><loc_68><loc_40><loc_91><loc_42></location>28, Socy. Benson's fallh In He Amorlcan farmer. Why he feels farmers_ Jeft alone can often solve their own problemg better than Waghington.</paragraph>
<paragraph><location><page_1><loc_68><loc_38><loc_91><loc_40></location>29, Your braln'$ vnreolizod powers, Seven new findings to help you use your brain more cfficiently_</paragraph>
<paragraph><location><page_1><loc_68><loc_37><loc_91><loc_38></location>J0 Brltaln'sindoslrucllble Man:' What Sir Winston Churchill is doing in retirement. "Old</paragraph>
<paragraph><location><page_1><loc_68><loc_35><loc_91><loc_37></location>31 Are lurles Blvíng away loo much money? Fantastic awarde juries hand because they confuse compassion with common sønse_ out</paragraph>
<paragraph><location><page_1><loc_68><loc_32><loc_91><loc_35></location>32. My last bost days on oarlh. Inher own words young mothcr, learning she had cancer tells how she decicled to malro thiø the 'best ycar of her lífe.</paragraph>
<paragraph><location><page_1><loc_68><loc_31><loc_91><loc_32></location>33. Foroløn-uid manla. How tho billiong we'vo given havo brought mainly disuppointment and highor taxes.</paragraph>
<paragraph><location><page_1><loc_68><loc_29><loc_91><loc_31></location>34. Out whoro lol planes oro born. Story of Edwnrd Air Forco Buge, whcro 10,000 men battle wind, Band and spced barricrø to kcep u8 gupremo in tho sly -</paragraph>
<paragraph><location><page_1><loc_68><loc_27><loc_91><loc_29></location>35 Llfo In thoso Unilod Slalos. Iumoroug anccdotos revcal ing quirks of humun nelure.</paragraph>
<paragraph><location><page_1><loc_68><loc_26><loc_91><loc_27></location>36, Man's most playful frlond: tho Land Oller. Intercsting facts about this amuBlng 4nlmul</paragraph>
<paragraph><location><page_1><loc_68><loc_24><loc_91><loc_26></location>37 Why nol foroløn - sorvlco carøor? Iow our Stute Doportmont ismaking foreign gorvicoattractivotoyoung mn</paragraph>
<paragraph><location><page_1><loc_68><loc_23><loc_91><loc_24></location>30.A now doal In Iho old flrohouso Iow ono-town got lower tuxes, grealer protection combininn fro und polico,</paragraph>
<paragraph><location><page_1><loc_68><loc_21><loc_91><loc_23></location>39 Crazy man on Crazy Horso, tho nJnn w/ogo Btatuo ol an Indínn will bo thc largeøt in higlovy. Mcet</paragraph>
<paragraph><location><page_1><loc_68><loc_20><loc_91><loc_21></location>40. Tholr businoss Ie dynamlto. Iow tho munulauturo at thig explosive has bcen mado ono o[ the snfept incustrics;</paragraph>
<paragraph><location><page_1><loc_68><loc_17><loc_91><loc_20></location>41. Hls bost cuslomors aro bablos. How kilehan strnincr and0 pint Df mnshod peug bccamo tho Gerber Producl# Co. 42, Smokry Mounlaln maglc. Why this, our mnoat. nncient mountaln rango Jng moro vlgitorg thun nny othor ,</paragraph>
<paragraph><location><page_1><loc_68><loc_15><loc_91><loc_17></location>43, Call for Mr: Emorgoncy Mcot the JEmorgoncy who got 8 million Now Yorkerø put ol trouhlo Polico,</paragraph>
<paragraph><location><page_1><loc_68><loc_14><loc_91><loc_15></location>44' Boauly by !ho mllo: Iow Jandacopo onglnecrg provo rondeldo planting is lileunving "8 Jveli 18 bonullful</paragraph>
<paragraph><location><page_1><loc_68><loc_13><loc_91><loc_14></location>45_ Humor In unlform, Btorlc8 ol tho [unny Bldo of Ilfo In our Armed Forcog, 'Truo</paragraph>
<paragraph><location><page_1><loc_68><loc_9><loc_91><loc_12></location>46_ Sovon oconomlc Fallaclos. Thc Amorican Economlc Doundutlon oxplodog mlnconcopliong "bout our cconomy. 47, Admlral ol Ihio Grook Oll Floot. Story of Stavros NInrchoe, who hnb won fortuno bolling 0n-und cnrrying~o1l.</paragraph>
<subtitle-level-1><location><page_1><loc_67><loc_85><loc_90><loc_89></location>Which six articles will readers of the October Digest like best ?</subtitle-level-1>
<paragraph><location><page_1><loc_68><loc_84><loc_90><loc_85></location>Norfolk's frlend (0 Iroubled teen-agers_ Story of the arthritic cripple to whom youngsters fock for advice.</paragraph>
<paragraph><location><page_1><loc_68><loc_82><loc_90><loc_84></location>2. The great Piltdown hoax. How this famed "missing link" in human evolution has been proved fraud from the start.</paragraph>
<paragraph><location><page_1><loc_68><loc_81><loc_90><loc_82></location>3. How to sharpen Your judgmert. Famed author Bertrand Russell offerg six rules to help you form sounder opinions.</paragraph>
<paragraph><location><page_1><loc_68><loc_79><loc_90><loc_81></location>4 My most unforgellable 'character. Fond memories of Connie Mack who led the 'Athletics for 50 years.</paragraph>
<paragraph><location><page_1><loc_68><loc_78><loc_90><loc_79></location>5 How to make peace af Ihe Pentogon. to end ruinOUS rivalry between our Army, Navy and Air Force. Stepg</paragraph>
<paragraph><location><page_1><loc_67><loc_76><loc_90><loc_78></location>6. Book condensalion: "High Wide Lonesome Hal Zorßaados pxaitig story of his adventurous; boyhood on and</paragraph>
<paragraph><location><page_1><loc_68><loc_74><loc_90><loc_76></location>7. Medicine's animal pionoers. How medical researchera learn from animals new ways to save human lives -</paragraph>
<paragraph><location><page_1><loc_68><loc_73><loc_90><loc_74></location>8- What Ihe mess' in Moscow means. Evidence that the Communist system is as unworkable as it is unnatural.</paragraph>
<paragraph><location><page_1><loc_68><loc_71><loc_90><loc_73></location>9 Master bridge builder. Introducing David Steinman, world leader in bridge design and construction.</paragraph>
<paragraph><location><page_1><loc_68><loc_70><loc_90><loc_71></location>10 College Iwo years sooner. Here's how extensive experi ments proved bright lOth-grader is ready for college.</paragraph>
<paragraph><location><page_1><loc_68><loc_68><loc_90><loc_70></location>11. Laughter the best medicine. Amusing experiences from everyday life.</paragraph>
<paragraph><location><page_1><loc_68><loc_66><loc_90><loc_68></location>12: What happens when we pray for others? Too often we pray only for ourselves: Here's how we true rewards of prayer when we pray for others_ gain</paragraph>
<paragraph><location><page_1><loc_68><loc_65><loc_90><loc_66></location>13. European vs. U.5. beaulies. Why European women are more glamorous to men,</paragraph>
<paragraph><location><page_1><loc_68><loc_63><loc_90><loc_65></location>!4. Trading stamps~bonus or bunkum? How much of their cost is included in the price you pay?</paragraph>
<paragraph><location><page_1><loc_68><loc_62><loc_90><loc_63></location>15. Living memorials inslead of flowers. A way tohonor the dead by serving the living</paragraph>
<paragraph><location><page_1><loc_68><loc_60><loc_90><loc_62></location>16 It pays to increase Your word power An entertaining quiz to build your vocabulary.</paragraph>
<paragraph><location><page_1><loc_68><loc_59><loc_90><loc_60></location>17_ Are we too soft on young criminals? Why the best way t0 cure juvenile' delinquency i8 to punish first offenders.</paragraph>
<paragraph><location><page_1><loc_68><loc_57><loc_90><loc_59></location>18. Medicine man on the Amazon How two devoted missionaries bring medical aid to jungle natives.</paragraph>
<paragraph><location><page_1><loc_68><loc_56><loc_90><loc_57></location>19_ Creatures in the night: The fascinating drama of nature that is enacted between dusk and dawn.</paragraph>
</document> </document>

2
tests/data_scanned/scanned_02.easyocr.json
View File

File diff suppressed because one or more lines are too long

179
tests/data_scanned/scanned_02.easyocr.md
View File

@ -1,181 +1,40 @@
You Can Win a Cash Award _ and Scholarship Money for Your in College
## Reader's Digest
Open to All College Students (Faculty, too! )
$41,000 CONTEST
Nothing to buy. nothing to write
and you may find you know more about than think! sou people
How well do you know human nature? Can you tell what subjects interest people most? Here is a chance to test your judgment ~show how an editor you areand you may win 85,000 for yourself, plus 85,000 in scholarship funds for college. good your
Why'do far more college graduates read Reader's Digest than any other magazine? What is it that makes the Digest the-most widely read magazine in the world with Il million bought each month in the United Stai plus 9 million abroad? is it read each month by at least 60 million people, in 12 languages Arabic, Danish, English, Finnish, French, German; Italian, Japanese, Norwegian; Portuguese, Spanish and Swedish? copies tes, Why
Its fun to try _ Maybe you can top other students in colleges across the country and you can match wits with thé editors of Reader's Digest.
Can you spot in a typical issue of Reader's Digest the universal human values that link scholars, statesmen, scientists, writers, businessmen, housewives? Can you pick out the articles that will be most popular with the average Digest reader?
## You may find _ you knor more about people than you think!
Heres all you do. Study the descriptions (at right) of the'articles in the October Reader's Digest ~or, better still, read complete articles in the issue itself (But you are not required to buy The Reader's Digest to enter the contest ) Then simply list the six articles-in order of preference -that you think readers of the magazine will lilce best. This will be compared with a nationwidesurvey conducted among a cross gection of Digest subscribers the
Follow the directions given below. Fill in the entry blank, paste it on 0 card, and get it into the mail before the deadline. Additional blanks are obtainable at boolrstore. your post . college
All entries must be postmarked not later than midniglt, October 25, 1956. Don't delay. In case of ties, the entry with the carliegt poetmark will win.
<!-- image --> <!-- image -->
Just pick in order the six- articles you think most readers of October Reader's Digest will like the best. ## Purpose
READER' S DIGEST CONTEST, Box 4, Great Neck, L 1., Now York In the epace oppogite the word "FIRST" write tho number of tho urticlo you thínk will bo tho 'moBt popular of all. Oppogite the word "SECOND' write the numbor of the articlo you think will ronk gecond in populority. Liat in thig way tho numbore 0f tho 8ix top articleb in tho ordor of thoir populority. (Noto:Uøeonly the numbers of orticleb chooge you Do not write the title of any article ) Clip and puste this coupon on @ Government post card . The Dean's Professional Development Award for Staff is to allow CLASS staff the opportunity to attend conferences and workshops in their field for the sole purpose of professional development. The intent is to defray costs associated with attendance. The maximum amount of the award is $2,000 per staff member. Up to four awards will be made per year, contingent upon the availability of funding. Staff members that are awarded must wait three years from the date of award notification before reapplying again.
Flrst ## Eligibility
Second All staff currently employed in CLASS are eligible.
Thlyd Founh ## What the Award Will Fund
FIllh Costs associated with conference/workshop including:
Slxih Airfare
Name Lodging
ddreas Meals
City_ Registration fees
Namc of college_ Ground Transportation
Ite popularity and influcnce arc world-wlde ## What the Award Will Not Fund
State expenses incurred outside of the scope of the proposed development activity. Any
## You CAN WIM : ## Granting Schedule
s5000 cash 1st prize 85000 plus for the scholarship fund of your college or_ Earliest Submission Date: August 1st
## 2ud s1000 cash prize Applications Due: October 1s
20. What Your sense of humor lells about you. What the jokes you like, the way you laugh reveal about you. Notification of Awards: November 1st
plus 81000 for the scholarship fund of your college or Please submit applications to CLASSGrt@uh edu by the deadline. Please write "Professional Development-Staff" in the subject line.
21 The sub ihat wouldn' stay down Stirring saga of the US.Sv Squalus' rescue from depth of 40 fathoms PLEASE NOTE: Please include a supporting letter from your Department Chair or Immediate Supervisor. Incomplete applications not be reviewed. Applications will be considered incomplete until all information has been received, at which time an email confirming receipt will be sent to you. will
TEN Any of $500 cash prizes plus 8500 for the scholarship fund of your college or
## Any %f 100 $10 prizes
Andifyour entry is the best from your college you will receive an extra award aI additional $10 in book credit at your college bookstore.
in book credit from your local bookstore college
## FOLLOW THESE EASY RULES
1. Read the descriptions in this adver tisement of the articleg that appear in October Reader'ø Digeet_ Or, better, read the complete articleg Then select the 6 that you tbink most readerg will like best.
3. This conlest is open only to studentg and faculty memberg in thc U, S. excluding employces of The Reader'ø Digeat, itg advertising agenciep and their familieg, It iø gubject to 'all federal, Btate and local lawg and rogulations. college
2On the entry blank at left, write the nunber of each article you select. Ligt them in what you think will be the order of popularity, from first to sixth place. Your gclectiong will be judged by comparigon with & national gurvey which ranks in order of popdlarity the 6 articleg that reàderg like begt. Fillin and mail the coupon. All entrieg mugt be poatmarked not later than midnight, October 25, 1956.
4. Only ono entry por person
5. In case of tles entricg poetmarked eaxlieet will win. Entrieg will bo judged by 0 E McIntyro;_ Inc., whogo decigion will be final. All ontrieg become property of The Reader's Digest; none returned.
6 All winners notified by' mail. List of if cagh-prizo winnorg mailed you enclogoagelf-addresscd, etampedenvolope
<!-- image -->
23 Doctors should lell palienIs the Iruth. When the doctor operated, exactly what did he do? Why written record of your medical bistory may someday save your life
22. Madame BuHerflyin bobby sox. How new freedoms have changed life for Japanese women; what the men think.
24 wonderful you are Here'8 why affection and admiration aren't much good unless expressed; why locked-up emotions eventually wither . "How
25. Harry_ Holt and hoarlful of children. Story of farmer who singlehandedly finds homes for bundredg of Korean war orphang.
26 Our lax laws mako usdishonest. How unlair tax lawg are cauging Berious moral deterioration_
27 _ Voneroal dlsoase now threat Io .Youth. How V.D. i8 epreading among teen-agers and gane advice to victims_
28, Socy. Benson's fallh In He Amorlcan farmer. Why he feels farmers_ Jeft alone can often solve their own problemg better than Waghington.
29, Your braln'$ vnreolizod powers, Seven new findings to help you use your brain more cfficiently_
J0 Brltaln'sindoslrucllble Man:' What Sir Winston Churchill is doing in retirement. "Old
31 Are lurles Blvíng away loo much money? Fantastic awarde juries hand because they confuse compassion with common sønse_ out
32. My last bost days on oarlh. Inher own words young mothcr, learning she had cancer tells how she decicled to malro thiø the 'best ycar of her lífe.
33. Foroløn-uid manla. How tho billiong we'vo given havo brought mainly disuppointment and highor taxes.
34. Out whoro lol planes oro born. Story of Edwnrd Air Forco Buge, whcro 10,000 men battle wind, Band and spced barricrø to kcep u8 gupremo in tho sly -
35 Llfo In thoso Unilod Slalos. Iumoroug anccdotos revcal ing quirks of humun nelure.
36, Man's most playful frlond: tho Land Oller. Intercsting facts about this amuBlng 4nlmul
37 Why nol foroløn - sorvlco carøor? Iow our Stute Doportmont ismaking foreign gorvicoattractivotoyoung mn
30.A now doal In Iho old flrohouso Iow ono-town got lower tuxes, grealer protection combininn fro und polico,
39 Crazy man on Crazy Horso, tho nJnn w/ogo Btatuo ol an Indínn will bo thc largeøt in higlovy. Mcet
40. Tholr businoss Ie dynamlto. Iow tho munulauturo at thig explosive has bcen mado ono o[ the snfept incustrics;
41. Hls bost cuslomors aro bablos. How kilehan strnincr and0 pint Df mnshod peug bccamo tho Gerber Producl# Co. 42, Smokry Mounlaln maglc. Why this, our mnoat. nncient mountaln rango Jng moro vlgitorg thun nny othor ,
43, Call for Mr: Emorgoncy Mcot the JEmorgoncy who got 8 million Now Yorkerø put ol trouhlo Polico,
44' Boauly by !ho mllo: Iow Jandacopo onglnecrg provo rondeldo planting is lileunving "8 Jveli 18 bonullful
45_ Humor In unlform, Btorlc8 ol tho [unny Bldo of Ilfo In our Armed Forcog, 'Truo
46_ Sovon oconomlc Fallaclos. Thc Amorican Economlc Doundutlon oxplodog mlnconcopliong "bout our cconomy. 47, Admlral ol Ihio Grook Oll Floot. Story of Stavros NInrchoe, who hnb won fortuno bolling 0n-und cnrrying~o1l.
## Which six articles will readers of the October Digest like best ?
Norfolk's frlend (0 Iroubled teen-agers_ Story of the arthritic cripple to whom youngsters fock for advice.
2. The great Piltdown hoax. How this famed "missing link" in human evolution has been proved fraud from the start.
3. How to sharpen Your judgmert. Famed author Bertrand Russell offerg six rules to help you form sounder opinions.
4 My most unforgellable 'character. Fond memories of Connie Mack who led the 'Athletics for 50 years.
5 How to make peace af Ihe Pentogon. to end ruinOUS rivalry between our Army, Navy and Air Force. Stepg
6. Book condensalion: "High Wide Lonesome Hal Zorßaados pxaitig story of his adventurous; boyhood on and
7. Medicine's animal pionoers. How medical researchera learn from animals new ways to save human lives -
8- What Ihe mess' in Moscow means. Evidence that the Communist system is as unworkable as it is unnatural.
9 Master bridge builder. Introducing David Steinman, world leader in bridge design and construction.
10 College Iwo years sooner. Here's how extensive experi ments proved bright lOth-grader is ready for college.
11. Laughter the best medicine. Amusing experiences from everyday life.
12: What happens when we pray for others? Too often we pray only for ourselves: Here's how we true rewards of prayer when we pray for others_ gain
13. European vs. U.5. beaulies. Why European women are more glamorous to men,
!4. Trading stamps~bonus or bunkum? How much of their cost is included in the price you pay?
15. Living memorials inslead of flowers. A way tohonor the dead by serving the living
16 It pays to increase Your word power An entertaining quiz to build your vocabulary.
17_ Are we too soft on young criminals? Why the best way t0 cure juvenile' delinquency i8 to punish first offenders.
18. Medicine man on the Amazon How two devoted missionaries bring medical aid to jungle natives.
19_ Creatures in the night: The fascinating drama of nature that is enacted between dusk and dawn.

2
tests/data_scanned/scanned_02.easyocr.pages.json
View File

File diff suppressed because one or more lines are too long

BIN
tests/data_scanned/scanned_02.pdf
View File

Binary file not shown.

BIN
tests/data_scanned/scanned_02.png
View File

Binary file not shown.
Side by Side Swipe Overlay

Before

Width:  |  Height:  |  Size: 710 KiB

After

Width:  |  Height:  |  Size: 216 KiB

25
tests/data_scanned/scanned_02.tesseract.doctags.txt Normal file
View File

@ -0,0 +1,25 @@
<document>
<figure>
<location><page_1><loc_16><loc_87><loc_82><loc_91></location>
</figure>
<subtitle-level-1><location><page_1><loc_11><loc_80><loc_20><loc_82></location>Purpose</subtitle-level-1>
<paragraph><location><page_1><loc_11><loc_69><loc_88><loc_80></location>The Dean's Professional Development Award for Staff is to allow CLASS staff the opportunity to attend conferences and workshops in their field for the sole purpose of professional development. The intent is to defray costs associated with attendance. The maximum amount of the award is $2,000 per staff member. Up to four awards will be made per year, contingent upon the availability of funding. Staff members that are awarded must wait three years from the date of award notification before reapplying again.</paragraph>
<subtitle-level-1><location><page_1><loc_12><loc_66><loc_21><loc_68></location>Eligibility</subtitle-level-1>
<paragraph><location><page_1><loc_12><loc_64><loc_51><loc_66></location>All staff currently employed in CLASS are eligible.</paragraph>
<subtitle-level-1><location><page_1><loc_12><loc_61><loc_37><loc_63></location>What the Award Will Fund</subtitle-level-1>
<paragraph><location><page_1><loc_12><loc_58><loc_56><loc_61></location>Costs associated with conference/workshop including:</paragraph>
<paragraph><location><page_1><loc_15><loc_57><loc_23><loc_58></location>Airfare</paragraph>
<paragraph><location><page_1><loc_15><loc_55><loc_24><loc_56></location>Lodging</paragraph>
<paragraph><location><page_1><loc_15><loc_53><loc_23><loc_54></location>Meals</paragraph>
<paragraph><location><page_1><loc_15><loc_51><loc_32><loc_53></location>Registration fees</paragraph>
<paragraph><location><page_1><loc_15><loc_49><loc_37><loc_51></location>Ground Transportation</paragraph>
<subtitle-level-1><location><page_1><loc_12><loc_46><loc_41><loc_48></location>What the Award Will Not Fund</subtitle-level-1>
<paragraph><location><page_1><loc_12><loc_43><loc_78><loc_46></location>expenses incurred outside of the scope of the proposed development activity. Any</paragraph>
<subtitle-level-1><location><page_1><loc_11><loc_40><loc_29><loc_43></location>Granting Schedule</subtitle-level-1>
<paragraph><location><page_1><loc_12><loc_38><loc_41><loc_40></location>Earliest Submission Date: August 1st</paragraph>
<paragraph><location><page_1><loc_11><loc_36><loc_36><loc_38></location>Applications Due: October 1s</paragraph>
<paragraph><location><page_1><loc_11><loc_36><loc_36><loc_38></location>Applications Due: October 1s</paragraph>
<paragraph><location><page_1><loc_12><loc_34><loc_42><loc_36></location>Notification of Awards: November 1st</paragraph>
<paragraph><location><page_1><loc_11><loc_28><loc_85><loc_32></location>Please submit applications to CLASSGrt@uh edu by the deadline. Please write "Professional Development-Staff" in the subject line.</paragraph>
<paragraph><location><page_1><loc_12><loc_19><loc_86><loc_27></location>PLEASE NOTE: Please include a supporting letter from your Department Chair or Immediate Supervisor. Incomplete applications not be reviewed. Applications will be considered incomplete until all information has been received, at which time an email confirming receipt will be sent to you. will</paragraph>
</document>

1
tests/data_scanned/scanned_02.tesseract.json Normal file
View File

File diff suppressed because one or more lines are too long

40
tests/data_scanned/scanned_02.tesseract.md Normal file
View File

@ -0,0 +1,40 @@
<!-- image -->
## Purpose
The Dean's Professional Development Award for Staff is to allow CLASS staff the opportunity to attend conferences and workshops in their field for the sole purpose of professional development. The intent is to defray costs associated with attendance. The maximum amount of the award is $2,000 per staff member. Up to four awards will be made per year, contingent upon the availability of funding. Staff members that are awarded must wait three years from the date of award notification before reapplying again.
## Eligibility
All staff currently employed in CLASS are eligible.
## What the Award Will Fund
Costs associated with conference/workshop including:
Airfare
Lodging
Meals
Registration fees
Ground Transportation
## What the Award Will Not Fund
expenses incurred outside of the scope of the proposed development activity. Any
## Granting Schedule
Earliest Submission Date: August 1st
Applications Due: October 1s
Notification of Awards: November 1st
Please submit applications to CLASSGrt@uh edu by the deadline. Please write "Professional Development-Staff" in the subject line.
PLEASE NOTE: Please include a supporting letter from your Department Chair or Immediate Supervisor. Incomplete applications not be reviewed. Applications will be considered incomplete until all information has been received, at which time an email confirming receipt will be sent to you. will

1
tests/data_scanned/scanned_02.tesseract.pages.json Normal file
View File

File diff suppressed because one or more lines are too long

164
tests/data_scanned/scanned_02.tesserocr.doctags.txt
View File

@ -1,145 +1,27 @@
<document> <document>
<paragraph><location><page_1><loc_7><loc_84><loc_48><loc_89></location>You Can Win a Cash Award and Scholarship Money for Your College in</paragraph>
<paragraph><location><page_1><loc_13><loc_70><loc_62><loc_73></location>Qpen to All College Students (Faauty, to!</paragraph>
<paragraph><location><page_1><loc_17><loc_63><loc_56><loc_67></location>...and you may find you know more about people than you think!</paragraph>
<paragraph><location><page_1><loc_13><loc_66><loc_55><loc_70></location>Nothing to buy...nothing to write</paragraph>
<paragraph><location><page_1><loc_7><loc_56><loc_45><loc_62></location>'How well do you know human nature? Can you tell what subjects interest people most? Here is a chance to test your judgment-show how good an editor you are-and you may win $5,000 for yourself, plus $5, 000 in scholarship funds for your college.</paragraph>
<paragraph><location><page_1><loc_7><loc_55><loc_7><loc_55></location>|</paragraph>
<paragraph><location><page_1><loc_45><loc_60><loc_45><loc_60></location>_</paragraph>
<subtitle-level-1><location><page_1><loc_48><loc_59><loc_64><loc_61></location>You CAN WIN:</subtitle-level-1>
<paragraph><location><page_1><loc_48><loc_58><loc_53><loc_59></location>$5000</paragraph>
<paragraph><location><page_1><loc_53><loc_58><loc_57><loc_59></location>cash</paragraph>
<paragraph><location><page_1><loc_57><loc_58><loc_59><loc_59></location>i**</paragraph>
<paragraph><location><page_1><loc_59><loc_58><loc_63><loc_59></location>prize</paragraph>
<paragraph><location><page_1><loc_48><loc_56><loc_63><loc_58></location>plus $5000 for the scholarship fund of your college or...</paragraph>
<subtitle-level-1><location><page_1><loc_48><loc_54><loc_63><loc_56></location>$1000 cash 2" prize</subtitle-level-1>
<paragraph><location><page_1><loc_48><loc_50><loc_64><loc_52></location>Any of TEN $500 cash. prizes plus $500 for the scholarship :.. fund of your college or...</paragraph>
<paragraph><location><page_1><loc_48><loc_52><loc_63><loc_54></location>plus $1000 for the scholarship fund of your college or...</paragraph>
<subtitle-level-1><location><page_1><loc_50><loc_48><loc_62><loc_49></location>Any of 100 $10 prizes</subtitle-level-1>
<paragraph><location><page_1><loc_50><loc_48><loc_51><loc_48></location>in</paragraph>
<paragraph><location><page_1><loc_51><loc_48><loc_53><loc_48></location>book</paragraph>
<paragraph><location><page_1><loc_54><loc_48><loc_57><loc_48></location>credit</paragraph>
<paragraph><location><page_1><loc_57><loc_48><loc_59><loc_48></location>from</paragraph>
<paragraph><location><page_1><loc_50><loc_47><loc_53><loc_47></location>Jocal</paragraph>
<paragraph><location><page_1><loc_53><loc_47><loc_56><loc_47></location>college</paragraph>
<paragraph><location><page_1><loc_57><loc_47><loc_62><loc_48></location>your bookstore</paragraph>
<paragraph><location><page_1><loc_62><loc_47><loc_63><loc_47></location>_</paragraph>
<paragraph><location><page_1><loc_47><loc_43><loc_65><loc_46></location>And if your entry is the best from your college you will receive an extra award -an additional $10 in book credit. at your college bookstore,</paragraph>
<subtitle-level-1><location><page_1><loc_48><loc_42><loc_63><loc_43></location>FOLLOW THESE EASY-RULES</subtitle-level-1>
<paragraph><location><page_1><loc_47><loc_38><loc_65><loc_42></location>1 . Read the descriptions in this ad vertisement of the articles that appear in» October Reader's Digest. Or, better, read the complete articles, Then select the 6 that you think most readers will like best.</paragraph>
<paragraph><location><page_1><loc_47><loc_30><loc_65><loc_38></location>2.On the entry blank at left, write the number of each article you select. List them in what you think. will be the: order of popularity, from first to sixth place. Your selections will be judged by comparison with a national survey which ranks in order of popularity the 6 articles that readers like best. Fill in -and mail the coupon. All entries must be postmarked not later than midnight, October 25, 1956:</paragraph>
<paragraph><location><page_1><loc_48><loc_25><loc_65><loc_30></location>3. This contest is open only to college students and faculty members in the U. S., excluding employees of The Reader's Digest, its advertising agencies, and their families, It is subject to 'all federal, state and local Jaws and regulations.</paragraph>
<paragraph><location><page_1><loc_48><loc_24><loc_61><loc_25></location>4, Only one entry per person.</paragraph>
<paragraph><location><page_1><loc_47><loc_20><loc_65><loc_24></location>5. In case of ties, entries postmarked earliest will win, Entries will be judged by, O. E. McIntyre, Inc., whose decision will be final. All entries become property of The Reader's Digest; none returned...</paragraph>
<paragraph><location><page_1><loc_48><loc_17><loc_65><loc_20></location>6. All winners notified by'mail. List of cash-prize winners mailed if you pe ,Stamped enveope,</paragraph>
<figure> <figure>
<location><page_1><loc_48><loc_10><loc_64><loc_15></location> <location><page_1><loc_17><loc_87><loc_82><loc_91></location>
</figure> </figure>
<paragraph><location><page_1><loc_46><loc_13><loc_47><loc_13></location>'.</paragraph> <subtitle-level-1><location><page_1><loc_12><loc_80><loc_20><loc_82></location>Purpose</subtitle-level-1>
<paragraph><location><page_1><loc_46><loc_9><loc_66><loc_11></location>Its Poptart and influence are world-wide</paragraph> <paragraph><location><page_1><loc_12><loc_69><loc_88><loc_80></location>The Dean's Professional Development Award for Staff is to allow CLASS staff the opportunity to attend conferences and workshops in their field for the sole purpose of professional development. The intent is to defray costs associated with attendance. The maximum amount of the award is $2,000 per staff member. Up to four awards will be made per year, contingent upon the availability of funding. Staff members that are awarded must wait three years from the date of award notification before reapplying again.</paragraph>
<paragraph><location><page_1><loc_62><loc_9><loc_62><loc_9></location>I</paragraph> <subtitle-level-1><location><page_1><loc_12><loc_66><loc_20><loc_68></location>Eligibility</subtitle-level-1>
<paragraph><location><page_1><loc_7><loc_53><loc_44><loc_56></location>It's fun to try. Maybe you can top other students in colleges across the country ... and you can match wits with the editors of Reader's Digest.</paragraph> <paragraph><location><page_1><loc_12><loc_64><loc_51><loc_65></location>All staff currently employed in CLASS are eligible.</paragraph>
<paragraph><location><page_1><loc_44><loc_49><loc_45><loc_50></location>-</paragraph> <subtitle-level-1><location><page_1><loc_12><loc_61><loc_37><loc_62></location>What the Award Will Fund</subtitle-level-1>
<paragraph><location><page_1><loc_7><loc_45><loc_45><loc_52></location>_ Why'do far more college graduates read Reader's Digest any other magazine? What is it that makes the Digest the-most widely read magazine in the world - with 11 million copies bought each month in the United States, plus 9 million abroad? Why is it read each month by at least 60 million people, in 12 languages-Arabic, Danish, English, Finnish, French, German, Italian, Japanese, Norwegian, Portuguese, Spanish and Swedish?</paragraph> <paragraph><location><page_1><loc_12><loc_59><loc_56><loc_60></location>Costs associated with conference/workshop including:</paragraph>
<paragraph><location><page_1><loc_8><loc_41><loc_44><loc_45></location>Can you spot in a typical issue of Reader's Digest the universal human values that link scholars, statesmen, scientists, writers, businessmen, housewives? Can you pick out the articles that will be most popular with the average 'Digest reader?</paragraph> <paragraph><location><page_1><loc_15><loc_57><loc_23><loc_58></location>e Airfare</paragraph>
<subtitle-level-1><location><page_1><loc_7><loc_38><loc_44><loc_40></location>You may find... you know more about people than you think!</subtitle-level-1> <paragraph><location><page_1><loc_15><loc_55><loc_24><loc_56></location>e Lodging</paragraph>
<paragraph><location><page_1><loc_8><loc_33><loc_44><loc_38></location>Here's all you do. Study the descriptions (at right) of the 'articles in the October Reader's Digest-or, better still, read the complete articles in the issue itself, (But you are not required to buy The Reader's Digest to enter the contest.) Then simply list the six articles-in order of preference-that you think readers of the magazine will like best. This will be compared with a nationwide survey conducted. among a cross séction of Digest subscribers.</paragraph> <paragraph><location><page_1><loc_15><loc_53><loc_23><loc_54></location>e Meals</paragraph>
<paragraph><location><page_1><loc_31><loc_29><loc_36><loc_30></location>midnight,</paragraph> <paragraph><location><page_1><loc_15><loc_51><loc_31><loc_52></location>e Registration fees</paragraph>
<paragraph><location><page_1><loc_7><loc_29><loc_44><loc_33></location>Follow the directions given below. Fill in the entry blank, piste it ona post card, and get it into the mail before the deadline. Additional blanks are obtainable at your college bookstore. 25,</paragraph> <paragraph><location><page_1><loc_15><loc_49><loc_36><loc_50></location>e Ground Transportation</paragraph>
<figure> <subtitle-level-1><location><page_1><loc_12><loc_46><loc_41><loc_47></location>What the Award Will Not Fund</subtitle-level-1>
<location><page_1><loc_7><loc_21><loc_22><loc_27></location> <paragraph><location><page_1><loc_12><loc_44><loc_77><loc_45></location>Any expenses incurred outside of the scope of the proposed development activity.</paragraph>
</figure> <subtitle-level-1><location><page_1><loc_12><loc_40><loc_29><loc_42></location>Granting Schedule</subtitle-level-1>
<paragraph><location><page_1><loc_8><loc_27><loc_44><loc_30></location>obtainable All entries must be postmarked not later than October 25, Don't delay. In case of ties, the entry with the earliest postimarls will win.</paragraph> <paragraph><location><page_1><loc_12><loc_39><loc_32><loc_40></location>Earliest Submission Date:</paragraph>
<paragraph><location><page_1><loc_42><loc_29><loc_44><loc_30></location>1956.</paragraph> <paragraph><location><page_1><loc_33><loc_38><loc_41><loc_40></location>August 1°</paragraph>
<paragraph><location><page_1><loc_23><loc_22><loc_45><loc_27></location>Just pick in order the six: articles you think most readers of October Reader's Digest will like the best.</paragraph> <paragraph><location><page_1><loc_12><loc_36><loc_26><loc_38></location>Applications Due:</paragraph>
<paragraph><location><page_1><loc_9><loc_13><loc_36><loc_21></location>READER'S DIGEST CONTEST, Box 4, Great Neck, L. 1, New York " , In the space opposite the word "FIRST" write the number of the article you think will be the most popular of all. Opposite the word "SECOND" write the number of the article you think will rank second in popularity, List in this -way the numbers of the six top articles in the order of their popularity. (Note: Use only the numbers of articles you choose. Do not write the title of any article.) Clip and Paste this coupon on a Government post card, ,</paragraph> <paragraph><location><page_1><loc_27><loc_37><loc_35><loc_38></location>October 1°</paragraph>
<paragraph><location><page_1><loc_9><loc_13><loc_12><loc_13></location>Name.</paragraph> <paragraph><location><page_1><loc_12><loc_35><loc_31><loc_36></location>Notification of Awards:</paragraph>
<paragraph><location><page_1><loc_9><loc_11><loc_11><loc_12></location>City,</paragraph> <paragraph><location><page_1><loc_31><loc_35><loc_42><loc_36></location>November 1°</paragraph>
<paragraph><location><page_1><loc_28><loc_11><loc_32><loc_12></location>State.</paragraph> <paragraph><location><page_1><loc_12><loc_28><loc_85><loc_32></location>Please submit applications to CLASSGrt@uh.edu by the deadline. Please write "Professional DevelopmentStaff" in the subject line.</paragraph>
<paragraph><location><page_1><loc_15><loc_10><loc_43><loc_11></location>en</paragraph> <paragraph><location><page_1><loc_12><loc_19><loc_86><loc_27></location>PLEASE NOTE: Please include a supporting letter from your Department Chair or Immediate Supervisor. Incomplete applications will not be reviewed. Applications will be considered incomplete until all information has been received, at which time an email confirming receipt will be sent to you.</paragraph>
<paragraph><location><page_1><loc_9><loc_10><loc_16><loc_11></location>Neanrne of COU</paragraph>
<paragraph><location><page_1><loc_22><loc_9><loc_22><loc_9></location>\</paragraph>
<paragraph><location><page_1><loc_37><loc_18><loc_44><loc_21></location>a First - Second</paragraph>
<paragraph><location><page_1><loc_37><loc_16><loc_44><loc_18></location>Thin. Fourth :</paragraph>
<paragraph><location><page_1><loc_45><loc_16><loc_46><loc_22></location>esenansmamemanet</paragraph>
<paragraph><location><page_1><loc_44><loc_10><loc_46><loc_16></location>© -° 4 _ewen esas aemenesan</paragraph>
<paragraph><location><page_1><loc_46><loc_9><loc_46><loc_9></location>|</paragraph>
<paragraph><location><page_1><loc_26><loc_13><loc_30><loc_14></location>Address,</paragraph>
<paragraph><location><page_1><loc_36><loc_16><loc_37><loc_17></location>-</paragraph>
<paragraph><location><page_1><loc_37><loc_16><loc_37><loc_16></location>-</paragraph>
<paragraph><location><page_1><loc_37><loc_15><loc_39><loc_16></location>Fifth</paragraph>
<paragraph><location><page_1><loc_37><loc_14><loc_39><loc_15></location>Sixth</paragraph>
<paragraph><location><page_1><loc_37><loc_14><loc_39><loc_15></location>Sixth</paragraph>
<paragraph><location><page_1><loc_41><loc_52><loc_44><loc_53></location>than</paragraph>
<paragraph><location><page_1><loc_47><loc_22><loc_47><loc_22></location>|</paragraph>
<paragraph><location><page_1><loc_47><loc_27><loc_47><loc_27></location>.</paragraph>
<paragraph><location><page_1><loc_58><loc_8><loc_58><loc_9></location>1</paragraph>
<paragraph><location><page_1><loc_59><loc_92><loc_61><loc_94></location>_</paragraph>
<paragraph><location><page_1><loc_65><loc_29><loc_65><loc_29></location>-</paragraph>
<paragraph><location><page_1><loc_65><loc_37><loc_65><loc_37></location>-</paragraph>
<paragraph><location><page_1><loc_67><loc_74><loc_67><loc_75></location>.</paragraph>
<paragraph><location><page_1><loc_67><loc_70><loc_67><loc_70></location>_</paragraph>
<paragraph><location><page_1><loc_67><loc_78><loc_67><loc_78></location>_</paragraph>
<paragraph><location><page_1><loc_68><loc_23><loc_68><loc_23></location>.</paragraph>
<subtitle-level-1><location><page_1><loc_68><loc_85><loc_90><loc_89></location>Which six articles will readers of the October Digest like best?</subtitle-level-1>
<paragraph><location><page_1><loc_68><loc_84><loc_90><loc_85></location>'T.. Norfolk's friend to troubled feen-agers. Story of the arthritic cripple to whom youngsters flock for advice.</paragraph>
<paragraph><location><page_1><loc_68><loc_82><loc_90><loc_84></location>2. The great Piltdown hoax. How this famed "missing link'' in human evolution has been proved a fraud from the start.</paragraph>
<paragraph><location><page_1><loc_68><loc_81><loc_90><loc_82></location>3. How fo sharpen your judgment, Famed author Bertrand Russell offers six rules to heip you form sounder opinions.</paragraph>
<paragraph><location><page_1><loc_68><loc_80><loc_68><loc_81></location>4.</paragraph>
<paragraph><location><page_1><loc_68><loc_79><loc_77><loc_81></location>My most unforgettable nie Mack-who led the</paragraph>
<paragraph><location><page_1><loc_81><loc_79><loc_87><loc_81></location>Fond memories for 50 years.</paragraph>
<paragraph><location><page_1><loc_77><loc_80><loc_81><loc_81></location>character.</paragraph>
<paragraph><location><page_1><loc_77><loc_79><loc_81><loc_80></location>Athletics</paragraph>
<paragraph><location><page_1><loc_68><loc_78><loc_90><loc_79></location>5. How to make peace at the Pentagon. Steps to end ruinous rivalry between our Army, Navy and Air Force,</paragraph>
<paragraph><location><page_1><loc_86><loc_75><loc_90><loc_76></location>rosarchere</paragraph>
<paragraph><location><page_1><loc_68><loc_76><loc_90><loc_79></location>Army, 6. Book condensation: "High, Wide and Lonesome."' Hal Borland's exciting story of his adventurous: boyhood on a Colorado prairie.</paragraph>
<paragraph><location><page_1><loc_68><loc_74><loc_90><loc_76></location>7. Medicine's animal pioneers. How medical learn from animals new ways to save human lives,</paragraph>
<paragraph><location><page_1><loc_68><loc_73><loc_90><loc_74></location>8. What the mess in Moscow means. Eyidence that the Communist system is as unworkable as it is unnatural.</paragraph>
<paragraph><location><page_1><loc_68><loc_71><loc_90><loc_73></location>9. Master bridge builder. Introducing David Steinman, world leader in bridge design and construction.</paragraph>
<paragraph><location><page_1><loc_68><loc_70><loc_90><loc_71></location>10. College two years sooner. Here's how extensive ex periments proved a bright 10th-grader is ready for college.</paragraph>
<paragraph><location><page_1><loc_68><loc_68><loc_90><loc_70></location>11. Laughter the best medicine. Amusing experiences from everyday life. others?</paragraph>
<paragraph><location><page_1><loc_68><loc_66><loc_90><loc_68></location>12. What happens when we pray for others? Too often we pray only for ourselves. Here's how we gain true rewards of prayer when we pray for others.</paragraph>
<paragraph><location><page_1><loc_68><loc_65><loc_90><loc_66></location>13. European vs. U.S, beaviies. Why European women are more glamorous to men,</paragraph>
<paragraph><location><page_1><loc_68><loc_63><loc_90><loc_65></location>14. Trading stamps-bonus or bunkum? How much of their cost is included in the price you pay?</paragraph>
<paragraph><location><page_1><loc_68><loc_62><loc_90><loc_63></location>15. Living memorials instead of flowers. A way toh honor the dead by serving the living.</paragraph>
<paragraph><location><page_1><loc_68><loc_60><loc_90><loc_62></location>16. ft pays to increase your word power. An entertaining quiz to build your vocabulary.</paragraph>
<paragraph><location><page_1><loc_68><loc_59><loc_90><loc_60></location>17, Are we too soft on young criminals? Why the best way to cure juvenile delinquency is to punish first offenders.</paragraph>
<paragraph><location><page_1><loc_68><loc_57><loc_90><loc_59></location>18, Medicine man on the Amazon. How two devoted missionaries bring medical aid to jungle natives.</paragraph>
<paragraph><location><page_1><loc_68><loc_56><loc_90><loc_57></location>19. Creatures in the night. The fascinating drama of nature that is enacted between dusk and dawn..</paragraph>
<paragraph><location><page_1><loc_68><loc_53><loc_90><loc_54></location>21. The sub that wouldn't stay down, Stirring saga of the U.S.So Squalus' rescue from a depth of 40 fathoms.</paragraph>
<paragraph><location><page_1><loc_68><loc_54><loc_90><loc_56></location>20. What your sense of humor fells about you. What the jokes you like, the way you laugh reveal about you.</paragraph>
<paragraph><location><page_1><loc_68><loc_51><loc_90><loc_53></location>22. Madame Butterfiyi in bobby sox. How new freedoms have changed life for Japanese women; what the men think.</paragraph>
<paragraph><location><page_1><loc_68><loc_49><loc_90><loc_51></location>23. Dactors should tell patients the truth. When the doctor operated, exactly what did he do? Why a written record of your medical history may someday save your life.</paragraph>
<paragraph><location><page_1><loc_68><loc_47><loc_90><loc_49></location>24. 'How wonderful you are..." Here's why affection and admiration aren't much good unless expressed; why locked-up emotions eventually wither.</paragraph>
<paragraph><location><page_1><loc_68><loc_45><loc_90><loc_47></location>25. Harry Holt and a heariful of children, Story of a farmer 'who singlehandedly finds homes for hundreds of Korean war orphans.</paragraph>
<paragraph><location><page_1><loc_68><loc_43><loc_90><loc_45></location>26. Our tax laws make us dishonest. How unfair tax laws are causing a serious moral deterioration.</paragraph>
<paragraph><location><page_1><loc_68><loc_42><loc_90><loc_43></location>27. Venercal disease now a threat to youth. How V.D. i is spreading among teen-agers-and sane advice to victims.</paragraph>
<paragraph><location><page_1><loc_68><loc_40><loc_91><loc_42></location>28. Secy. Benson's faith in the Amorican former. Why he" feels farmers, left alone, can often solve their own probe lems better than Washington. i</paragraph>
<paragraph><location><page_1><loc_68><loc_38><loc_90><loc_40></location>29, Your brain's unrealized powers. Seven new findings to help you use your brain more efficiently.</paragraph>
<paragraph><location><page_1><loc_68><loc_37><loc_90><loc_38></location>30. Britain's indestructible "'Old Man.'? What Sir Winston Churchill is doing in retirement,</paragraph>
<paragraph><location><page_1><loc_68><loc_35><loc_68><loc_35></location>_</paragraph>
<paragraph><location><page_1><loc_68><loc_35><loc_90><loc_37></location>31. Are juries giving away too much money? Fantastic awards juries hand out because they confuse compassion with common sense,</paragraph>
<paragraph><location><page_1><loc_68><loc_33><loc_90><loc_34></location>32. My last bost days on earth. In her own words a young mother, learning she had cancer, tells how she decided to make this the "'best year of her life."</paragraph>
<paragraph><location><page_1><loc_68><loc_31><loc_90><loc_32></location>33. Foreign-aid mania. How the billions we've given have brought mainly disappointment and higher taxes.</paragraph>
<paragraph><location><page_1><loc_68><loc_29><loc_90><loc_31></location>34, Out whero fet planes are born. Story of Edward Air Force Base, where 10,000 men. battle wind, sand and speed barriers to keep us supreme in tho sky. %</paragraph>
<paragraph><location><page_1><loc_68><loc_26><loc_91><loc_28></location>36, Man's most playful friend: tho Lend 'Otter. Interesting 'facts about this amusing fnimal.</paragraph>
<paragraph><location><page_1><loc_68><loc_27><loc_91><loc_29></location>35. Life In thege United States. Humorous anecdotes reveal. ing quirks of human nature,</paragraph>
<paragraph><location><page_1><loc_68><loc_24><loc_91><loc_26></location>37. Why not a foreign service careor? 'How our State De. partment ismaking foreign sorviceattractive toy oung men,</paragraph>
<paragraph><location><page_1><loc_68><loc_23><loc_91><loc_24></location>38. A now deal in tho old firehouse. How one-town got lower taxes, greater protection combining firo und police,</paragraph>
<paragraph><location><page_1><loc_68><loc_21><loc_91><loc_23></location>39. Crazy man on Crazy Horse, Mect the man whoso statue of an Indian will be the largest in history,</paragraph>
<paragraph><location><page_1><loc_68><loc_20><loc_91><loc_21></location>40. Tholr business is dynamite. How the manufacture of this explosive has been made one of the safest industries,</paragraph>
<paragraph><location><page_1><loc_68><loc_17><loc_91><loc_20></location>41. His best customers ara bablos, How a kitchen strainer and a pint of mashed peas became the Gerber Produets Co, 42, Smoky Mountain magic, Why this, our most. ancient mountain range, has more visitors than any other,</paragraph>
<paragraph><location><page_1><loc_80><loc_15><loc_84><loc_16></location>landacape</paragraph>
<paragraph><location><page_1><loc_68><loc_15><loc_91><loc_17></location>43. Call for Mr. Emorgoncy, Meot the Io Smergoncy Polico, who got 8 million New Yorkers put of troublo,</paragraph>
<paragraph><location><page_1><loc_68><loc_14><loc_91><loc_15></location>44, Boauly by tho mile, Now engineers prove rondsido planting is lifesaving as well as beautiful,</paragraph>
<paragraph><location><page_1><loc_68><loc_13><loc_91><loc_14></location>45. Humor in uniform, True stories of tho funny side of life in our Armed Forcas,</paragraph>
<paragraph><location><page_1><loc_68><loc_10><loc_91><loc_12></location>46. Sovon economic Fallacies. The Amorican 'Economie Foundation explodes misconceptions xbout our economy, 47. Admiral of tho Greak Ol! Floot, Story of Stavros Ninrchos, who has won afortuno betting on-and enrrying-oll,</paragraph>
<paragraph><location><page_1><loc_73><loc_9><loc_74><loc_9></location>"</paragraph>
<paragraph><location><page_1><loc_87><loc_80><loc_90><loc_81></location>of Con-</paragraph>
<paragraph><location><page_1><loc_91><loc_67><loc_91><loc_67></location>-</paragraph>
<paragraph><location><page_1><loc_91><loc_63><loc_91><loc_63></location>~</paragraph>
<paragraph><location><page_1><loc_91><loc_23><loc_91><loc_23></location>.</paragraph>
<paragraph><location><page_1><loc_94><loc_17><loc_95><loc_19></location>-Sarnn, Pitt 27</paragraph>
<paragraph><location><page_1><loc_95><loc_16><loc_95><loc_17></location>eae</paragraph>
<paragraph><location><page_1><loc_96><loc_77><loc_96><loc_80></location>a</paragraph>
</document> </document>

2
tests/data_scanned/scanned_02.tesserocr.json
View File

File diff suppressed because one or more lines are too long

269
tests/data_scanned/scanned_02.tesserocr.md
View File

@ -1,271 +1,46 @@
You Can Win a Cash Award and Scholarship Money for Your College in
Qpen to All College Students (Faauty, to!
...and you may find you know more about people than you think!
Nothing to buy...nothing to write
'How well do you know human nature? Can you tell what subjects interest people most? Here is a chance to test your judgment-show how good an editor you are-and you may win $5,000 for yourself, plus $5, 000 in scholarship funds for your college.
|
_
## You CAN WIN:
$5000
cash
i**
prize
plus $5000 for the scholarship fund of your college or...
## $1000 cash 2" prize
Any of TEN $500 cash. prizes plus $500 for the scholarship :.. fund of your college or...
plus $1000 for the scholarship fund of your college or...
## Any of 100 $10 prizes
in
book
credit
from
Jocal
college
your bookstore
_
And if your entry is the best from your college you will receive an extra award -an additional $10 in book credit. at your college bookstore,
## FOLLOW THESE EASY-RULES
1 . Read the descriptions in this ad vertisement of the articles that appear in» October Reader's Digest. Or, better, read the complete articles, Then select the 6 that you think most readers will like best.
2.On the entry blank at left, write the number of each article you select. List them in what you think. will be the: order of popularity, from first to sixth place. Your selections will be judged by comparison with a national survey which ranks in order of popularity the 6 articles that readers like best. Fill in -and mail the coupon. All entries must be postmarked not later than midnight, October 25, 1956:
3. This contest is open only to college students and faculty members in the U. S., excluding employees of The Reader's Digest, its advertising agencies, and their families, It is subject to 'all federal, state and local Jaws and regulations.
4, Only one entry per person.
5. In case of ties, entries postmarked earliest will win, Entries will be judged by, O. E. McIntyre, Inc., whose decision will be final. All entries become property of The Reader's Digest; none returned...
6. All winners notified by'mail. List of cash-prize winners mailed if you pe ,Stamped enveope,
<!-- image --> <!-- image -->
'. ## Purpose
Its Poptart and influence are world-wide The Dean's Professional Development Award for Staff is to allow CLASS staff the opportunity to attend conferences and workshops in their field for the sole purpose of professional development. The intent is to defray costs associated with attendance. The maximum amount of the award is $2,000 per staff member. Up to four awards will be made per year, contingent upon the availability of funding. Staff members that are awarded must wait three years from the date of award notification before reapplying again.
I ## Eligibility
It's fun to try. Maybe you can top other students in colleges across the country ... and you can match wits with the editors of Reader's Digest. All staff currently employed in CLASS are eligible.
- ## What the Award Will Fund
_ Why'do far more college graduates read Reader's Digest any other magazine? What is it that makes the Digest the-most widely read magazine in the world - with 11 million copies bought each month in the United States, plus 9 million abroad? Why is it read each month by at least 60 million people, in 12 languages-Arabic, Danish, English, Finnish, French, German, Italian, Japanese, Norwegian, Portuguese, Spanish and Swedish? Costs associated with conference/workshop including:
Can you spot in a typical issue of Reader's Digest the universal human values that link scholars, statesmen, scientists, writers, businessmen, housewives? Can you pick out the articles that will be most popular with the average 'Digest reader? e Airfare
## You may find... you know more about people than you think! e Lodging
Here's all you do. Study the descriptions (at right) of the 'articles in the October Reader's Digest-or, better still, read the complete articles in the issue itself, (But you are not required to buy The Reader's Digest to enter the contest.) Then simply list the six articles-in order of preference-that you think readers of the magazine will like best. This will be compared with a nationwide survey conducted. among a cross séction of Digest subscribers. e Meals
midnight, e Registration fees
Follow the directions given below. Fill in the entry blank, piste it ona post card, and get it into the mail before the deadline. Additional blanks are obtainable at your college bookstore. 25, e Ground Transportation
## What the Award Will Not Fund
<!-- image --> Any expenses incurred outside of the scope of the proposed development activity.
obtainable All entries must be postmarked not later than October 25, Don't delay. In case of ties, the entry with the earliest postimarls will win. ## Granting Schedule
1956. Earliest Submission Date:
Just pick in order the six: articles you think most readers of October Reader's Digest will like the best. August 1°
READER'S DIGEST CONTEST, Box 4, Great Neck, L. 1, New York " , In the space opposite the word "FIRST" write the number of the article you think will be the most popular of all. Opposite the word "SECOND" write the number of the article you think will rank second in popularity, List in this -way the numbers of the six top articles in the order of their popularity. (Note: Use only the numbers of articles you choose. Do not write the title of any article.) Clip and Paste this coupon on a Government post card, , Applications Due:
Name. October 1°
City, Notification of Awards:
State. November 1°
en Please submit applications to CLASSGrt@uh.edu by the deadline. Please write "Professional DevelopmentStaff" in the subject line.
Neanrne of COU PLEASE NOTE: Please include a supporting letter from your Department Chair or Immediate Supervisor. Incomplete applications will not be reviewed. Applications will be considered incomplete until all information has been received, at which time an email confirming receipt will be sent to you.
\
a First - Second
Thin. Fourth :
esenansmamemanet
© -° 4 _ewen esas aemenesan
|
Address,
-
Fifth
Sixth
than
|
.
1
_
-
.
_
.
## Which six articles will readers of the October Digest like best?
'T.. Norfolk's friend to troubled feen-agers. Story of the arthritic cripple to whom youngsters flock for advice.
2. The great Piltdown hoax. How this famed "missing link'' in human evolution has been proved a fraud from the start.
3. How fo sharpen your judgment, Famed author Bertrand Russell offers six rules to heip you form sounder opinions.
4.
My most unforgettable nie Mack-who led the
Fond memories for 50 years.
character.
Athletics
5. How to make peace at the Pentagon. Steps to end ruinous rivalry between our Army, Navy and Air Force,
rosarchere
Army, 6. Book condensation: "High, Wide and Lonesome."' Hal Borland's exciting story of his adventurous: boyhood on a Colorado prairie.
7. Medicine's animal pioneers. How medical learn from animals new ways to save human lives,
8. What the mess in Moscow means. Eyidence that the Communist system is as unworkable as it is unnatural.
9. Master bridge builder. Introducing David Steinman, world leader in bridge design and construction.
10. College two years sooner. Here's how extensive ex periments proved a bright 10th-grader is ready for college.
11. Laughter the best medicine. Amusing experiences from everyday life. others?
12. What happens when we pray for others? Too often we pray only for ourselves. Here's how we gain true rewards of prayer when we pray for others.
13. European vs. U.S, beaviies. Why European women are more glamorous to men,
14. Trading stamps-bonus or bunkum? How much of their cost is included in the price you pay?
15. Living memorials instead of flowers. A way toh honor the dead by serving the living.
16. ft pays to increase your word power. An entertaining quiz to build your vocabulary.
17, Are we too soft on young criminals? Why the best way to cure juvenile delinquency is to punish first offenders.
18, Medicine man on the Amazon. How two devoted missionaries bring medical aid to jungle natives.
19. Creatures in the night. The fascinating drama of nature that is enacted between dusk and dawn..
21. The sub that wouldn't stay down, Stirring saga of the U.S.So Squalus' rescue from a depth of 40 fathoms.
20. What your sense of humor fells about you. What the jokes you like, the way you laugh reveal about you.
22. Madame Butterfiyi in bobby sox. How new freedoms have changed life for Japanese women; what the men think.
23. Dactors should tell patients the truth. When the doctor operated, exactly what did he do? Why a written record of your medical history may someday save your life.
24. 'How wonderful you are..." Here's why affection and admiration aren't much good unless expressed; why locked-up emotions eventually wither.
25. Harry Holt and a heariful of children, Story of a farmer 'who singlehandedly finds homes for hundreds of Korean war orphans.
26. Our tax laws make us dishonest. How unfair tax laws are causing a serious moral deterioration.
27. Venercal disease now a threat to youth. How V.D. i is spreading among teen-agers-and sane advice to victims.
28. Secy. Benson's faith in the Amorican former. Why he" feels farmers, left alone, can often solve their own probe lems better than Washington. i
29, Your brain's unrealized powers. Seven new findings to help you use your brain more efficiently.
30. Britain's indestructible "'Old Man.'? What Sir Winston Churchill is doing in retirement,
_
31. Are juries giving away too much money? Fantastic awards juries hand out because they confuse compassion with common sense,
32. My last bost days on earth. In her own words a young mother, learning she had cancer, tells how she decided to make this the "'best year of her life."
33. Foreign-aid mania. How the billions we've given have brought mainly disappointment and higher taxes.
34, Out whero fet planes are born. Story of Edward Air Force Base, where 10,000 men. battle wind, sand and speed barriers to keep us supreme in tho sky. %
36, Man's most playful friend: tho Lend 'Otter. Interesting 'facts about this amusing fnimal.
35. Life In thege United States. Humorous anecdotes reveal. ing quirks of human nature,
37. Why not a foreign service careor? 'How our State De. partment ismaking foreign sorviceattractive toy oung men,
38. A now deal in tho old firehouse. How one-town got lower taxes, greater protection combining firo und police,
39. Crazy man on Crazy Horse, Mect the man whoso statue of an Indian will be the largest in history,
40. Tholr business is dynamite. How the manufacture of this explosive has been made one of the safest industries,
41. His best customers ara bablos, How a kitchen strainer and a pint of mashed peas became the Gerber Produets Co, 42, Smoky Mountain magic, Why this, our most. ancient mountain range, has more visitors than any other,
landacape
43. Call for Mr. Emorgoncy, Meot the Io Smergoncy Polico, who got 8 million New Yorkers put of troublo,
44, Boauly by tho mile, Now engineers prove rondsido planting is lifesaving as well as beautiful,
45. Humor in uniform, True stories of tho funny side of life in our Armed Forcas,
46. Sovon economic Fallacies. The Amorican 'Economie Foundation explodes misconceptions xbout our economy, 47. Admiral of tho Greak Ol! Floot, Story of Stavros Ninrchos, who has won afortuno betting on-and enrrying-oll,
"
of Con-
-
~
.
-Sarnn, Pitt 27
eae
a

2
tests/data_scanned/scanned_02.tesserocr.pages.json
View File

File diff suppressed because one or more lines are too long

21
tests/data_scanned/scanned_03.easyocr.doctags.txt
View File

@ -1,21 +0,0 @@
<document>
<subtitle-level-1><location><page_1><loc_3><loc_83><loc_95><loc_97></location>Students around the country are noticing LaSalle</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_4><loc_78><loc_92><loc_83></location>Some of them have questions you can answer. only</subtitle-level-1>
<paragraph><location><page_1><loc_4><loc_66><loc_46><loc_77></location>Alumni are the most credible source of information for many prospective students and their parents. That's we re inviting you to become part of the La Salle Explorer Alumni Recruiting Network (LEARN). There are several ways you can share your La Salle experiences with students and parents: why</paragraph>
<paragraph><location><page_1><loc_7><loc_61><loc_45><loc_65></location>Represent the university at college nights and receptions in your area.</paragraph>
<paragraph><location><page_1><loc_7><loc_59><loc_44><loc_61></location>Call students who are interested in La Salle.</paragraph>
<paragraph><location><page_1><loc_7><loc_53><loc_43><loc_59></location>Write a letter on company stationery about your work experience and La Salle education. your</paragraph>
<paragraph><location><page_1><loc_7><loc_49><loc_44><loc_53></location>Visit high schools to discuss your profession and your preparation at La Salle.</paragraph>
<paragraph><location><page_1><loc_4><loc_40><loc_47><loc_47></location>If you are interested in helping the university recruit tomorrow's graduates, please call Maryhelen Driscoll, associate director of admissions, at 215/951-1500. Or return the coupon below.</paragraph>
<paragraph><location><page_1><loc_6><loc_32><loc_45><loc_37></location>Id like to play a role in recruiting students to La Salle.</paragraph>
<paragraph><location><page_1><loc_6><loc_29><loc_10><loc_30></location>Name</paragraph>
<paragraph><location><page_1><loc_6><loc_25><loc_12><loc_27></location>Address</paragraph>
<paragraph><location><page_1><loc_6><loc_22><loc_16><loc_24></location>City /State/Zip</paragraph>
<paragraph><location><page_1><loc_6><loc_19><loc_19><loc_21></location>Telephone (home)</paragraph>
<paragraph><location><page_1><loc_14><loc_16><loc_18><loc_17></location>(work)</paragraph>
<paragraph><location><page_1><loc_6><loc_11><loc_31><loc_16></location>Mail to: Office of Admissions La Salle University Philadelphia, PA 19141-1199</paragraph>
<paragraph><location><page_1><loc_9><loc_6><loc_42><loc_10></location>LA SALLE ExPLORER ALUMNI RECRUITING NETWORK</paragraph>
<figure>
<location><page_1><loc_55><loc_7><loc_92><loc_75></location>
</figure>
</document>

1
tests/data_scanned/scanned_03.easyocr.json
View File

@ -1 +0,0 @@
{"_name": "", "type": "pdf-document", "description": {"logs": []}, "file-info": {"filename": "scanned_03.pdf", "document-hash": "d3118d34c5cee331763fea26f416acca5b18d774858219fedfab39701eed809b", "#-pages": 1, "page-hashes": [{"hash": "c6a54f3a62cec9cb8bd23964df5de476dd8a37c655f76a9bb0e754448837e809", "model": "default", "page": 1}]}, "main-text": [{"prov": [{"bbox": [41.66666793823242, 1351.0, 1174.0, 1573.0], "page": 1, "span": [0, 48]}], "text": "Students around the country are noticing LaSalle", "type": "subtitle-level-1", "name": "Section-header"}, {"prov": [{"bbox": [43.66666793823242, 1278.67724609375, 1145.6666259765625, 1350.9893798828125], "page": 1, "span": [0, 48]}], "text": "Some of them have questions you can answer. only", "type": "subtitle-level-1", "name": "Section-header"}, {"prov": [{"bbox": [47.33333206176758, 1078.3333740234375, 572.0, 1253.0433349609375], "page": 1, "span": [0, 294]}], "text": "Alumni are the most credible source of information for many prospective students and their parents. That's we re inviting you to become part of the La Salle Explorer Alumni Recruiting Network (LEARN). There are several ways you can share your La Salle experiences with students and parents: why", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [88.01899719238281, 989.6666870117188, 558.0, 1055.3333740234375], "page": 1, "span": [0, 71]}], "text": "Represent the university at college nights and receptions in your area.", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [88.30470275878906, 962.0, 541.6666870117188, 987.6666870117188], "page": 1, "span": [0, 45]}], "text": "Call students who are interested in La Salle.", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [88.37541198730469, 870.3333129882812, 536.0, 954.6666870117188], "page": 1, "span": [0, 92]}], "text": "Write a letter on company stationery about your work experience and La Salle education. your", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [88.19837188720703, 800.3333129882812, 548.0, 864.0], "page": 1, "span": [0, 79]}], "text": "Visit high schools to discuss your profession and your preparation at La Salle.", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [48.0, 652.3333129882812, 582.0, 771.6666870117188], "page": 1, "span": [0, 188]}], "text": "If you are interested in helping the university recruit tomorrow's graduates, please call Maryhelen Driscoll, associate director of admissions, at 215/951-1500. Or return the coupon below.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [71.33333587646484, 529.6666870117188, 562.0, 608.0], "page": 1, "span": [0, 58]}], "text": "Id like to play a role in recruiting students to La Salle.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [76.33333587646484, 465.0, 129.6666717529297, 484.33544921875], "page": 1, "span": [0, 4]}], "text": "Name", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [75.0, 412.3333435058594, 147.6666717529297, 437.3333435058594], "page": 1, "span": [0, 7]}], "text": "Address", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [75.66666412353516, 361.0, 198.0, 385.3333435058594], "page": 1, "span": [0, 15]}], "text": "City /State/Zip", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [75.66666412353516, 311.3333435058594, 232.3333282470703, 335.6666564941406], "page": 1, "span": [0, 16]}], "text": "Telephone (home)", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [167.6666717529297, 263.3333435058594, 228.3333282470703, 284.6666564941406], "page": 1, "span": [0, 6]}], "text": "(work)", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [77.0, 182.6666717529297, 387.6666564941406, 252.6666717529297], "page": 1, "span": [0, 77]}], "text": "Mail to: Office of Admissions La Salle University Philadelphia, PA 19141-1199", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [109.15133666992188, 92.33333587646484, 526.0, 155.0], "page": 1, "span": [0, 43]}], "text": "LA SALLE ExPLORER ALUMNI RECRUITING NETWORK", "type": "paragraph", "name": "Text"}, {"name": "Picture", "type": "figure", "$ref": "#/figures/0"}, {"prov": [{"bbox": [680.6666870117188, 53.33333206176758, 1128.0, 79.33333587646484], "page": 1, "span": [0, 61]}], "text": "La Salle appears in many of the nation 's leading guidebooks.", "type": "page-footer", "name": "Page-footer"}], "figures": [{"prov": [{"bbox": [685.5791015625, 107.46976470947266, 1146.682373046875, 1225.0693359375], "page": 1, "span": [0, 0]}], "text": "", "type": "figure"}], "tables": [], "equations": [], "footnotes": [], "page-dimensions": [{"height": 1630.0, "page": 1, "width": 1241.0}], "page-footers": [], "page-headers": []}

34
tests/data_scanned/scanned_03.easyocr.md
View File

@ -1,34 +0,0 @@
## Students around the country are noticing LaSalle
## Some of them have questions you can answer. only
Alumni are the most credible source of information for many prospective students and their parents. That's we re inviting you to become part of the La Salle Explorer Alumni Recruiting Network (LEARN). There are several ways you can share your La Salle experiences with students and parents: why
Represent the university at college nights and receptions in your area.
Call students who are interested in La Salle.
Write a letter on company stationery about your work experience and La Salle education. your
Visit high schools to discuss your profession and your preparation at La Salle.
If you are interested in helping the university recruit tomorrow's graduates, please call Maryhelen Driscoll, associate director of admissions, at 215/951-1500. Or return the coupon below.
Id like to play a role in recruiting students to La Salle.
Name
Address
City /State/Zip
Telephone (home)
(work)
Mail to: Office of Admissions La Salle University Philadelphia, PA 19141-1199
LA SALLE ExPLORER ALUMNI RECRUITING NETWORK
<!-- image -->

1
tests/data_scanned/scanned_03.easyocr.pages.json
View File

File diff suppressed because one or more lines are too long

BIN
tests/data_scanned/scanned_03.pdf
View File

Binary file not shown.

BIN
tests/data_scanned/scanned_03.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 1.6 MiB

21
tests/data_scanned/scanned_03.tesserocr.doctags.txt
View File

@ -1,21 +0,0 @@
<document>
<subtitle-level-1><location><page_1><loc_4><loc_83><loc_93><loc_96></location>Students around the country are noticing La Salle.</subtitle-level-1>
<subtitle-level-1><location><page_1><loc_4><loc_79><loc_92><loc_82></location>Some of them have questions only you can answer.</subtitle-level-1>
<paragraph><location><page_1><loc_4><loc_66><loc_46><loc_77></location>Alumni are the most credible source of information for many prospective students and their parents. That's why we're inviting you to become part of the La Salle Explorer Alumni Recruiting Network (L.E.A.R.N.). There are several ways you can share your La Salle experiences with students and parents:</paragraph>
<paragraph><location><page_1><loc_7><loc_61><loc_45><loc_64></location>e Represent the university at college nights and receptions in your area.</paragraph>
<paragraph><location><page_1><loc_7><loc_59><loc_43><loc_61></location>e Call students who are interested in La Salle.</paragraph>
<paragraph><location><page_1><loc_7><loc_54><loc_43><loc_58></location>e Write a letter on company stationery about your work experience and your La Salle education.</paragraph>
<paragraph><location><page_1><loc_7><loc_49><loc_44><loc_53></location>¢ Visit high schools to discuss your profession and your preparation at La Salle.</paragraph>
<paragraph><location><page_1><loc_4><loc_40><loc_47><loc_47></location>If you are interested in helping the university recruit tomorrow's graduates, please call Maryhelen Driscoll, associate director of admissions, at 215/951-1500. Or return the coupon below.</paragraph>
<paragraph><location><page_1><loc_6><loc_33><loc_45><loc_37></location>I'd like to play a role in recruiting students to La Salle.</paragraph>
<paragraph><location><page_1><loc_6><loc_29><loc_10><loc_30></location>Name</paragraph>
<paragraph><location><page_1><loc_6><loc_26><loc_12><loc_27></location>Address</paragraph>
<paragraph><location><page_1><loc_6><loc_22><loc_16><loc_24></location>City/State/Zip</paragraph>
<paragraph><location><page_1><loc_6><loc_19><loc_19><loc_21></location>Telephone (home)</paragraph>
<paragraph><location><page_1><loc_14><loc_16><loc_18><loc_17></location>(work)</paragraph>
<paragraph><location><page_1><loc_6><loc_11><loc_31><loc_15></location>oe . . Mail to: Office of Admissions La Salle University Philadelphia, PA 19141-1199</paragraph>
<paragraph><location><page_1><loc_9><loc_6><loc_42><loc_9></location>LA SALLE EXPLORER ALUMNI RECRUITING NETWORK</paragraph>
<figure>
<location><page_1><loc_55><loc_7><loc_92><loc_75></location>
</figure>
</document>

1
tests/data_scanned/scanned_03.tesserocr.json
View File

@ -1 +0,0 @@
{"_name": "", "type": "pdf-document", "description": {"logs": []}, "file-info": {"filename": "scanned_03.pdf", "document-hash": "d3118d34c5cee331763fea26f416acca5b18d774858219fedfab39701eed809b", "#-pages": 1, "page-hashes": [{"hash": "c6a54f3a62cec9cb8bd23964df5de476dd8a37c655f76a9bb0e754448837e809", "model": "default", "page": 1}]}, "main-text": [{"prov": [{"bbox": [49.032997131347656, 1359.27392578125, 1160.3310546875, 1561.4920654296875], "page": 1, "span": [0, 50]}], "text": "Students around the country are noticing La Salle.", "type": "subtitle-level-1", "name": "Section-header"}, {"prov": [{"bbox": [50.13961410522461, 1285.441650390625, 1139.0716552734375, 1340.6978759765625], "page": 1, "span": [0, 48]}], "text": "Some of them have questions only you can answer.", "type": "subtitle-level-1", "name": "Section-header"}, {"prov": [{"bbox": [48.61845397949219, 1081.7232666015625, 565.8486938476562, 1253.0433349609375], "page": 1, "span": [0, 299]}], "text": "Alumni are the most credible source of information for many prospective students and their parents. That's why we're inviting you to become part of the La Salle Explorer Alumni Recruiting Network (L.E.A.R.N.). There are several ways you can share your La Salle experiences with students and parents:", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [88.01899719238281, 996.06640625, 555.45751953125, 1051.1669921875], "page": 1, "span": [0, 73]}], "text": "e Represent the university at college nights and receptions in your area.", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [88.30470275878906, 963.8839111328125, 538.364501953125, 987.3241577148438], "page": 1, "span": [0, 47]}], "text": "e Call students who are interested in La Salle.", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [88.37541198730469, 872.5484619140625, 532.9567260742188, 951.9700317382812], "page": 1, "span": [0, 94]}], "text": "e Write a letter on company stationery about your work experience and your La Salle education.", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [88.19837188720703, 805.2804565429688, 545.2137451171875, 860.2967529296875], "page": 1, "span": [0, 81]}], "text": "\u00a2 Visit high schools to discuss your profession and your preparation at La Salle.", "type": "paragraph", "name": "List-item"}, {"prov": [{"bbox": [49.349178314208984, 654.9208984375, 579.4306640625, 767.715576171875], "page": 1, "span": [0, 188]}], "text": "If you are interested in helping the university recruit tomorrow's graduates, please call Maryhelen Driscoll, associate director of admissions, at 215/951-1500. Or return the coupon below.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [75.11003112792969, 532.2311401367188, 557.2638549804688, 601.1927490234375], "page": 1, "span": [0, 59]}], "text": "I'd like to play a role in recruiting students to La Salle.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [76.95063781738281, 467.0712890625, 127.2648696899414, 484.33544921875], "page": 1, "span": [0, 4]}], "text": "Name", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [77.28118133544922, 417.2911071777344, 144.09515380859375, 433.92181396484375], "page": 1, "span": [0, 7]}], "text": "Address", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [76.7642822265625, 363.6166076660156, 195.35006713867188, 384.2001647949219], "page": 1, "span": [0, 14]}], "text": "City/State/Zip", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [76.54296875, 313.0050964355469, 229.5892333984375, 334.3070373535156], "page": 1, "span": [0, 16]}], "text": "Telephone (home)", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [170.29672241210938, 264.3503723144531, 225.14401245117188, 284.4351501464844], "page": 1, "span": [0, 6]}], "text": "(work)", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [77.09578704833984, 185.03317260742188, 383.94482421875, 251.23171997070312], "page": 1, "span": [0, 84]}], "text": "oe . . Mail to: Office of Admissions La Salle University Philadelphia, PA 19141-1199", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [109.15133666992188, 96.73243713378906, 524.3324584960938, 153.5020751953125], "page": 1, "span": [0, 43]}], "text": "LA SALLE EXPLORER ALUMNI RECRUITING NETWORK", "type": "paragraph", "name": "Text"}, {"name": "Picture", "type": "figure", "$ref": "#/figures/0"}, {"prov": [{"bbox": [682.0344848632812, 55.896629333496094, 1127.4478759765625, 75.76051330566406], "page": 1, "span": [0, 60]}], "text": "La Salle appears in many of the nation's leading guidebooks.", "type": "page-footer", "name": "Page-footer"}], "figures": [{"prov": [{"bbox": [685.5791015625, 107.46976470947266, 1146.682373046875, 1225.0693359375], "page": 1, "span": [0, 0]}], "text": "", "type": "figure"}], "tables": [], "equations": [], "footnotes": [], "page-dimensions": [{"height": 1630.0, "page": 1, "width": 1241.0}], "page-footers": [], "page-headers": []}

34
tests/data_scanned/scanned_03.tesserocr.md
View File

@ -1,34 +0,0 @@
## Students around the country are noticing La Salle.
## Some of them have questions only you can answer.
Alumni are the most credible source of information for many prospective students and their parents. That's why we're inviting you to become part of the La Salle Explorer Alumni Recruiting Network (L.E.A.R.N.). There are several ways you can share your La Salle experiences with students and parents:
e Represent the university at college nights and receptions in your area.
e Call students who are interested in La Salle.
e Write a letter on company stationery about your work experience and your La Salle education.
¢ Visit high schools to discuss your profession and your preparation at La Salle.
If you are interested in helping the university recruit tomorrow's graduates, please call Maryhelen Driscoll, associate director of admissions, at 215/951-1500. Or return the coupon below.
I'd like to play a role in recruiting students to La Salle.
Name
Address
City/State/Zip
Telephone (home)
(work)
oe . . Mail to: Office of Admissions La Salle University Philadelphia, PA 19141-1199
LA SALLE EXPLORER ALUMNI RECRUITING NETWORK
<!-- image -->

1
tests/data_scanned/scanned_03.tesserocr.pages.json
View File

File diff suppressed because one or more lines are too long

14
tests/data_scanned/scanned_04.easyocr.doctags.txt
View File

@ -1,14 +0,0 @@
<document>
<paragraph><location><page_1><loc_15><loc_87><loc_88><loc_92></location>Ihe parameters used for all theoretical graphs were, unless stated otherwise:</paragraph>
<paragraph><location><page_1><loc_42><loc_83><loc_65><loc_85></location>(h) 9 0.1 mm Gap</paragraph>
<paragraph><location><page_1><loc_31><loc_80><loc_67><loc_82></location>Wire Radius (R) 0.825 mm</paragraph>
<paragraph><location><page_1><loc_23><loc_77><loc_69><loc_79></location>Wire Yield Stress (x) 245 MN/ 0<2</paragraph>
<paragraph><location><page_1><loc_31><loc_74><loc_63><loc_76></location>Iube Iength (L) 60 mm</paragraph>
<paragraph><location><page_1><loc_24><loc_71><loc_67><loc_74></location>50 Js/n2</paragraph>
<paragraph><location><page_1><loc_23><loc_71><loc_51><loc_73></location>Polymer Viscosity (n)</paragraph>
<paragraph><location><page_1><loc_19><loc_68><loc_75><loc_71></location>Critical Shear Stress (Ic) = 0.6 x N/a2 706</paragraph>
<paragraph><location><page_1><loc_14><loc_55><loc_95><loc_66></location>Ihese values can be considered typical and were also used when estimating the experinental error (see Chapter 6.01) _ Io aid comparison between the effects of changing various parameters the sane graph axes were used throughout.</paragraph>
<paragraph><location><page_1><loc_14><loc_46><loc_87><loc_54></location>Assumptions were mnade, further to those already des cribed in Chapter II, when calculating the theoretical results explained above.</paragraph>
<paragraph><location><page_1><loc_14><loc_23><loc_89><loc_46></location>Firstly the viscosity was assuned to be constant when in fact it is reduçed by shear stress and increased by pressure. Ihe effect of shear stress on viscosity can be measured (see 8) and a mean value may be assumed . It is impossible, however, to neasure the combined effect of both high shear stress and high pressure although there is reason to believe they may, to some extent, cancel each other out. Ihis problem is discussed more fully later. Fig.</paragraph>
<paragraph><location><page_1><loc_14><loc_6><loc_95><loc_22></location>It was also assuned that the transition from no slip to total takes place instantaneously at the velocity Vc when in practice it is more likely to take place over a small range of velocities. Ihis gives the theoretical graphs a sharp peak where as the effect of the slip taking place over a range of velocity would be to "round-off" this peak. slip</paragraph>
</document>

1
tests/data_scanned/scanned_04.easyocr.json
View File

@ -1 +0,0 @@
{"_name": "", "type": "pdf-document", "description": {"logs": []}, "file-info": {"filename": "scanned_04.pdf", "document-hash": "551a2a04bd4b1e1ae357cc78600f5500f60ae99b7a31cf47982a69ee0a682516", "#-pages": 1, "page-hashes": [{"hash": "d2932ec29e3c749a6436c7d1fcbba528f99844f68bebc082f862cc3a88727e29", "model": "default", "page": 1}]}, "main-text": [{"prov": [{"bbox": [535.00927734375, 1651.6666259765625, 626.1068115234375, 1677.0], "page": 1, "span": [0, 4]}], "text": "42 9", "type": "page-header", "name": "Page-header"}, {"prov": [{"bbox": [172.0, 1511.0, 1030.6666259765625, 1589.6666259765625], "page": 1, "span": [0, 77]}], "text": "Ihe parameters used for all theoretical graphs were, unless stated otherwise:", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [489.3140563964844, 1429.0186767578125, 763.3333129882812, 1466.9813232421875], "page": 1, "span": [0, 16]}], "text": "(h) 9 0.1 mm Gap", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [366.6666564941406, 1383.3333740234375, 791.7745361328125, 1417.0], "page": 1, "span": [0, 24]}], "text": "Wire Radius (R) 0.825 mm", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [275.6666564941406, 1333.0, 813.9666748046875, 1371.9569091796875], "page": 1, "span": [0, 33]}], "text": "Wire Yield Stress (x) 245 MN/ 0<2", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [366.3333435058594, 1279.6666259765625, 747.0, 1316.3333740234375], "page": 1, "span": [0, 21]}], "text": "Iube Iength (L) 60 mm", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [278.57855224609375, 1231.0, 793.6666870117188, 1275.0], "page": 1, "span": [0, 8]}], "text": "50 Js/n2", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [275.3333435058594, 1228.0, 595.6666870117188, 1266.6666259765625], "page": 1, "span": [0, 21]}], "text": "Polymer Viscosity (n)", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [218.0, 1182.3333740234375, 882.6666870117188, 1224.6666259765625], "page": 1, "span": [0, 43]}], "text": "Critical Shear Stress (Ic) = 0.6 x N/a2 706", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [168.0, 951.0, 1115.502685546875, 1137.6666259765625], "page": 1, "span": [0, 228]}], "text": "Ihese values can be considered typical and were also used when estimating the experinental error (see Chapter 6.01) _ Io aid comparison between the effects of changing various parameters the sane graph axes were used throughout.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [169.6666717529297, 801.6666870117188, 1028.3333740234375, 940.3333129882812], "page": 1, "span": [0, 132]}], "text": "Assumptions were mnade, further to those already des cribed in Chapter II, when calculating the theoretical results explained above.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [167.0, 398.6666564941406, 1045.3333740234375, 787.6666870117188], "page": 1, "span": [0, 452]}], "text": "Firstly the viscosity was assuned to be constant when in fact it is redu\u00e7ed by shear stress and increased by pressure. Ihe effect of shear stress on viscosity can be measured (see 8) and a mean value may be assumed . It is impossible, however, to neasure the combined effect of both high shear stress and high pressure although there is reason to believe they may, to some extent, cancel each other out. Ihis problem is discussed more fully later. Fig.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [164.3333282470703, 99.0, 1119.043212890625, 383.6666564941406], "page": 1, "span": [0, 348]}], "text": "It was also assuned that the transition from no slip to total takes place instantaneously at the velocity Vc when in practice it is more likely to take place over a small range of velocities. Ihis gives the theoretical graphs a sharp peak where as the effect of the slip taking place over a range of velocity would be to \"round-off\" this peak. slip", "type": "paragraph", "name": "Text"}], "figures": [], "tables": [], "equations": [], "footnotes": [], "page-dimensions": [{"height": 1729.0, "page": 1, "width": 1177.0}], "page-footers": [], "page-headers": []}

23
tests/data_scanned/scanned_04.easyocr.md
View File

@ -1,23 +0,0 @@
Ihe parameters used for all theoretical graphs were, unless stated otherwise:
(h) 9 0.1 mm Gap
Wire Radius (R) 0.825 mm
Wire Yield Stress (x) 245 MN/ 0<2
Iube Iength (L) 60 mm
50 Js/n2
Polymer Viscosity (n)
Critical Shear Stress (Ic) = 0.6 x N/a2 706
Ihese values can be considered typical and were also used when estimating the experinental error (see Chapter 6.01) _ Io aid comparison between the effects of changing various parameters the sane graph axes were used throughout.
Assumptions were mnade, further to those already des cribed in Chapter II, when calculating the theoretical results explained above.
Firstly the viscosity was assuned to be constant when in fact it is reduçed by shear stress and increased by pressure. Ihe effect of shear stress on viscosity can be measured (see 8) and a mean value may be assumed . It is impossible, however, to neasure the combined effect of both high shear stress and high pressure although there is reason to believe they may, to some extent, cancel each other out. Ihis problem is discussed more fully later. Fig.
It was also assuned that the transition from no slip to total takes place instantaneously at the velocity Vc when in practice it is more likely to take place over a small range of velocities. Ihis gives the theoretical graphs a sharp peak where as the effect of the slip taking place over a range of velocity would be to "round-off" this peak. slip

1
tests/data_scanned/scanned_04.easyocr.pages.json
View File

File diff suppressed because one or more lines are too long

BIN
tests/data_scanned/scanned_04.pdf
View File

Binary file not shown.

BIN
tests/data_scanned/scanned_04.png
View File

Binary file not shown.
Side by Side

Before

Width:  |  Height:  |  Size: 112 KiB

13
tests/data_scanned/scanned_04.tesserocr.doctags.txt
View File

@ -1,13 +0,0 @@
<document>
<paragraph><location><page_1><loc_15><loc_88><loc_88><loc_92></location>The parameters used for all theoretical graphs were, unless stated otherwise:</paragraph>
<paragraph><location><page_1><loc_42><loc_83><loc_65><loc_85></location>Gap (h) = O.1 mm</paragraph>
<paragraph><location><page_1><loc_31><loc_80><loc_67><loc_82></location>Wire Radius (R) = 0.825 mm</paragraph>
<paragraph><location><page_1><loc_24><loc_77><loc_68><loc_79></location>Wire Yield Stress (Y) = 245 MNym@</paragraph>
<paragraph><location><page_1><loc_31><loc_74><loc_63><loc_76></location>Tube Length (IL) = 60 mm</paragraph>
<paragraph><location><page_1><loc_24><loc_71><loc_67><loc_74></location>Polymer Viscosity (n) = 50 Ns /m*</paragraph>
<paragraph><location><page_1><loc_19><loc_69><loc_75><loc_71></location>Critical Shear Stress(Tc) = 0.6 x 10° N/we</paragraph>
<paragraph><location><page_1><loc_15><loc_55><loc_95><loc_66></location>These values can be considered 'typical' and were also used when estimating the experimental error (see Chapter 6.01). To aid comparison between the effects of changing various parameters the same graph axes were used throughout.</paragraph>
<paragraph><location><page_1><loc_15><loc_47><loc_87><loc_54></location>Assumptions were made, further to those already described in Chapter II, when calculating the theoretical results explained above.</paragraph>
<paragraph><location><page_1><loc_14><loc_23><loc_89><loc_45></location>Firstly the viscosity was assumed to be constant when in fact it is reduced by shear stress and increased by pressure. The effect of shear stress on viscosity can be measured (see Fig. 8) and a mean value may be assumed. It is impossible, however, to measure the combined effect of both high shear stress and high pressure although there is reason to believe they may, to some extent, cancel each other out. This problem is discussed more fully later.</paragraph>
<paragraph><location><page_1><loc_14><loc_6><loc_95><loc_22></location>It was also assumed that the transition from no slip to total slip takes place instantaneously at the velocity Vc when in practice it is more likely to take place over a small range of velocities. This gives the theoretical graphs a sharp peak where as the effect of the slip taking place over a range of velocity would be to round-off this peak.</paragraph>
</document>

1
tests/data_scanned/scanned_04.tesserocr.json
View File

@ -1 +0,0 @@
{"_name": "", "type": "pdf-document", "description": {"logs": []}, "file-info": {"filename": "scanned_04.pdf", "document-hash": "551a2a04bd4b1e1ae357cc78600f5500f60ae99b7a31cf47982a69ee0a682516", "#-pages": 1, "page-hashes": [{"hash": "d2932ec29e3c749a6436c7d1fcbba528f99844f68bebc082f862cc3a88727e29", "model": "default", "page": 1}]}, "main-text": [{"prov": [{"bbox": [535.00927734375, 1654.9765625, 626.1068115234375, 1675.8818359375], "page": 1, "span": [0, 6]}], "text": "~ 4d -", "type": "page-header", "name": "Page-header"}, {"prov": [{"bbox": [174.1952362060547, 1515.218994140625, 1034.0, 1586.716064453125], "page": 1, "span": [0, 77]}], "text": "The parameters used for all theoretical graphs were, unless stated otherwise:", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [491.0345153808594, 1434.818603515625, 762.4005126953125, 1464.32080078125], "page": 1, "span": [0, 16]}], "text": "Gap (h) = O.1 mm", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [369.16650390625, 1385.7452392578125, 791.7745361328125, 1413.839599609375], "page": 1, "span": [0, 26]}], "text": "Wire Radius (R) = 0.825 mm", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [278.53167724609375, 1335.5279541015625, 805.1238403320312, 1371.9569091796875], "page": 1, "span": [0, 33]}], "text": "Wire Yield Stress (Y) = 245 MNym@", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [369.3974914550781, 1283.9993896484375, 745.410400390625, 1313.46240234375], "page": 1, "span": [0, 24]}], "text": "Tube Length (IL) = 60 mm", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [278.57855224609375, 1233.7530517578125, 790.2716064453125, 1272.5389404296875], "page": 1, "span": [0, 33]}], "text": "Polymer Viscosity (n) = 50 Ns /m*", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [219.99317932128906, 1184.6124267578125, 878.9542846679688, 1221.7412109375], "page": 1, "span": [0, 42]}], "text": "Critical Shear Stress(Tc) = 0.6 x 10\u00b0 N/we", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [171.25897216796875, 958.4537353515625, 1115.502685546875, 1135.67138671875], "page": 1, "span": [0, 229]}], "text": "These values can be considered 'typical' and were also used when estimating the experimental error (see Chapter 6.01). To aid comparison between the effects of changing various parameters the same graph axes were used throughout.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [172.10269165039062, 808.0, 1027.4852294921875, 933.71337890625], "page": 1, "span": [0, 130]}], "text": "Assumptions were made, further to those already described in Chapter II, when calculating the theoretical results explained above.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [169.87054443359375, 405.0, 1044.957275390625, 782.91357421875], "page": 1, "span": [0, 451]}], "text": "Firstly the viscosity was assumed to be constant when in fact it is reduced by shear stress and increased by pressure. The effect of shear stress on viscosity can be measured (see Fig. 8) and a mean value may be assumed. It is impossible, however, to measure the combined effect of both high shear stress and high pressure although there is reason to believe they may, to some extent, cancel each other out. This problem is discussed more fully later.", "type": "paragraph", "name": "Text"}, {"prov": [{"bbox": [169.0041046142578, 102.8302001953125, 1119.043212890625, 381.3938903808594], "page": 1, "span": [0, 346]}], "text": "It was also assumed that the transition from no slip to total slip takes place instantaneously at the velocity Vc when in practice it is more likely to take place over a small range of velocities. This gives the theoretical graphs a sharp peak where as the effect of the slip taking place over a range of velocity would be to round-off this peak.", "type": "paragraph", "name": "Text"}], "figures": [], "tables": [], "equations": [], "footnotes": [], "page-dimensions": [{"height": 1729.0, "page": 1, "width": 1177.0}], "page-footers": [], "page-headers": []}

21
tests/data_scanned/scanned_04.tesserocr.md
View File

@ -1,21 +0,0 @@
The parameters used for all theoretical graphs were, unless stated otherwise:
Gap (h) = O.1 mm
Wire Radius (R) = 0.825 mm
Wire Yield Stress (Y) = 245 MNym@
Tube Length (IL) = 60 mm
Polymer Viscosity (n) = 50 Ns /m*
Critical Shear Stress(Tc) = 0.6 x 10° N/we
These values can be considered 'typical' and were also used when estimating the experimental error (see Chapter 6.01). To aid comparison between the effects of changing various parameters the same graph axes were used throughout.
Assumptions were made, further to those already described in Chapter II, when calculating the theoretical results explained above.
Firstly the viscosity was assumed to be constant when in fact it is reduced by shear stress and increased by pressure. The effect of shear stress on viscosity can be measured (see Fig. 8) and a mean value may be assumed. It is impossible, however, to measure the combined effect of both high shear stress and high pressure although there is reason to believe they may, to some extent, cancel each other out. This problem is discussed more fully later.
It was also assumed that the transition from no slip to total slip takes place instantaneously at the velocity Vc when in practice it is more likely to take place over a small range of velocities. This gives the theoretical graphs a sharp peak where as the effect of the slip taking place over a range of velocity would be to round-off this peak.

1
tests/data_scanned/scanned_04.tesserocr.pages.json
View File

File diff suppressed because one or more lines are too long