mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-12-08 20:58:11 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: Support AsciiDoc and Markdown input format (#168)

Browse Source 
* updated the base-model and added the asciidoc_backend

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* updated the asciidoc backend

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* Ensure all models work only on valid pages (#158)

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* ci: run ci also on forks (#160)


---------

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
Signed-off-by: Michele Dolfi <97102151+dolfim-ibm@users.noreply.github.com>

* fix: fix legacy doc ref (#162)

Signed-off-by: Panos Vagenas <35837085+vagenas@users.noreply.github.com>

* docs: typo fix (#155)

* Docs: Typo fix

- Corrected spelling of invidual to automatic

Signed-off-by: ABHISHEK FADAKE <31249309+fadkeabhi@users.noreply.github.com>

* add synchronize event for forks

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

---------

Signed-off-by: ABHISHEK FADAKE <31249309+fadkeabhi@users.noreply.github.com>
Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
Co-authored-by: Michele Dolfi <dol@zurich.ibm.com>

* feat: add coverage_threshold to skip OCR for small images (#161)

* feat: add coverage_threshold to skip OCR for small images

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* filter individual boxes

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* rename option

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

---------

Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>

* chore: bump version to 2.1.0 [skip ci]

* adding tests for asciidocs

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* first working asciidoc parser

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* reformatted the code

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* fixed the mypy

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* adding test_02.asciidoc

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* Drafting Markdown backend via Marko library

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* work in progress on MD backend

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* md_backend produces docling document with headers, paragraphs, lists

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Improvements in md parsing

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Detecting and assembling tables in markdown in temporary buffers

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Added initial docling table support to md_backend

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Cleaned code, improved logging for MD

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Fixes MyPy requirements, and rest of pre-commit

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Fixed example run_md, added origin info to md_backend

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* working on asciidocs, struggling with ImageRef

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* able to parse the captions and image uri's

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* fixed the mypy

Signed-off-by: Peter Staar <taa@zurich.ibm.com>

* Update all backends with proper filename in DocumentOrigin

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Update to docling-core v2.1.0

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Fixes for MD Backend, to avoid duplicated text inserts into docling doc

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Fix styling

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>

* Added support for code blocks and fenced code in MD

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* cleaned prints

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Added proper processing of in-line textual elements for MD backend

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Fixed issues with duplicated paragraphs and incorrect lists in pptx

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

* Fixed issue with group ordeering in pptx backend, added gebug log into run with formats

Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>

---------

Signed-off-by: Peter Staar <taa@zurich.ibm.com>
Signed-off-by: Christoph Auer <cau@zurich.ibm.com>
Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
Signed-off-by: Michele Dolfi <97102151+dolfim-ibm@users.noreply.github.com>
Signed-off-by: Panos Vagenas <35837085+vagenas@users.noreply.github.com>
Signed-off-by: ABHISHEK FADAKE <31249309+fadkeabhi@users.noreply.github.com>
Signed-off-by: Maksym Lysak <mly@zurich.ibm.com>
Co-authored-by: Peter Staar <taa@zurich.ibm.com>
Co-authored-by: Michele Dolfi <97102151+dolfim-ibm@users.noreply.github.com>
Co-authored-by: Panos Vagenas <35837085+vagenas@users.noreply.github.com>
Co-authored-by: ABHISHEK FADAKE <31249309+fadkeabhi@users.noreply.github.com>
Co-authored-by: Michele Dolfi <dol@zurich.ibm.com>
Co-authored-by: github-actions[bot] <github-actions[bot]@users.noreply.github.com>
Co-authored-by: Maksym Lysak <mly@zurich.ibm.com>
This commit is contained in:
Christoph Auer
2024-10-23 16:14:26 +02:00
committed by GitHub
parent 3496b4838f
commit 3023f18ba0
52 changed files with 3731 additions and 3517 deletions
Download Patch File Download Diff File Expand all files Collapse all files

206
tests/data/groundtruth/docling_v2/2203.01017v2.md
View File

@@ -14,8 +14,6 @@ The occurrence of tables in documents is ubiquitous. They often summarise quanti
Tables organize valuable content in a concise and compact representation. This content is extremely valuable for systems such as search engines, Knowledge Graph's, etc, since they enhance their predictive capabilities. Unfortunately, tables come in a large variety of shapes and sizes. Furthermore, they can have complex column/row-header configurations, multiline rows, different variety of separation lines, missing entries, etc. As such, the correct identification of the table-structure from an image is a nontrivial task. In this paper, we present a new table-structure identification model. The latter improves the latest end-toend deep learning model (i.e. encoder-dual-decoder from PubTabNet) in two significant ways. First, we introduce a new object detection decoder for table-cells. In this way, we can obtain the content of the table-cells from programmatic PDF's directly from the PDF source and avoid the training of the custom OCR decoders. This architectural change leads to more accurate table-content extraction and allows us to tackle non-english tables. Second, we replace the LSTM decoders with transformer based decoders. This upgrade improves significantly the previous state-of-the-art tree-editing-distance-score (TEDS) from 91% to 98.5% on simple tables and from 88.7% to 95% on complex tables.
b. Red-annotation of bounding boxes, Blue-predictions by TableFormer
<!-- image -->
@@ -26,7 +24,6 @@ Structure predicted by TableFormer:
Figure 1: Picture of a table with subtle, complex features such as (1) multi-column headers, (2) cell with multi-row text and (3) cells with no content. Image from PubTabNet evaluation set, filename: 'PMC2944238 004 02'.
| 0 | 1 | 1 | 2 1 | 2 1 | |
|-----|-----|-----|-------|-------|----|
| 3 | 4 | 5 3 | 6 | 7 | |
@@ -46,13 +43,10 @@ In this paper, we want to address these weaknesses and present a robust table-st
To meet the design criteria listed above, we developed a new model called TableFormer and a synthetically generated table structure dataset called SynthTabNet $^{1}$. In particular, our contributions in this work can be summarised as follows:
· We propose TableFormer , a transformer based model that predicts tables structure and bounding boxes for the table content simultaneously in an end-to-end approach.
· Across all benchmark datasets TableFormer significantly outperforms existing state-of-the-art metrics, while being much more efficient in training and inference to existing works.
· We present SynthTabNet a synthetically generated dataset, with various appearance styles and complexity.
· An augmented dataset based on PubTabNet [37], FinTabNet [36], and TableBank [17] with generated ground-truth for reproducibility.
- · We propose TableFormer , a transformer based model that predicts tables structure and bounding boxes for the table content simultaneously in an end-to-end approach.
- · Across all benchmark datasets TableFormer significantly outperforms existing state-of-the-art metrics, while being much more efficient in training and inference to existing works.
- · We present SynthTabNet a synthetically generated dataset, with various appearance styles and complexity.
- · An augmented dataset based on PubTabNet [37], FinTabNet [36], and TableBank [17] with generated ground-truth for reproducibility.
The paper is structured as follows. In Sec. 2, we give a brief overview of the current state-of-the-art. In Sec. 3, we describe the datasets on which we train. In Sec. 4, we introduce the TableFormer model-architecture and describe
@@ -80,7 +74,6 @@ We rely on large-scale datasets such as PubTabNet [37], FinTabNet [36], and Tabl
Figure 2: Distribution of the tables across different table dimensions in PubTabNet + FinTabNet datasets
<!-- image -->
balance in the previous datasets.
@@ -101,7 +94,6 @@ In this regard, we have prepared four synthetic datasets, each one containing 15
Table 1: Both "Combined-Tabnet" and "CombinedTabnet" are variations of the following: (*) The CombinedTabnet dataset is the processed combination of PubTabNet and Fintabnet. (**) The combined dataset is the processed combination of PubTabNet, Fintabnet and TableBank.
| | Tags | Bbox | Size | Format |
|--------------------|--------|--------|--------|----------|
| PubTabNet | 3 | 3 | 509k | PNG |
@@ -127,12 +119,10 @@ CNN Backbone Network. A ResNet-18 CNN is the backbone that receives the table im
Figure 3: TableFormer takes in an image of the PDF and creates bounding box and HTML structure predictions that are synchronized. The bounding boxes grabs the content from the PDF and inserts it in the structure.
<!-- image -->
Figure 4: Given an input image of a table, the Encoder produces fixed-length features that represent the input image. The features are then passed to both the Structure Decoder and Cell BBox Decoder . During training, the Structure Decoder receives 'tokenized tags' of the HTML code that represent the table structure. Afterwards, a transformer encoder and decoder architecture is employed to produce features that are received by a linear layer, and the Cell BBox Decoder. The linear layer is applied to the features to predict the tags. Simultaneously, the Cell BBox Decoder selects features referring to the data cells (' < td > ', ' < ') and passes them through an attention network, an MLP, and a linear layer to predict the bounding boxes.
<!-- image -->
forming classification, and adding an adaptive pooling layer of size 28*28. ResNet by default downsamples the image resolution by 32 and then the encoded image is provided to both the Structure Decoder , and Cell BBox Decoder .
@@ -195,7 +185,6 @@ Structure. As shown in Tab. 2, TableFormer outperforms all SOTA methods across d
Table 2: Structure results on PubTabNet (PTN), FinTabNet (FTN), TableBank (TB) and SynthTabNet (STN).
| Model | Dataset | Simple | TEDS Complex | All |
|-------------|-----------|----------|----------------|-------|
| EDD | PTN | 91.1 | 88.7 | 89.9 |
@@ -217,7 +206,6 @@ our Cell BBox Decoder accuracy for cells with a class label of 'content' only us
Table 3: Cell Bounding Box detection results on PubTabNet, and FinTabNet. PP: Post-processing.
| Model | Dataset | mAP | mAP (PP) |
|-------------|-------------|-------|------------|
| EDD+BBox | PubTabNet | 79.2 | 82.7 |
@@ -228,7 +216,6 @@ Cell Content. In this section, we evaluate the entire pipeline of recovering a t
Table 4: Results of structure with content retrieved using cell detection on PubTabNet. In all cases the input is PDF documents with cropped tables.
| Model | Simple | TEDS Complex | All |
|-------------|----------|----------------|-------|
| Tabula | 78 | 57.8 | 67.9 |
@@ -264,7 +251,6 @@ Structure predicted by TableFormer, with superimposed matched PDF cell text:
Text is aligned to match original for ease of viewing
| | Shares (in millions) | Shares (in millions) | Weighted Average Grant Date Fair Value | Weighted Average Grant Date Fair Value |
|--------------------------|------------------------|------------------------|------------------------------------------|------------------------------------------|
| | RS U s | PSUs | RSUs | PSUs |
@@ -276,14 +262,12 @@ Text is aligned to match original for ease of viewing
Figure 5: One of the benefits of TableFormer is that it is language agnostic, as an example, the left part of the illustration demonstrates TableFormer predictions on previously unseen language (Japanese). Additionally, we see that TableFormer is robust to variability in style and content, right side of the illustration shows the example of the TableFormer prediction from the FinTabNet dataset.
<!-- image -->
<!-- image -->
Figure 6: An example of TableFormer predictions (bounding boxes and structure) from generated SynthTabNet table.
<!-- image -->
## 5.5. Qualitative Analysis
@@ -296,89 +280,54 @@ In this paper, we presented TableFormer an end-to-end transformer based approach
## References
[1] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-
- [1] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-
<!-- image -->
end object detection with transformers. In Andrea Vedaldi, Horst Bischof, Thomas Brox, and Jan-Michael Frahm, editors, Computer Vision - ECCV 2020 , pages 213-229, Cham, 2020. Springer International Publishing. 5
- end object detection with transformers. In Andrea Vedaldi, Horst Bischof, Thomas Brox, and Jan-Michael Frahm, editors, Computer Vision - ECCV 2020 , pages 213-229, Cham, 2020. Springer International Publishing. 5
- [2] Zewen Chi, Heyan Huang, Heng-Da Xu, Houjin Yu, Wanxuan Yin, and Xian-Ling Mao. Complicated table structure recognition. arXiv preprint arXiv:1908.04729 , 2019. 3
- [3] Bertrand Couasnon and Aurelie Lemaitre. Recognition of Tables and Forms , pages 647-677. Springer London, London, 2014. 2
- [4] Herv'e D'ejean, Jean-Luc Meunier, Liangcai Gao, Yilun Huang, Yu Fang, Florian Kleber, and Eva-Maria Lang. ICDAR 2019 Competition on Table Detection and Recognition (cTDaR), Apr. 2019. http://sac.founderit.com/. 2
- [5] Basilios Gatos, Dimitrios Danatsas, Ioannis Pratikakis, and Stavros J Perantonis. Automatic table detection in document images. In International Conference on Pattern Recognition and Image Analysis , pages 609-618. Springer, 2005. 2
- [6] Max Gobel, Tamir Hassan, Ermelinda Oro, and Giorgio Orsi. Icdar 2013 table competition. In 2013 12th International Conference on Document Analysis and Recognition , pages 1449-1453, 2013. 2
- [7] EA Green and M Krishnamoorthy. Recognition of tables using table grammars. procs. In Symposium on Document Analysis and Recognition (SDAIR'95) , pages 261-277. 2
- [8] Khurram Azeem Hashmi, Alain Pagani, Marcus Liwicki, Didier Stricker, and Muhammad Zeshan Afzal. Castabdetectors: Cascade network for table detection in document images with recursive feature pyramid and switchable atrous convolution. Journal of Imaging , 7(10), 2021. 1
- [9] Kaiming He, Georgia Gkioxari, Piotr Dollar, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE International Conference on Computer Vision (ICCV) , Oct 2017. 1
- [10] Yelin He, X. Qi, Jiaquan Ye, Peng Gao, Yihao Chen, Bingcong Li, Xin Tang, and Rong Xiao. Pingan-vcgroup's solution for icdar 2021 competition on scientific table image recognition to latex. ArXiv , abs/2105.01846, 2021. 2
- [11] Jianying Hu, Ramanujan S Kashi, Daniel P Lopresti, and Gordon Wilfong. Medium-independent table detection. In Document Recognition and Retrieval VII , volume 3967, pages 291-302. International Society for Optics and Photonics, 1999. 2
- [12] Matthew Hurst. A constraint-based approach to table structure derivation. In Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 2 , ICDAR '03, page 911, USA, 2003. IEEE Computer Society. 2
- [13] Thotreingam Kasar, Philippine Barlas, Sebastien Adam, Cl'ement Chatelain, and Thierry Paquet. Learning to detect tables in scanned document images using line information. In 2013 12th International Conference on Document Analysis and Recognition , pages 1185-1189. IEEE, 2013. 2
- [14] Pratik Kayal, Mrinal Anand, Harsh Desai, and Mayank Singh. Icdar 2021 competition on scientific table image recognition to latex, 2021. 2
- [15] Harold W Kuhn. The hungarian method for the assignment problem. Naval research logistics quarterly , 2(1-2):83-97, 1955. 6
[2] Zewen Chi, Heyan Huang, Heng-Da Xu, Houjin Yu, Wanxuan Yin, and Xian-Ling Mao. Complicated table structure recognition. arXiv preprint arXiv:1908.04729 , 2019. 3
[3] Bertrand Couasnon and Aurelie Lemaitre. Recognition of Tables and Forms , pages 647-677. Springer London, London, 2014. 2
[4] Herv'e D'ejean, Jean-Luc Meunier, Liangcai Gao, Yilun Huang, Yu Fang, Florian Kleber, and Eva-Maria Lang. ICDAR 2019 Competition on Table Detection and Recognition (cTDaR), Apr. 2019. http://sac.founderit.com/. 2
[5] Basilios Gatos, Dimitrios Danatsas, Ioannis Pratikakis, and Stavros J Perantonis. Automatic table detection in document images. In International Conference on Pattern Recognition and Image Analysis , pages 609-618. Springer, 2005. 2
[6] Max Gobel, Tamir Hassan, Ermelinda Oro, and Giorgio Orsi. Icdar 2013 table competition. In 2013 12th International Conference on Document Analysis and Recognition , pages 1449-1453, 2013. 2
[7] EA Green and M Krishnamoorthy. Recognition of tables using table grammars. procs. In Symposium on Document Analysis and Recognition (SDAIR'95) , pages 261-277. 2
[8] Khurram Azeem Hashmi, Alain Pagani, Marcus Liwicki, Didier Stricker, and Muhammad Zeshan Afzal. Castabdetectors: Cascade network for table detection in document images with recursive feature pyramid and switchable atrous convolution. Journal of Imaging , 7(10), 2021. 1
[9] Kaiming He, Georgia Gkioxari, Piotr Dollar, and Ross Girshick. Mask r-cnn. In Proceedings of the IEEE International Conference on Computer Vision (ICCV) , Oct 2017. 1
[10] Yelin He, X. Qi, Jiaquan Ye, Peng Gao, Yihao Chen, Bingcong Li, Xin Tang, and Rong Xiao. Pingan-vcgroup's solution for icdar 2021 competition on scientific table image recognition to latex. ArXiv , abs/2105.01846, 2021. 2
[11] Jianying Hu, Ramanujan S Kashi, Daniel P Lopresti, and Gordon Wilfong. Medium-independent table detection. In Document Recognition and Retrieval VII , volume 3967, pages 291-302. International Society for Optics and Photonics, 1999. 2
[12] Matthew Hurst. A constraint-based approach to table structure derivation. In Proceedings of the Seventh International Conference on Document Analysis and Recognition - Volume 2 , ICDAR '03, page 911, USA, 2003. IEEE Computer Society. 2
[13] Thotreingam Kasar, Philippine Barlas, Sebastien Adam, Cl'ement Chatelain, and Thierry Paquet. Learning to detect tables in scanned document images using line information. In 2013 12th International Conference on Document Analysis and Recognition , pages 1185-1189. IEEE, 2013. 2
[14] Pratik Kayal, Mrinal Anand, Harsh Desai, and Mayank Singh. Icdar 2021 competition on scientific table image recognition to latex, 2021. 2
[15] Harold W Kuhn. The hungarian method for the assignment problem. Naval research logistics quarterly , 2(1-2):83-97, 1955. 6
[16] Girish Kulkarni, Visruth Premraj, Vicente Ordonez, Sagnik Dhar, Siming Li, Yejin Choi, Alexander C. Berg, and Tamara L. Berg. Babytalk: Understanding and generating simple image descriptions. IEEE Transactions on Pattern Analysis and Machine Intelligence , 35(12):2891-2903, 2013. 4
[17] Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou, and Zhoujun Li. Tablebank: A benchmark dataset for table detection and recognition, 2019. 2, 3
[18] Yiren Li, Zheng Huang, Junchi Yan, Yi Zhou, Fan Ye, and Xianhui Liu. Gfte: Graph-based financial table extraction. In Alberto Del Bimbo, Rita Cucchiara, Stan Sclaroff, Giovanni Maria Farinella, Tao Mei, Marco Bertini, Hugo Jair Escalante, and Roberto Vezzani, editors, Pattern Recognition. ICPR International Workshops and Challenges , pages 644-658, Cham, 2021. Springer International Publishing. 2, 3
[19] Nikolaos Livathinos, Cesar Berrospi, Maksym Lysak, Viktor Kuropiatnyk, Ahmed Nassar, Andre Carvalho, Michele Dolfi, Christoph Auer, Kasper Dinkla, and Peter Staar. Robust pdf document conversion using recurrent neural networks. Proceedings of the AAAI Conference on Artificial Intelligence , 35(17):15137-15145, May 2021. 1
[20] Rujiao Long, Wen Wang, Nan Xue, Feiyu Gao, Zhibo Yang, Yongpan Wang, and Gui-Song Xia. Parsing table structures in the wild. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 944-952, 2021. 2
[21] Shubham Singh Paliwal, D Vishwanath, Rohit Rahul, Monika Sharma, and Lovekesh Vig. Tablenet: Deep learning model for end-to-end table detection and tabular data extraction from scanned document images. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 128-133. IEEE, 2019. 1
[22] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and Soumith Chintala. Pytorch: An imperative style, high-performance deep learning library. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alch'e-Buc, E. Fox, and R. Garnett, editors, Advances in Neural Information Processing Systems 32 , pages 8024-8035. Curran Associates, Inc., 2019. 6
[23] Devashish Prasad, Ayan Gadpal, Kshitij Kapadni, Manish Visave, and Kavita Sultanpure. Cascadetabnet: An approach for end to end table detection and structure recognition from image-based documents. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops , pages 572-573, 2020. 1
[24] Shah Rukh Qasim, Hassan Mahmood, and Faisal Shafait. Rethinking table recognition using graph neural networks. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 142-147. IEEE, 2019. 3
[25] Hamid Rezatofighi, Nathan Tsoi, JunYoung Gwak, Amir Sadeghian, Ian Reid, and Silvio Savarese. Generalized intersection over union: A metric and a loss for bounding box regression. In Proceedings of the IEEE/CVF Conference on
- [16] Girish Kulkarni, Visruth Premraj, Vicente Ordonez, Sagnik Dhar, Siming Li, Yejin Choi, Alexander C. Berg, and Tamara L. Berg. Babytalk: Understanding and generating simple image descriptions. IEEE Transactions on Pattern Analysis and Machine Intelligence , 35(12):2891-2903, 2013. 4
- [17] Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, Ming Zhou, and Zhoujun Li. Tablebank: A benchmark dataset for table detection and recognition, 2019. 2, 3
- [18] Yiren Li, Zheng Huang, Junchi Yan, Yi Zhou, Fan Ye, and Xianhui Liu. Gfte: Graph-based financial table extraction. In Alberto Del Bimbo, Rita Cucchiara, Stan Sclaroff, Giovanni Maria Farinella, Tao Mei, Marco Bertini, Hugo Jair Escalante, and Roberto Vezzani, editors, Pattern Recognition. ICPR International Workshops and Challenges , pages 644-658, Cham, 2021. Springer International Publishing. 2, 3
- [19] Nikolaos Livathinos, Cesar Berrospi, Maksym Lysak, Viktor Kuropiatnyk, Ahmed Nassar, Andre Carvalho, Michele Dolfi, Christoph Auer, Kasper Dinkla, and Peter Staar. Robust pdf document conversion using recurrent neural networks. Proceedings of the AAAI Conference on Artificial Intelligence , 35(17):15137-15145, May 2021. 1
- [20] Rujiao Long, Wen Wang, Nan Xue, Feiyu Gao, Zhibo Yang, Yongpan Wang, and Gui-Song Xia. Parsing table structures in the wild. In Proceedings of the IEEE/CVF International Conference on Computer Vision , pages 944-952, 2021. 2
- [21] Shubham Singh Paliwal, D Vishwanath, Rohit Rahul, Monika Sharma, and Lovekesh Vig. Tablenet: Deep learning model for end-to-end table detection and tabular data extraction from scanned document images. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 128-133. IEEE, 2019. 1
- [22] Adam Paszke, Sam Gross, Francisco Massa, Adam Lerer, James Bradbury, Gregory Chanan, Trevor Killeen, Zeming Lin, Natalia Gimelshein, Luca Antiga, Alban Desmaison, Andreas Kopf, Edward Yang, Zachary DeVito, Martin Raison, Alykhan Tejani, Sasank Chilamkurthy, Benoit Steiner, Lu Fang, Junjie Bai, and Soumith Chintala. Pytorch: An imperative style, high-performance deep learning library. In H. Wallach, H. Larochelle, A. Beygelzimer, F. d'Alch'e-Buc, E. Fox, and R. Garnett, editors, Advances in Neural Information Processing Systems 32 , pages 8024-8035. Curran Associates, Inc., 2019. 6
- [23] Devashish Prasad, Ayan Gadpal, Kshitij Kapadni, Manish Visave, and Kavita Sultanpure. Cascadetabnet: An approach for end to end table detection and structure recognition from image-based documents. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition Workshops , pages 572-573, 2020. 1
- [24] Shah Rukh Qasim, Hassan Mahmood, and Faisal Shafait. Rethinking table recognition using graph neural networks. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 142-147. IEEE, 2019. 3
- [25] Hamid Rezatofighi, Nathan Tsoi, JunYoung Gwak, Amir Sadeghian, Ian Reid, and Silvio Savarese. Generalized intersection over union: A metric and a loss for bounding box regression. In Proceedings of the IEEE/CVF Conference on
Computer Vision and Pattern Recognition , pages 658-666, 2019. 6
[26] Sebastian Schreiber, Stefan Agne, Ivo Wolf, Andreas Dengel, and Sheraz Ahmed. Deepdesrt: Deep learning for detection and structure recognition of tables in document images. In 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR) , volume 01, pages 11621167, 2017. 1
- [26] Sebastian Schreiber, Stefan Agne, Ivo Wolf, Andreas Dengel, and Sheraz Ahmed. Deepdesrt: Deep learning for detection and structure recognition of tables in document images. In 2017 14th IAPR International Conference on Document Analysis and Recognition (ICDAR) , volume 01, pages 11621167, 2017. 1
- [27] Sebastian Schreiber, Stefan Agne, Ivo Wolf, Andreas Dengel, and Sheraz Ahmed. Deepdesrt: Deep learning for detection and structure recognition of tables in document images. In 2017 14th IAPR international conference on document analysis and recognition (ICDAR) , volume 1, pages 1162-1167. IEEE, 2017. 3
- [28] Faisal Shafait and Ray Smith. Table detection in heterogeneous documents. In Proceedings of the 9th IAPR International Workshop on Document Analysis Systems , pages 6572, 2010. 2
- [29] Shoaib Ahmed Siddiqui, Imran Ali Fateh, Syed Tahseen Raza Rizvi, Andreas Dengel, and Sheraz Ahmed. Deeptabstr: Deep learning based table structure recognition. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 1403-1409. IEEE, 2019. 3
- [30] Peter W J Staar, Michele Dolfi, Christoph Auer, and Costas Bekas. Corpus conversion service: A machine learning platform to ingest documents at scale. In Proceedings of the 24th ACM SIGKDD , KDD '18, pages 774-782, New York, NY, USA, 2018. ACM. 1
- [31] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin. Attention is all you need. In I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors, Advances in Neural Information Processing Systems 30 , pages 5998-6008. Curran Associates, Inc., 2017. 5
- [32] Oriol Vinyals, Alexander Toshev, Samy Bengio, and Dumitru Erhan. Show and tell: A neural image caption generator. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , June 2015. 2
- [33] Wenyuan Xue, Qingyong Li, and Dacheng Tao. Res2tim: reconstruct syntactic structures from table images. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 749-755. IEEE, 2019. 3
- [34] Wenyuan Xue, Baosheng Yu, Wen Wang, Dacheng Tao, and Qingyong Li. Tgrnet: A table graph reconstruction network for table structure recognition. arXiv preprint arXiv:2106.10598 , 2021. 3
- [35] Quanzeng You, Hailin Jin, Zhaowen Wang, Chen Fang, and Jiebo Luo. Image captioning with semantic attention. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 4651-4659, 2016. 4
- [36] Xinyi Zheng, Doug Burdick, Lucian Popa, Peter Zhong, and Nancy Xin Ru Wang. Global table extractor (gte): A framework for joint table identification and cell structure recognition using visual context. Winter Conference for Applications in Computer Vision (WACV) , 2021. 2, 3
- [37] Xu Zhong, Elaheh ShafieiBavani, and Antonio Jimeno Yepes. Image-based table recognition: Data, model,
[27] Sebastian Schreiber, Stefan Agne, Ivo Wolf, Andreas Dengel, and Sheraz Ahmed. Deepdesrt: Deep learning for detection and structure recognition of tables in document images. In 2017 14th IAPR international conference on document analysis and recognition (ICDAR) , volume 1, pages 1162-1167. IEEE, 2017. 3
[28] Faisal Shafait and Ray Smith. Table detection in heterogeneous documents. In Proceedings of the 9th IAPR International Workshop on Document Analysis Systems , pages 6572, 2010. 2
[29] Shoaib Ahmed Siddiqui, Imran Ali Fateh, Syed Tahseen Raza Rizvi, Andreas Dengel, and Sheraz Ahmed. Deeptabstr: Deep learning based table structure recognition. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 1403-1409. IEEE, 2019. 3
[30] Peter W J Staar, Michele Dolfi, Christoph Auer, and Costas Bekas. Corpus conversion service: A machine learning platform to ingest documents at scale. In Proceedings of the 24th ACM SIGKDD , KDD '18, pages 774-782, New York, NY, USA, 2018. ACM. 1
[31] Ashish Vaswani, Noam Shazeer, Niki Parmar, Jakob Uszkoreit, Llion Jones, Aidan N Gomez, Ł ukasz Kaiser, and Illia Polosukhin. Attention is all you need. In I. Guyon, U. V. Luxburg, S. Bengio, H. Wallach, R. Fergus, S. Vishwanathan, and R. Garnett, editors, Advances in Neural Information Processing Systems 30 , pages 5998-6008. Curran Associates, Inc., 2017. 5
[32] Oriol Vinyals, Alexander Toshev, Samy Bengio, and Dumitru Erhan. Show and tell: A neural image caption generator. In Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) , June 2015. 2
[33] Wenyuan Xue, Qingyong Li, and Dacheng Tao. Res2tim: reconstruct syntactic structures from table images. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 749-755. IEEE, 2019. 3
[34] Wenyuan Xue, Baosheng Yu, Wen Wang, Dacheng Tao, and Qingyong Li. Tgrnet: A table graph reconstruction network for table structure recognition. arXiv preprint arXiv:2106.10598 , 2021. 3
[35] Quanzeng You, Hailin Jin, Zhaowen Wang, Chen Fang, and Jiebo Luo. Image captioning with semantic attention. In Proceedings of the IEEE conference on computer vision and pattern recognition , pages 4651-4659, 2016. 4
[36] Xinyi Zheng, Doug Burdick, Lucian Popa, Peter Zhong, and Nancy Xin Ru Wang. Global table extractor (gte): A framework for joint table identification and cell structure recognition using visual context. Winter Conference for Applications in Computer Vision (WACV) , 2021. 2, 3
[37] Xu Zhong, Elaheh ShafieiBavani, and Antonio Jimeno Yepes. Image-based table recognition: Data, model,
and evaluation. In Andrea Vedaldi, Horst Bischof, Thomas Brox, and Jan-Michael Frahm, editors, Computer Vision ECCV 2020 , pages 564-580, Cham, 2020. Springer International Publishing. 2, 3, 7
[38] Xu Zhong, Jianbin Tang, and Antonio Jimeno Yepes. Publaynet: Largest dataset ever for document layout analysis. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 1015-1022, 2019. 1
- and evaluation. In Andrea Vedaldi, Horst Bischof, Thomas Brox, and Jan-Michael Frahm, editors, Computer Vision ECCV 2020 , pages 564-580, Cham, 2020. Springer International Publishing. 2, 3, 7
- [38] Xu Zhong, Jianbin Tang, and Antonio Jimeno Yepes. Publaynet: Largest dataset ever for document layout analysis. In 2019 International Conference on Document Analysis and Recognition (ICDAR) , pages 1015-1022, 2019. 1
## TableFormer: Table Structure Understanding with Transformers Supplementary Material
@@ -400,15 +349,11 @@ ances in regard to their size, structure, style and content. Every synthetic dat
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 (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.
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, a 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.
- 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.
- 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, a 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
@@ -416,52 +361,40 @@ Although TableFormer can predict the table structure and the bounding boxes for
Figure 7: Distribution of the tables across different dimensions per dataset. Simple vs complex tables per dataset and split, strict vs non strict html structures per dataset and table complexity, missing bboxes per dataset and table complexity.
<!-- image -->
· TableFormer output does not include the table cell content.
· There are occasional inaccuracies in the predictions of the bounding boxes.
- · TableFormer output does not include the table cell content.
- · There are occasional inaccuracies in the predictions of the bounding boxes.
However, it is possible to mitigate those limitations by combining the TableFormer predictions with the information already present inside a programmatic PDF document. More specifically, PDF documents can be seen as a sequence of PDF cells where each cell is described by its content and bounding box. If we are able to associate the PDF cells with the predicted table cells, we can directly link the PDF cell content to the table cell structure and use the PDF bounding boxes to correct misalignments in the predicted table cell bounding boxes.
Here is a step-by-step description of the prediction postprocessing:
1. Get the minimal grid dimensions - number of rows and columns for the predicted table structure. This represents the most granular grid for the underlying table structure.
2. Generate pair-wise matches between the bounding boxes of the PDF cells and the predicted cells. The Intersection Over Union (IOU) metric is used to evaluate the quality of the matches.
3. Use a carefully selected IOU threshold to designate the matches as "good" ones and "bad" ones.
3.a. If all IOU scores in a column are below the threshold, discard all predictions (structure and bounding boxes) for that column.
4. Find the best-fitting content alignment for the predicted cells with good IOU per each column. The alignment of the column can be identified by the following formula:
- 1. Get the minimal grid dimensions - number of rows and columns for the predicted table structure. This represents the most granular grid for the underlying table structure.
- 2. Generate pair-wise matches between the bounding boxes of the PDF cells and the predicted cells. The Intersection Over Union (IOU) metric is used to evaluate the quality of the matches.
- 3. Use a carefully selected IOU threshold to designate the matches as "good" ones and "bad" ones.
- 3.a. If all IOU scores in a column are below the threshold, discard all predictions (structure and bounding boxes) for that column.
- 4. Find the best-fitting content alignment for the predicted cells with good IOU per each column. The alignment of the column can be identified by the following formula:
alignment = arg min c { D$_{c}$ } D$_{c}$ = max { x$_{c}$ } - min { x$_{c}$ } (4)
where c is one of { left, centroid, right } and x$_{c}$ is the xcoordinate for the corresponding point.
5. Use the alignment computed in step 4, to compute the median x -coordinate for all table columns and the me-
- 5. Use the alignment computed in step 4, to compute the median x -coordinate for all table columns and the me-
dian cell size for all table cells. The usage of median during the computations, helps to eliminate outliers caused by occasional column spans which are usually wider than the normal.
6. Snap all cells with bad IOU to their corresponding median x -coordinates and cell sizes.
7. Generate a new set of pair-wise matches between the corrected bounding boxes and PDF cells. This time use a modified version of the IOU metric, where the area of the intersection between the predicted and PDF cells is divided by the PDF cell area. In case there are multiple matches for the same PDF cell, the prediction with the higher score is preferred. This covers the cases where the PDF cells are smaller than the area of predicted or corrected prediction cells.
8. In some rare occasions, we have noticed that TableFormer can confuse a single column as two. When the postprocessing steps are applied, this results with two predicted columns pointing to the same PDF column. In such case we must de-duplicate the columns according to highest total column intersection score.
9. Pick up the remaining orphan cells. There could be cases, when after applying all the previous post-processing steps, some PDF cells could still remain without any match to predicted cells. However, it is still possible to deduce the correct matching for an orphan PDF cell by mapping its bounding box on the geometry of the grid. This mapping decides if the content of the orphan cell will be appended to an already matched table cell, or a new table cell should be created to match with the orphan.
- 6. Snap all cells with bad IOU to their corresponding median x -coordinates and cell sizes.
- 7. Generate a new set of pair-wise matches between the corrected bounding boxes and PDF cells. This time use a modified version of the IOU metric, where the area of the intersection between the predicted and PDF cells is divided by the PDF cell area. In case there are multiple matches for the same PDF cell, the prediction with the higher score is preferred. This covers the cases where the PDF cells are smaller than the area of predicted or corrected prediction cells.
- 8. In some rare occasions, we have noticed that TableFormer can confuse a single column as two. When the postprocessing steps are applied, this results with two predicted columns pointing to the same PDF column. In such case we must de-duplicate the columns according to highest total column intersection score.
- 9. Pick up the remaining orphan cells. There could be cases, when after applying all the previous post-processing steps, some PDF cells could still remain without any match to predicted cells. However, it is still possible to deduce the correct matching for an orphan PDF cell by mapping its bounding box on the geometry of the grid. This mapping decides if the content of the orphan cell will be appended to an already matched table cell, or a new table cell should be created to match with the orphan.
9a. Compute the top and bottom boundary of the horizontal band for each grid row (min/max y coordinates per row).
9b. Intersect the orphan's bounding box with the row bands, and map the cell to the closest grid row.
9c. Compute the left and right boundary of the vertical band for each grid column (min/max x coordinates per column).
9d. Intersect the orphan's bounding box with the column bands, and map the cell to the closest grid column.
9e. If the table cell under the identified row and column is not empty, extend its content with the content of the or-
- 9b. Intersect the orphan's bounding box with the row bands, and map the cell to the closest grid row.
- 9c. Compute the left and right boundary of the vertical band for each grid column (min/max x coordinates per column).
- 9d. Intersect the orphan's bounding box with the column bands, and map the cell to the closest grid column.
- 9e. If the table cell under the identified row and column is not empty, extend its content with the content of the or-
phan cell.
@@ -473,27 +406,22 @@ Figure 8: Example of a table with multi-line header.
Figure 9: Example of a table with big empty distance between cells.
<!-- image -->
Figure 10: Example of a complex table with empty cells.
<!-- image -->
Figure 14: Example with multi-line text.
<!-- image -->
Figure 11: Simple table with different style and empty cells.
<!-- image -->
Figure 12: Simple table predictions and post processing.
<!-- image -->
<!-- image -->
@@ -502,22 +430,18 @@ Figure 12: Simple table predictions and post processing.
Figure 13: Table predictions example on colorful table.
<!-- image -->
Figure 16: Example of how post-processing helps to restore mis-aligned bounding boxes prediction artifact.
<!-- image -->
<!-- image -->
Figure 15: Example with triangular table.
<!-- image -->
Figure 17: Example of long table. End-to-end example from initial PDF cells to prediction of bounding boxes, post processing and prediction of structure.
<!-- image -->