mirrors/docling
1
0
Fork 0
mirror of https://github.com/DS4SD/docling.git synced 2025-12-09 05:08:14 +00:00
Code Issues Packages Projects Releases Wiki Activity

feat: Implement new reading-order model (#916)

Browse Source 
* Implement new reading-order model, replacing DS GLM model (WIP)

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

* Update reading-order model branch

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

* Update lockfile [skip ci]

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

* Add captions, footnotes and merges [skip ci]

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

* Updates for reading-order implementation

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

* Updates for reading-order implementation

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

* Update tests and lockfile

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

* Fixes, update tests

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

* Add normalization, update tests again

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

* Update tests with code

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

* Push final lockfile

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

* sanitize text

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

* Inlcude furniture, Update tests with furniture

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

* Fix content_layer assignment

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

* chore: Delete empty file docling/models/ds_glm_model.py

Signed-off-by: Nikos Livathinos <nli@zurich.ibm.com>

---------

Signed-off-by: Christoph Auer <cau@zurich.ibm.com>
Signed-off-by: Michele Dolfi <dol@zurich.ibm.com>
Signed-off-by: Nikos Livathinos <nli@zurich.ibm.com>
Co-authored-by: Michele Dolfi <dol@zurich.ibm.com>
Co-authored-by: Nikos Livathinos <nli@zurich.ibm.com>
This commit is contained in:
Christoph Auer
2025-02-20 17:51:17 +01:00
committed by GitHub
parent c031a7ae47
commit c93e36988f
67 changed files with 757 additions and 871 deletions
Download Patch File Download Diff File Expand all files Collapse all files

14
tests/data/groundtruth/docling_v2/2206.01062.md
View File

@@ -20,7 +20,7 @@ Accurate document layout analysis is a key requirement for highquality PDF docum
Permission to make digital or hard copies of part or all of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights for third-party components of this work must be honored. For all other uses, contact the owner/author(s).
KDD '22, August 14-18, 2022, Washington, DC, USA
KDD ’22, August 14-18, 2022, Washington, DC, USA
© 2022 Copyright held by the owner/author(s).
@@ -119,9 +119,7 @@ Figure 3: Corpus Conversion Service annotation user interface. The PDF page is s
we distributed the annotation workload and performed continuous quality controls. Phase one and two required a small team of experts only. For phases three and four, a group of 40 dedicated annotators were assembled and supervised.
Phase 1: Data selection and preparation. Our inclusion criteria for documents were described in Section 3. A large effort went into ensuring that all documents are free to use. The data sources
include publication repositories such as arXiv$^{3}$, government offices, company websites as well as data directory services for financial reports and patents. Scanned documents were excluded wherever possible because they can be rotated or skewed. This would not allow us to perform annotation with rectangular bounding-boxes and therefore complicate the annotation process.
Phase 1: Data selection and preparation. Our inclusion criteria for documents were described in Section 3. A large effort went into ensuring that all documents are free to use. The data sources include publication repositories such as arXiv$^{3}$, government offices, company websites as well as data directory services for financial reports and patents. Scanned documents were excluded wherever possible because they can be rotated or skewed. This would not allow us to perform annotation with rectangular bounding-boxes and therefore complicate the annotation process.
Preparation work included uploading and parsing the sourced PDF documents in the Corpus Conversion Service (CCS) [22], a cloud-native platform which provides a visual annotation interface and allows for dataset inspection and analysis. The annotation interface of CCS is shown in Figure 3. The desired balance of pages between the different document categories was achieved by selective subsampling of pages with certain desired properties. For example, we made sure to include the title page of each document and bias the remaining page selection to those with figures or tables. The latter was achieved by leveraging pre-trained object detection models from PubLayNet, which helped us estimate how many figures and tables a given page contains.
@@ -144,8 +142,6 @@ The complete annotation guideline is over 100 pages long and a detailed descript
Phase 3: Training. After a first trial with a small group of people, we realised that providing the annotation guideline and a set of random practice pages did not yield the desired quality level for layout annotation. Therefore we prepared a subset of pages with two different complexity levels, each with a practice and an exam part. 974 pages were reference-annotated by one proficient core team member. Annotation staff were then given the task to annotate the same subsets (blinded from the reference). By comparing the annotations of each staff member with the reference annotations, we could quantify how closely their annotations matched the reference. Only after passing two exam levels with high annotation quality, staff were admitted into the production phase. Practice iterations
Figure 4: Examples of plausible annotation alternatives for the same page. Criteria in our annotation guideline can resolve cases A to C, while the case D remains ambiguous.
<!-- image -->
05237a14f2524e3f53c8454b074409d05078038a6a36b770fcc8ec7e540deae0
@@ -192,8 +188,6 @@ In Table 2, we present baseline experiments (given in mAP) on Mask R-CNN [12], F
Table 3: Performance of a Mask R-CNN R50 network in mAP@0.5-0.95 scores trained on DocLayNet with different class label sets. The reduced label sets were obtained by either down-mapping or dropping labels.
Table 4: Performance of a Mask R-CNN R50 network with document-wise and page-wise split for different label sets. Naive page-wise split will result in GLYPH&lt;tildelow&gt; 10% point improvement.
| Class-count | 11 | 6 | 5 | 4 |
|----------------|------|---------|---------|---------|
| Caption | 68 | Text | Text | Text |
@@ -217,6 +211,8 @@ One of the fundamental questions related to any dataset is if it is "large enoug
The choice and number of labels can have a significant effect on the overall model performance. Since PubLayNet, DocBank and DocLayNet all have different label sets, it is of particular interest to understand and quantify this influence of the label set on the model performance. We investigate this by either down-mapping labels into more common ones (e.g. Caption → Text ) or excluding them from the annotations entirely. Furthermore, it must be stressed that all mappings and exclusions were performed on the data before model training. In Table 3, we present the mAP scores for a Mask R-CNN R50 network on different label sets. Where a label is down-mapped, we show its corresponding label, otherwise it was excluded. We present three different label sets, with 6, 5 and 4 different labels respectively. The set of 5 labels contains the same labels as PubLayNet. However, due to the different definition of
Table 4: Performance of a Mask R-CNN R50 network with document-wise and page-wise split for different label sets. Naive page-wise split will result in GLYPH&lt;tildelow&gt; 10% point improvement.
| Class-count | 11 | 11 | 5 | 5 |
|----------------|------|------|-----|------|
| Split | Doc | Page | Doc | Page |
@@ -302,13 +298,13 @@ Figure 6: Example layout predictions on selected pages from the DocLayNet test-s
Diaconu, Mai Thanh Minh, Marc, albinxavi, fatih, oleg, and wanghao yang. ultralytics/yolov5: v6.0 - yolov5n nano models, roboflow integration, tensorflow export, opencv dnn support, October 2021.
- [20] Shoubin Li, Xuyan Ma, Shuaiqun Pan, Jun Hu, Lin Shi, and Qing Wang. Vtlayout: Fusion of visual and text features for document layout analysis, 2021.
- [14] Nicolas Carion, Francisco Massa, Gabriel Synnaeve, Nicolas Usunier, Alexander Kirillov, and Sergey Zagoruyko. End-to-end object detection with transformers. CoRR , abs/2005.12872, 2020.
- [15] Mingxing Tan, Ruoming Pang, and Quoc V. Le. Efficientdet: Scalable and efficient object detection. CoRR , abs/1911.09070, 2019.
- [16] Tsung-Yi Lin, Michael Maire, Serge J. Belongie, Lubomir D. Bourdev, Ross B. Girshick, James Hays, Pietro Perona, Deva Ramanan, Piotr Dollár, and C. Lawrence Zitnick. Microsoft COCO: common objects in context, 2014.
- [17] Yuxin Wu, Alexander Kirillov, Francisco Massa, Wan-Yen Lo, and Ross Girshick. Detectron2, 2019.
- [18] Nikolaos Livathinos, Cesar Berrospi, Maksym Lysak, Viktor Kuropiatnyk, Ahmed Nassar, Andre Carvalho, Michele Dolfi, Christoph Auer, Kasper Dinkla, and Peter W. J. Staar. Robust pdf document conversion using recurrent neural networks. In Proceedings of the 35th Conference on Artificial Intelligence , AAAI, pages 1513715145, feb 2021.
- [19] Yiheng Xu, Minghao Li, Lei Cui, Shaohan Huang, Furu Wei, and Ming Zhou. Layoutlm: Pre-training of text and layout for document image understanding. In Proceedings of the 26th ACM SIGKDD International Conference on Knowledge Discovery and Data Mining , KDD, pages 1192-1200, New York, USA, 2020. Association for Computing Machinery.
- [20] Shoubin Li, Xuyan Ma, Shuaiqun Pan, Jun Hu, Lin Shi, and Qing Wang. Vtlayout: Fusion of visual and text features for document layout analysis, 2021.
- [21] Peng Zhang, Can Li, Liang Qiao, Zhanzhan Cheng, Shiliang Pu, Yi Niu, and Fei Wu. Vsr: A unified framework for document layout analysis combining vision, semantics and relations, 2021.
- [22] 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 International Conference on Knowledge Discovery and Data Mining , KDD, pages 774-782. ACM, 2018.
- [23] Connor Shorten and Taghi M. Khoshgoftaar. A survey on image data augmentation for deep learning. Journal of Big Data , 6(1):60, 2019.