OCR-IDL: OCR Annotations for Industry Document Library Dataset

Ali Furkan Biten; Rubèn Tito; Lluis Gomez; Ernest Valveny; Dimosthenis Karatzas

doi:10.1007/978-3-031-25069-9_16

OCR-IDL: OCR Annotations for Industry Document Library Dataset

Ali Furkan Biten^*, Rubèn Tito, Lluis Gomez, Ernest Valveny, Dimosthenis Karatzas

^*Corresponding author for this work

Department of Computer Science

Computer Vision Center (CVC)

Research output: Chapter in Book › Chapter › Research › peer-review

Abstract

Pretraining has proven successful in Document Intelligence tasks where deluge of documents are used to pretrain the models only later to be finetuned on downstream tasks. One of the problems of the pretraining approaches is the inconsistent usage of pretraining data with different OCR engines leading to incomparable results between models. In other words, it is not obvious whether the performance gain is coming from diverse usage of amount of data and distinct OCR engines or from the proposed models. To remedy the problem, we make public the OCR annotations for IDL documents using commercial OCR engine given their superior performance over open source OCR models. It is our hope that OCR-IDL can be a starting point for future works on Document Intelligence. All of our data and its collection process with the annotations can be found in https://github.com/furkanbiten/idl_data.

Original language	English
Title of host publication	Computer Vision – ECCV 2022 Workshops, Proceedings
Editors	Leonid Karlinsky, Tomer Michaeli, Ko Nishino
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	241-252
Number of pages	12
ISBN (Print)	9783031250682
DOIs	https://doi.org/10.1007/978-3-031-25069-9_16
Publication status	Published - 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	13804 LNCS
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Access to Document

10.1007/978-3-031-25069-9_16

Cite this

Biten, A. F., Tito, R., Gomez, L., Valveny, E., & Karatzas, D. (2023). OCR-IDL: OCR Annotations for Industry Document Library Dataset. In L. Karlinsky, T. Michaeli, & K. Nishino (Eds.), Computer Vision – ECCV 2022 Workshops, Proceedings (pp. 241-252). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13804 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-25069-9_16

Biten, Ali Furkan ; Tito, Rubèn ; Gomez, Lluis et al. / OCR-IDL : OCR Annotations for Industry Document Library Dataset. Computer Vision – ECCV 2022 Workshops, Proceedings. editor / Leonid Karlinsky ; Tomer Michaeli ; Ko Nishino. Springer Science and Business Media Deutschland GmbH, 2023. pp. 241-252 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inbook{36566f537bd94150b8afdc26f309026e,

title = "OCR-IDL: OCR Annotations for Industry Document Library Dataset",

abstract = "Pretraining has proven successful in Document Intelligence tasks where deluge of documents are used to pretrain the models only later to be finetuned on downstream tasks. One of the problems of the pretraining approaches is the inconsistent usage of pretraining data with different OCR engines leading to incomparable results between models. In other words, it is not obvious whether the performance gain is coming from diverse usage of amount of data and distinct OCR engines or from the proposed models. To remedy the problem, we make public the OCR annotations for IDL documents using commercial OCR engine given their superior performance over open source OCR models. It is our hope that OCR-IDL can be a starting point for future works on Document Intelligence. All of our data and its collection process with the annotations can be found in https://github.com/furkanbiten/idl_data.",

author = "Biten, {Ali Furkan} and Rub{\`e}n Tito and Lluis Gomez and Ernest Valveny and Dimosthenis Karatzas",

note = "Funding Information: Acknowledgments. This work has been supported by projects PDC2021-121512-I00, PLEC2021-00785, PID2020-116298GB-I00, ACE034/21/000084, the CERCA Programme / Generalitat de Catalunya, AGAUR project 2019PROD00090 (BeARS), the Ramon y Cajal RYC2020-030777-I / AEI / 10.13039/501100011033 and PhD scholarship from UAB (B18P0073). Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.",

year = "2023",

doi = "10.1007/978-3-031-25069-9_16",

language = "English",

isbn = "9783031250682",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "241--252",

editor = "Leonid Karlinsky and Tomer Michaeli and Ko Nishino",

booktitle = "Computer Vision – ECCV 2022 Workshops, Proceedings",

}

Biten, AF, Tito, R, Gomez, L , Valveny, E & Karatzas, D 2023, OCR-IDL: OCR Annotations for Industry Document Library Dataset. in L Karlinsky, T Michaeli & K Nishino (eds), Computer Vision – ECCV 2022 Workshops, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 13804 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 241-252. https://doi.org/10.1007/978-3-031-25069-9_16

OCR-IDL : OCR Annotations for Industry Document Library Dataset. / Biten, Ali Furkan; Tito, Rubèn; Gomez, Lluis et al.

Computer Vision – ECCV 2022 Workshops, Proceedings. ed. / Leonid Karlinsky; Tomer Michaeli; Ko Nishino. Springer Science and Business Media Deutschland GmbH, 2023. p. 241-252 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 13804 LNCS).

Research output: Chapter in Book › Chapter › Research › peer-review

TY - CHAP

T1 - OCR-IDL

T2 - OCR Annotations for Industry Document Library Dataset

AU - Biten, Ali Furkan

AU - Tito, Rubèn

AU - Gomez, Lluis

AU - Valveny, Ernest

AU - Karatzas, Dimosthenis

N1 - Funding Information: Acknowledgments. This work has been supported by projects PDC2021-121512-I00, PLEC2021-00785, PID2020-116298GB-I00, ACE034/21/000084, the CERCA Programme / Generalitat de Catalunya, AGAUR project 2019PROD00090 (BeARS), the Ramon y Cajal RYC2020-030777-I / AEI / 10.13039/501100011033 and PhD scholarship from UAB (B18P0073). Publisher Copyright: © 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.

PY - 2023

Y1 - 2023

N2 - Pretraining has proven successful in Document Intelligence tasks where deluge of documents are used to pretrain the models only later to be finetuned on downstream tasks. One of the problems of the pretraining approaches is the inconsistent usage of pretraining data with different OCR engines leading to incomparable results between models. In other words, it is not obvious whether the performance gain is coming from diverse usage of amount of data and distinct OCR engines or from the proposed models. To remedy the problem, we make public the OCR annotations for IDL documents using commercial OCR engine given their superior performance over open source OCR models. It is our hope that OCR-IDL can be a starting point for future works on Document Intelligence. All of our data and its collection process with the annotations can be found in https://github.com/furkanbiten/idl_data.

AB - Pretraining has proven successful in Document Intelligence tasks where deluge of documents are used to pretrain the models only later to be finetuned on downstream tasks. One of the problems of the pretraining approaches is the inconsistent usage of pretraining data with different OCR engines leading to incomparable results between models. In other words, it is not obvious whether the performance gain is coming from diverse usage of amount of data and distinct OCR engines or from the proposed models. To remedy the problem, we make public the OCR annotations for IDL documents using commercial OCR engine given their superior performance over open source OCR models. It is our hope that OCR-IDL can be a starting point for future works on Document Intelligence. All of our data and its collection process with the annotations can be found in https://github.com/furkanbiten/idl_data.

UR - http://www.scopus.com/inward/record.url?scp=85151054239&partnerID=8YFLogxK

UR - https://www.mendeley.com/catalogue/19c29c9b-85b1-3d58-8a35-de45420b765f/

U2 - 10.1007/978-3-031-25069-9_16

DO - 10.1007/978-3-031-25069-9_16

M3 - Chapter

AN - SCOPUS:85151054239

SN - 9783031250682

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 241

EP - 252

BT - Computer Vision – ECCV 2022 Workshops, Proceedings

A2 - Karlinsky, Leonid

A2 - Michaeli, Tomer

A2 - Nishino, Ko

PB - Springer Science and Business Media Deutschland GmbH

ER -

Biten AF, Tito R, Gomez L , Valveny E , Karatzas D. OCR-IDL: OCR Annotations for Industry Document Library Dataset. In Karlinsky L, Michaeli T, Nishino K, editors, Computer Vision – ECCV 2022 Workshops, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 241-252. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). Epub 2023 Feb 14. doi: 10.1007/978-3-031-25069-9_16

OCR-IDL: OCR Annotations for Industry Document Library Dataset

Abstract

Publication series

Access to Document

Other files and links

Fingerprint

Cite this