Record link: 
http://hdl.handle.net/20.500.12708/192601
-
Title: 
SCI-3000: A Dataset for Figure, Table and Caption Extraction from Scientific PDFs
en
Citation: 
Darmanovic, F., Hanbury, A., & Zlabinger, M. (2023). SCI-3000: A Dataset for Figure, Table and Caption Extraction from Scientific PDFs. In G. Fink, R. Jain, K. Kise, & R. Zanibbi (Eds.), Document Analysis and Recognition - ICDAR 2023 : 17th International Conference, San José, CA, USA, August 21–26, 2023, Proceedings, Part I (pp. 234–251). Springer Cham. https://doi.org/10.1007/978-3-031-41676-7_14
-
Publisher DOI: 
10.1007/978-3-031-41676-7_14
-
Publication Type: 
Inproceedings - Full-Paper Contribution
en
Language: 
English
-
Authors: 
Darmanovic, Filip 
Hanbury, Allan 
Zlabinger, Markus 
-
Organisational Unit: 
E194-04 - Forschungsbereich Data Science 
-
Published in: 
Document Analysis and Recognition - ICDAR 2023 : 17th International Conference, San José, CA, USA, August 21–26, 2023, Proceedings, Part I 
-
ISBN: 
978-3-031-41676-7
-
Volume: 
14187
-
DOI of the book: 
10.1007/978-3-031-41676-7
-
Date (published): 
19-Aug-2023
-
Event name: 
The 17th International Conference on Document Analysis and Recognition - ICDAR2023 
en
Event date: 
21-Aug-2023 - 26-Aug-2023
-
Event place: 
San José, Kalifornien, United States of America (the)
-
Number of Pages: 
18
-
Publisher: 
Springer Cham
-
Keywords: 
Caption Extraction; Figure Extraction; Table Extraction
en
Abstract: 
Extracting figures and similar visual elements from PDFs of scientific publications is important but non-trivial, and progress is impeded by a lack of datasets for evaluation and machine learning. In this work, we describe and publish the SCI-3000 dataset, containing 3 000 PDFs of scientific publications (34 791 pages) with annotations of figures, tables, and corresponding captions, from the fields of computer science, biomedicine, chemistry, physics, and technology. We demonstrate the use of the dataset to benchmark two figure, table, and caption extraction approaches from recent literature: one rule-based and one deep learning-based.
en
Research Areas: 
Information Systems Engineering: 100%
-
Science Branch: 
1020 - Informatik: 100%
-
Appears in Collections:
Conference Paper

Show full item record

Page view(s)

163
checked on Jan 23, 2024

Download(s)

6
checked on Jan 23, 2024

Google ScholarTM

Check