Publication Details

Importance of Textlines in Historical Document Classification

KIŠŠ, M.; KOHÚT, J.; BENEŠ, K.; HRADIŠ, M. Importance of Textlines in Historical Document Classification. In Uchida, S., Barney, E., Eglin, V. (eds) Document Analysis Systems. Lecture Notes in Computer Science. La Rochelle: Springer Nature Switzerland AG, 2022. p. 158-170. ISBN: 978-3-031-06554-5.
Czech title
Význam řádků textu při klasifikaci historických dokumentů
Type
conference paper
Language
English
Authors
Kišš Martin, Ing. (DCGM)
Kohút Jan, Ing. (DCGM)
Beneš Karel, Ing. (DCGM)
Hradiš Michal, Ing., Ph.D. (DCGM)
URL
Keywords

Historical document classification, Script and font classification, Document
origin localization, Document dating.

Abstract

This paper describes a system prepared at Brno University of Technology for ICDAR
2021 Competition on Historical Document Classification, experiments leading to
its design, and the main findings. The solved tasks include script and font
classification, document origin localization, and dating. We combined patch-level
and line-level approaches, where the line-level system utilizes an existing,
publicly available page layout analysis engine. In both systems, neural networks
provide local predictions which are combined into page-level decisions, and the
results of both systems are fused using linear or log-linear interpolation. We
propose loss functions suitable for weakly supervised classification problem
where multiple possible labels are provided, and we propose loss functions
suitable for interval regression in the dating task. The line-level system
significantly improves results in script and font classification and in the
dating task. The full system achieved 98.48%, 88.84%, and 79.69% accuracy in the
font, script, and location classification tasks respectively. In the dating task,
our system achieved a mean absolute error of 21.91 years. Our system achieved the
best results in all tasks and became the overall winner of the competition.

Published
2022
Pages
158–170
Proceedings
Uchida, S., Barney, E., Eglin, V. (eds) Document Analysis Systems
Series
Lecture Notes in Computer Science
Volume
13237
Conference
International Workshop on Document Analysis System, La Rochelle, FR
ISBN
978-3-031-06554-5
Publisher
Springer Nature Switzerland AG
Place
La Rochelle
DOI
10.1007/978-3-031-06555-2_11
UT WoS
000870314500011
EID Scopus
2-s2.0-85131141303
BibTeX 
@inproceedings{BUT178121,
  author="Martin {Kišš} and Jan {Kohút} and Karel {Beneš} and Michal {Hradiš}",
  title="Importance of Textlines in Historical Document Classification",
  booktitle="Uchida, S., Barney, E., Eglin, V. (eds) Document Analysis Systems",
  year="2022",
  series="Lecture Notes in Computer Science",
  volume="13237",
  pages="158--170",
  publisher="Springer Nature Switzerland AG",
  address="La Rochelle",
  doi="10.1007/978-3-031-06555-2\{_}11",
  isbn="978-3-031-06554-5",
  url="https://pero.fit.vutbr.cz/publications"
}
Back to top