Publication detail

SoftCTC-semi-supervised learning for text recognition using soft pseudo-labels

KIŠŠ, M. HRADIŠ, M. BENEŠ, K. BUCHAL, P. KULA, M.

Original Title

SoftCTC-semi-supervised learning for text recognition using soft pseudo-labels

Type

journal article in Web of Science

Language

English

Original Abstract

This paper explores semi-supervised training for sequence tasks, such as optical character recognition or automatic speech recognition. We propose a novel loss function-SoftCTC-which is an extension of CTC allowing to consider multiple transcription variants at the same time. This allows to omit the confidence-based filtering step which is otherwise a crucial component of pseudo-labeling approaches to semi-supervised learning. We demonstrate the effectiveness of our method on a challenging handwriting recognition task and conclude that SoftCTC matches the performance of a finely tuned filtering-based pipeline. We also evaluated SoftCTC in terms of computational efficiency, concluding that it is significantly more efficient than a nave CTC-based approach for training on multiple transcription variants, and we make our GPU implementation public.

Keywords

CTC, SoftCTC, OCR, Text recognition, Confusion networks

Authors

KIŠŠ, M.; HRADIŠ, M.; BENEŠ, K.; BUCHAL, P.; KULA, M.

Released

6. 10. 2023

ISBN

1433-2825

Periodical

International Journal on Document Analysis and Recognition

Year of study

2024

Number

27

State

Federal Republic of Germany

Pages from

177

Pages to

193

Pages count

17

URL

https://link.springer.com/article/10.1007/s10032-023-00452-9

BibTex

@article{BUT185136,
  author="Martin {Kišš} and Michal {Hradiš} and Karel {Beneš} and Petr {Buchal} and Michal {Kula}",
  title="SoftCTC-semi-supervised learning for text recognition using soft pseudo-labels",
  journal="International Journal on Document Analysis and Recognition",
  year="2023",
  volume="2024",
  number="27",
  pages="177--193",
  doi="10.1007/s10032-023-00452-9",
  issn="1433-2825",
  url="https://link.springer.com/article/10.1007/s10032-023-00452-9"
}