Showing 1–1 of 1 results for author: Buchal, P

SoftCTC -- Semi-Supervised Learning for Text Recognition using Soft Pseudo-Labels

Authors: Martin Kišš, Michal Hradiš, Karel Beneš, Petr Buchal, Michal Kula

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 $\unicode{x2013}$ SoftCTC $\unicode{x2013}$ 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 co… ▽ More This paper explores semi-supervised training for sequence tasks, such as Optical Character Recognition or Automatic Speech Recognition. We propose a novel loss function $\unicode{x2013}$ SoftCTC $\unicode{x2013}$ 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 naïve CTC-based approach for training on multiple transcription variants, and we make our GPU implementation public. △ Less

Submitted 19 September, 2023; v1 submitted 5 December, 2022; originally announced December 2022.

Comments: 21 pages, 8 figures, 6 tables, accepted to International Journal on Document Analysis and Recognition (IJDAR)

MSC Class: 68T07; 68T10