English

FEDS -- Filtered Edit Distance Surrogate

Computer Vision and Pattern Recognition 2021-05-27 v2

Abstract

This paper proposes a procedure to train a scene text recognition model using a robust learned surrogate of edit distance. The proposed method borrows from self-paced learning and filters out the training examples that are hard for the surrogate. The filtering is performed by judging the quality of the approximation, using a ramp function, enabling end-to-end training. Following the literature, the experiments are conducted in a post-tuning setup, where a trained scene text recognition model is tuned using the learned surrogate of edit distance. The efficacy is demonstrated by improvements on various challenging scene text datasets such as IIIT-5K, SVT, ICDAR, SVTP, and CUTE. The proposed method provides an average improvement of 11.2%11.2 \% on total edit distance and an error reduction of 9.5%9.5\% on accuracy.

Keywords

Cite

@article{arxiv.2103.04635,
  title  = {FEDS -- Filtered Edit Distance Surrogate},
  author = {Yash Patel and Jiri Matas},
  journal= {arXiv preprint arXiv:2103.04635},
  year   = {2021}
}

Comments

ICDAR 2021 camera-ready version

R2 v1 2026-06-23T23:52:07.317Z