method: DB_threshold2_TRBA_CocoValid2022-07-20

Authors: Yoonsik Kim, Taeho Kil, Seonghyeon Kim, Sukmin Seo

Affiliation: Clova AI OCR Team, NAVER/LINE Corp.

Description: The detector is based on Differentiable Binarization [1]. The recognizer is TRBA from WIW [2].
TRBA denotes TPS + ResNet Backbone + BiLSTM + Attention. The models were not jointly trained. Since DB does not output upvector, we rotated the detected region according to the aspect ratio. Cocotext has label noises (not case sensitive), and thus, we cleansed the dataset using the teacher model. Therefore, we used synthetic dataset (ST) and challenge-provided real datasets.

@article{kim2022deer, title={DEER: Detection-agnostic End-to-End Recognizer for Scene Text Spotting}, author={Kim, Seonghyeon and Shin, Seung and Kim, Yoonsik and Cho, Han-Cheol and Kil, Taeho and Surh, Jaeheung and Park, Seunghyun and Lee, Bado and Baek, Youngmin}, journal={arXiv preprint arXiv:2203.05122}, year={2022} }

@inproceedings{baek2019STRcomparisons, title={What Is Wrong With Scene Text Recognition Model Comparisons? Dataset and Model Analysis}, author={Baek, Jeonghun and Kim, Geewook and Lee, Junyeop and Park, Sungrae and Han, Dongyoon and Yun, Sangdoo and Oh, Seong Joon and Lee, Hwalsuk}, booktitle = {International Conference on Computer Vision (ICCV)}, year={2019}, pubstate={published}, tppubtype={inproceedings} }

Source code