Scene Text Recognition via Gated Cascade Attention

Siwei Wang, Yongtao Wang, Xiaoran Qin, Qijie Zhao, Zhi Tang

Abstract

Scene text recognition is very challenging due to the complex background, low resolution, perspective distortion and curved placement, etc. Most of the state-of-the-art methods adopt the attention-based encoder-decoder framework, and usually get suboptimal recognition performance for challenging text images due to the misalignment between attention region and target character region. In this paper, a novel module, named Gated Cascade Attention Module (GCAM), is proposed to increase the alignment precision of attention in a cascade way. Moreover, a channel and spatial attention module is introduced into the encoder to extract more discriminative features for text recognition. By assembling these two modules, a novel scene text recognizer is developed, and extensive experiments demonstrate it can achieve state-of-the-art results on multiple benchmarks of regular and irregular text images.
BibTeX
@inproceedings{wang2019scene,
  author = {Siwei Wang and Yongtao Wang and Xiaoran Qin and Qijie Zhao and Zhi Tang},
  title = {{Scene Text Recognition via Gated Cascade Attention}},
  booktitle = {ICME},
  year = {2019},
  doi = {10.1109/icme.2019.00179},
}

Loading related content...

Figure