Unconstrained Text Detection in Manga: A New Dataset and Baseline

Abstract

The detection and recognition of unconstrained text is an open problem in research. Text in comic books has unusual styles that raise many challenges for text detection. This work aims to identify text characters at a pixel level in a comic genre with highly sophisticated text styles: Japanese manga. To overcome the lack of a manga dataset with individual character level annotations, we create our own. Most of the literature in text detection use bounding box metrics, which are unsuitable for pixel-level evaluation. Thus, we implemented special metrics to evaluate performance. Using these resources, we designed and evaluated a deep network model, outperforming current methods for text detection in manga in most metrics.

Cite

Text

Del Gobbo and Herrera. "Unconstrained Text Detection in Manga: A New Dataset and Baseline." European Conference on Computer Vision Workshops, 2020. doi:10.1007/978-3-030-67070-2_38

Markdown

[Del Gobbo and Herrera. "Unconstrained Text Detection in Manga: A New Dataset and Baseline." European Conference on Computer Vision Workshops, 2020.](https://mlanthology.org/eccvw/2020/gobbo2020eccvw-unconstrained/) doi:10.1007/978-3-030-67070-2_38

BibTeX

@inproceedings{gobbo2020eccvw-unconstrained,
  title     = {{Unconstrained Text Detection in Manga: A New Dataset and Baseline}},
  author    = {Del Gobbo, Julián and Herrera, Rosana Matuk},
  booktitle = {European Conference on Computer Vision Workshops},
  year      = {2020},
  pages     = {629-646},
  doi       = {10.1007/978-3-030-67070-2_38},
  url       = {https://mlanthology.org/eccvw/2020/gobbo2020eccvw-unconstrained/}
}