zyddnys / manga-image-translator

Translate manga/image 一键翻译各类图片内文字 https://cotrans.touhou.ai/ (no longer working)

mindee / doctr

docTR (Document Text Recognition) - a seamless, high-performing & accessible library for OCR-related tasks powered by Deep Learning. Ongoing development and maintenance by t2k.

cs-chan / Total-Text-Dataset

Total Text Dataset. It consists of 1555 images with more than 3 different text orientations: Horizontal, Multi-Oriented, and Curved, one of a kind.