Ocr 数据集

• ICDAR(International Conference on Document Analysis and Recognition)专注文字和字符识别的会议，每两年举办一次(更新一次数据)。包含自然场景下的中英文检测和识别的标注信息，下载数据时需要注册。每次都会发布新主题的数据集，如：

It is about character and symbol recognition, printed/handwritten text recognition, graphics analysis and recognition, document analysis, document understanding, historical documents and digital libraries, document based forensics, camera and video based scene text analysis

1. CVPR 2020年Document Visual Question Answering" (DocVQA) challenge
2. 2019 Robust Reading Challenge on Scene Text Visual Question Answering, focuses on a specific type of Visual Question Answering task, where understanding the textual information in a scene is necessary in order to give an answer 
3. 2019 Multi-lingual scene text detection and recognition(MLT)
4. 2019 Large-scale Street View Text with Partial