PP-OCR: A Practical Ultra Lightweight OCR System

PP-OCR is an ultra-lightweight open-sourced optical character recognition (OCR) system.

- A 3.5M English and Chinese OCR model + a 2.8M English digital OCR model

- Supports multi-language recognition like Korean, Japanese, German, and French

- Released under Apache 2.0 license

Paper: https://arxiv.org/abs/2009.09941

GitHub: https://github.com/PaddlePaddle/PaddleOCR