xiaofengShi/CHINESE-OCR
A TensorFlow/keras/pytorch implementation for natural scene text detection and end-to-end Chinese OCR recognition.
Not currently ranked — collecting fresh signals.
star history
This project implements scene text detection and recognition using deep neural networks. It combines three models: a VGG16-based classifier for text orientation detection (0, 90, 180, 270 degrees), a CTPN (CNN+RNN) network for text region detection, and a CRNN (CNN+GRU/LSTM+CTC) network for end-to-end variable-length text recognition. The system supports both GPU and CPU environments with Python 3.6+ and TensorFlow 1.7.
Frequently asked
- What is xiaofengShi/CHINESE-OCR?
- A TensorFlow/keras/pytorch implementation for natural scene text detection and end-to-end Chinese OCR recognition.
- Is CHINESE-OCR open source?
- Yes — xiaofengShi/CHINESE-OCR is an open-source project tracked on heatdrop.
- What language is CHINESE-OCR written in?
- xiaofengShi/CHINESE-OCR is primarily written in Python.
- How popular is CHINESE-OCR?
- xiaofengShi/CHINESE-OCR has 3k stars on GitHub.
- Where can I find CHINESE-OCR?
- xiaofengShi/CHINESE-OCR is on GitHub at https://github.com/xiaofengShi/CHINESE-OCR.