https://github.com/soloice/Chinese-Character-Recognition

This project shows how to use CNN to perform Chinese character recognition, a much more complicated task compared to MNIST digit recognition.
https://github.com/soloice/Chinese-Character-Recognition

Last synced: 29 days ago
JSON representation

This project shows how to use CNN to perform Chinese character recognition, a much more complicated task compared to MNIST digit recognition.

• Host: GitHub
• URL: https://github.com/soloice/Chinese-Character-Recognition
• Owner: soloice
• Created: 2017-02-11T15:10:33.000Z (almost 8 years ago)
• Default Branch: master
• Last Pushed: 2017-05-06T06:41:55.000Z (over 7 years ago)
• Last Synced: 2024-08-04T00:11:00.874Z (4 months ago)
• Language: Python
• Size: 186 KB
• Stars: 201
• Watchers: 12
• Forks: 76
• Open Issues: 1
• Metadata Files:
  • Readme: README.md

Awesome Lists containing this project

• awesome-captcha - soloice/Chinese-Character-Recognition - ten projekt pokazuje jak używać CNN, aby wykonać rozpoznawanie chińskich znaków, jest to znacznie bardziej skomplikowane zadanie w porównaniu do rozpoznawania cyfr MNIST. (Łamanie / Chińskie)

README

        
# Chinese-Character-Recognition

This tensorflow project shows how to use CNN to perform Chinese character recognition, a much more complicated task compared to digit recognition.

This code is based on a similar [repo](https://github.com/burness/tensorflow-101/tree/master/chinese_hand_write_rec/src), but is cleaner and outperforms that one: This code defines a data iterator class to deal with the data preprocessing and use 2 different input pipeline for training and testing data. 

Moreover, Training 16000 steps (~30 hour on my workstation with 8 CPU cores and 32 GB RAM) achieves a top 1 accuracy of 92.50% and a top 3 accuracy of 97.48%. If you have a GPU card, it should be much more faster.

The repo owner above also kindly shared the [preprocessed dataset](https://pan.baidu.com/s/1o84jIrg#list/path=%2F)

For training, run:

---------------

`python chinese_character_recognition_bn.py --mode=train --max_steps=16002 --eval_steps=100 --save_steps=500`

For validation, run:

--------------

`python chinese_character_recognition_bn.py --mode=validation`

Note that the network hasn't fully convergenced yet after 16000 mini-batches of training (though almost), so training for a longer time should be able to improve the performance furthur. As reported in [this project report](http://cs231n.stanford.edu/reports/zyh_project.pdf), the network has the potential to achieve a top 1 accuracy of 95% when properly trained (maybe even better, since in the aforementioned paper they didn't use batch normalization, which should improve generalization capacity of the network). So if you have a powerful GPU, run more training steps.

This is [my checkpoint](https://pan.baidu.com/s/1o7CJrBW) at step 16000. If you want to try it on your laptop or something else, feel free to do so!

I also attached the learning curve of `chinese_character_recognition_bn.py`:

![accuracy-bn](https://github.com/soloice/Chinese-Character-Recognition/blob/master/accuracy-bn-16k.PNG)

![loss-bn](https://github.com/soloice/Chinese-Character-Recognition/blob/master/loss-bn-16k.PNG)