• Stars
    star
    110
  • Rank 316,770 (Top 7 %)
  • Language
    Python
  • License
    MIT License
  • Created about 7 years ago
  • Updated about 7 years ago

Reviews

There are no reviews yet. Be the first to send feedback to the community and the maintainers!

Repository Details

End-to-end variable length Captcha recognition using CNN+RNN+Attention/CTC (pytorch implementation). 端到端的不定长验证码识别

CaptchaRecognition

End-to-end variable length Captcha recognition using CNN+RNN+Attention or CTC. 端到端的不定长验证码识别

encoder: CNN+RNN or CNN

decoder: two types of attention +no attention

目前encoder端可以选择使用CNN+RNN或CNN;decoder端有两种attention方式+不使用attention。

update(2017-10-27)

  • RNN+CTC added. 新增了 CNN+RNN + CTC 的验证码识别。

Usage

Put your font file in fonts directory, and change the font file's name in line 173 of GenCaptcha.py.

To generate training data in data/, run python GenCaptcha.py.

Run python main.py to train.

把字体文件放入fonts文件夹,并修改GenCaptcha.py中第173行的字体文件名。

运行 python GenCaptcha.py ,在data/下生成数据集captcha.npz和captcha.vocab_dict。

(GenCaptcha.py中还提供了生成tfrecord文件的函数。)

运行 python main.py训练。

Using CTC

Install warp_ctc for pytorch, and run python ctcmain.py.

安装pytorch下的warp_ctc,

按照上一节的步骤生成数据后,运行 python ctcmain.py.

Results

results of using CTC (accuracy reaches 0.95 on test set):