TensorFlow.js Captcha Recognition

This project aims to train and use TensorFlow.js to perform captcha recognition in the browser.

Table of Contents

• Introduction
• Usage
• License

Introduction

In this project, we leverage the power of TensorFlow to train a machine learning model for captcha recognition. By using TensorFlow.js, we can perform captcha recognition directly in the browser without the need for server-side processing.

Usage

Note: In this project, all captcha images are located in the data folder and are labeled by their filenames, with the following deformation characteristics:

1. The characters are surrounded by a lot of noise
2. The individual characters themselves are slightly distorted.

To implement the captcha recognition functionality in your browser, follow these steps:

1. prepare and label the captcha images for training
2. run 0-preprocess.py to preprocess the captcha images
3. run 1-generate_dataset.py to generate the dataset
4. run 2-train_model.py to train the model
5. run 3-test_model.py to test the model
6. run 4-convert_tfjs_model.py to convert model for tensorflow.js

Now you have the model in the web/tfjs_model folder, start a server under the web directory, you can now perform captcha recognition in the browser.

License

Distributed under the GPLv3+ License. See LICENSE for more information.