Embedded Files
This project implements a captcha solver to download the captchas from eBay Website and solve the captchas. The eBay captchas are easy to solve that has less noise in captcha codes.
Here are the steps we followed
1. Download the captcha images from eBay Website. We have collected 1500 captcha images.
2. Manually labeled the images with captcha codes and split it into train and test dataset.
3. Train the images to generate a model and stored the trained model into the directory.
python file: train.py
4. Evaluate the captcha images using the trained model.
python file: eval.py
5. Graphs of the Accuracy and Loss will be stored in the directory: Outputs/Graphs
In this part-2, we have generated captchas that are not easily solvable (more noise). So we have implemented a deep learning algorithm to break it and show how hard the captchas are?
This kind of captchas is mostly used for security purposes where the captchas can only be solved by humans and avoid robots.
1. Generate the captcha images with more noise.
We have collected a 20,000 captcha image with the help of PHP code: captcha.php
2. Split the collected images into the train and test dataset.
Python code to split the images : captcha_splitting.py
Train images directory : dataset/train
Test images directory : dataset/test
3. Train the images to generate a model and stored the trained model into the directory.
Python code to train the images: train.py
The trained model will be saved into the following directory: Outputs/Model/model.h5
4. Evaluate the captcha images using the trained model.
Python code for evaluating the captchas: eval.py
5. Graphs of the Accuracy and Loss will be stored in the directory: Outputs/Graphs
PC Requirements:
1. Intel i5 6th gen or latest.
2. At least 6GB ram.
3. Minimum 500Gb hard disk (SSD is an advantage.)
Request source code for academic purpose, fill REQUEST FORM or contact +91 7904568456 by WhatsApp or info@verilogcourseteam.com, fee applicable.
SIMULATION VIDEO DEMO-Easy Captcha (eBay Dataset)
SIMULATION VIDEO DEMO-Hard Captcha (Customized Dataset)
Report abuse