benjaminwegener / captchasolver

based on https://medium.com/@ageitgey/how-to-break-a-captcha-system-in-15-minutes-with-machine-learning-dbebb035a710 by https://github.com/ageitgey

CaptchaSolver Jupyter Notebook

based on https://medium.com/@ageitgey/how-to-break-a-captcha-system-in-15-minutes-with-machine-learning-dbebb035a710 by https://github.com/ageitgey

DEMO

Before you get started

To run these scripts, you need the following installed:

1. Python 3
2. OpenCV 3 w/ Python extensions

• I highly recommend these OpenCV installation guides: https://www.pyimagesearch.com/opencv-tutorials-resources-guides/

3. The python libraries listed in requirements.txt

• Try running "pip3 install -r requirements.txt"

Step 1: Extract single letters from CAPTCHA images

Run:

python3 extract_single_letters_from_captchas.py

The results will be stored in the "extracted_letter_images" folder.

Step 2: Train the neural network to recognize single letters

Run:

python3 train_model.py

This will write out "captcha_model.hdf5" and "model_labels.dat"

Step 3: Use the model to solve CAPTCHAs!

Run:

python3 solve_captchas_with_model.py

captchasolver's Issues

[Question] 6-letter captchas

I've tried your CaptchaSolver on 6-letter captchas with a little more obfuscation:

Sadly extract_single_letters_from_captchas.py produces 75 % wrong letters - do I have to change the script to cope with 1/6 length compared to 1/4?