capchan

Tool to solve CAPTCHA

• Creates and trains a model based on provided greyscale images
• Uses greyscale model against other images to determine image contents

Installation:

python3 -m venv capchan
source capchan/bin/activate
pip install --upgrade pip
pip install -r requirements.txt 

Depending on Python Version, Machine Architecture and Hardware, install the specific TensorFlow package from https://www.tensorflow.org/install/pip#package_location

pip install --upgrade https://storage.googleapis.com/tensorflow/versions/*.whl

Docker Options: https://www.tensorflow.org/install/docker

CAPTCHA Forest and Image_Classification

Captcha_One

Project structure of main script and other resources used to complete challenge:

.
├── szizzleIMON.py
│   ├── ImageSplit.py
│   ├── ImageDetermine.py
│   ├── Images
│   │...outputs...
│   ├── output.txt
│   ├── image.png
│   ├── split_image_<num>.png
├── ImageGet.py
├── hex2png.py
└── bashywashy.sh

Remember to change dir paths for argparse and other scripts. Main script:

python3 szizzleIMON.py -s domain.local -v --debug

Test Scripts:

python3 ImageDetermine.py --split_images_folder ~/Captcha_One/ --data_folder Images/
python3 ImageGet.py -s domain.local -v --debug
python3 ImageSplit.py --image_path image.png
python3 hex2png.py 
./bashywashy.sh 

Captcha_Two

Project structure of main script and other resources used to complete challenge:

.
├── szizzleIMONGHARD.py
│   ├── ImageSplit.py
│   ├── nyadetermine.py
│   ├── nyasomething.py
│   ├── nyamove.py
│   ├── nyamodel.py
│   │...outputs...
│   ├── Data
│   ├── Val
│   ├── model.h5
│   ├── output.txt
│   ├── image.png
│   ├── split_image_<num>.png
├── ExtraData
├── ImageDetermine.py
├── ImageSplit.py 
├── model_graph.png
├── network_graph
├── nyaHowAc.py 
└── bashywasher.sh

Remeber to change dir paths for argparse and manually change desired image output files for getting test/val data. Main script:

./bashywasher.sh "python3 nyasomething.py -s domain.local -v --debug" 15
python3 nyamove.py 
python3 nyamodel.py

python3 szizzleIMONGRHARD.py -s domain.local -v --debug

You will need to add a Images/ folder for Captcha_One, inside the folder needs to be every letter of the alphabet as a folder, within that the single image can be places: Images/A/A/png

For Captcha_Two you will need a folder called Raw/ to place images farmed. Two more folders Data/ and Val/ similar to Captcha_One's Images/. The model will also have to be made from all of this. Data/A/*.png, Val/A/*.png, Raw/*.png and model.h5

If you have more than 10 000+ images total, then you're doing something right. image.png, output.txt and split_images_<number>.png will be made from the scripts.