This is a project for the course of Machine Learning in Cyber Security. It's aim is to produce a machine learning model capable of identifying tampered regions of images testin also its performances against black box adversarial attacks.
Clone this repo:
git clone https://github.com/GianlucaDeStefano/forgerydetectionproject.gitCreate a conda environment to run it:
conda create --name tf_gpu tensorflow-gpu
Activate the environment:
conda activate tf_gpu
Install the requirements using pip:
pip install -r requirements.txt
To train a model execute:
python train.py.
├── Datasets # Contains the builder classes to download and save datasets
├── Generators # Contains the generator classes to load datasets from memory
├── Logs # Folder used to save log data during training
├── Models # Contains the implementation of the classes to build and train the model
├── train.py # Main endpoint to use to train a model
└── README.md
Pull requests are welcome. For major changes, please open an issue first to discuss what you would like to change.
Please make sure to update tests as appropriate.