This project is a deepfake image detection system that utilizes the VGG16 model to classify images as real or fake. The application is built using Flask, allowing users to upload images and receive predictions on their authenticity.
├── .gitignore
├── README.md
├── app.py
├── requirement.txt
├── static
│ ├── favicon.ico
│ ├── logo.png
│ ├── styles
│ │ └── main.css
│ └── uploads
│ ├── deep r.jpg
│ ├── fake_8467.jpg
│ ├── fake_8475.jpg
│ └── real r.jpg
├── templates
│ ├── about.html
│ ├── base.html
│ ├── contact.html
│ ├── index.html
│ └── result.html
├── test_model.py
└── train_model.py
Python 3.x Virtual environment (optional but recommended)
git clone https://github.com/mouleshgs/deepfake-detection
cd deepfake-detectionpython -m venv venv
source venv/bin/activate # On Windows: venv\Scripts\activatepip install -r requirement.txt
Run the following command to train the deepfake detection model using VGG16:
python train_model.py
To test the trained model:
python test_model.py
Running the Flask Application
python app.py
Then, open a browser and visit http://127.0.0.1:5000/ to access the web interface.
- Architecture: VGG16
- Pretrained Weights: ImageNet
- Customization: The model is fine-tuned with additional dense layers for binary classification (real vs. fake images).
Upload an image via the web interface. Receive a prediction on whether the image is real or fake. View model results and details.