Deepfake Image Detection Tool Using Xception Architecture

This project is a web-based Deepfake Image Detection Tool developed for Apex Broadcasting Network Ltd to verify the authenticity of digital images before publication. The system uses a deep learning model based on the Xception architecture to accurately distinguish between real and manipulated images.

Features

Image upload and deepfake detection
Xception-based deep learning detection engine
Confidence score for each prediction
Deepfake literacy and awareness content
Secure image handling with CSRF protection and rate limiting

Technology Stack

Backend: Python (Flask)
Frontend: HTML, CSS, JavaScript
Deep Learning Model: Xception (TensorFlow/Keras)
Security: CSRF protection, rate limiting, secure headers

Project Structure

app.py – Flask backend and detection logic
templates/index.html – User interface
static/ – Images and frontend assets
best_xception_model_finetuned.keras – Trained model
uploads/ – Temporary image storage

How to Run Locally

pip install -r requirements.txt
python app.py
"# deepfake-tool"