An Approach to Fake Media Detection: Integrating Deepfake Analysis, Image Forensics, and Frequency Domain Analysis

This project is a thesis-based initiative to develop a unified framework combining:

• Image Forensics
• Frequency Domain Analysis
• Deepfake Detection

The goal is to create a system that improves over time, learning from user-submitted data to enhance detection of tampered and AI-generated images.

---

🧠 Framework: MMMN

• Built as a modular Python package
• Provides functions for image analysis using:
  • Metadata extraction
  • Frequency (FFT, DCT) features
  • Deep learning detection
• Can be reused in CLI apps, Flask, or notebooks

---

🌐 WebApp Interface (Flask)

• Frontend for uploading and analyzing images
• Stores results and feedback into a MySQL database
• Supports long-term retraining by collecting labeled image samples

---

📁 Folder Structure

thesis_project/
├── MMMN/            # Core detection framework
├── webapp/          # Flask-based web interface
├── requirements.txt
└── README.md