---

🤝 Contributors

• Mustafeez Shaikh — Machine Learning Engineer (Lead)
  Deep learning model development, frame-level analysis, forgery reasoning logic, and ML-backend integration.

• Sanskriti Singh — Documentation & Testing (Lead)*

  Documentation management, testing workflows, validation, and report preparation.

• Sakib Goundi — Frontend Developer
  React-based UI development, user interaction, and frontend integration.

• Sanjana Jinaral — Backend Developer
  Backend API development and integration with machine learning inference.

---

🚀 Project Overview

The DeepFake Detection System is an AI-powered web application designed to identify manipulated (deepfake) videos using deep learning and computer vision techniques.
The system performs frame-level analysis and provides an explainable confidence-based result indicating whether a video is real or fake.

---

🎯 Objectives

• Detect deepfake videos with high accuracy
• Perform frame-wise facial analysis
• Provide forgery reasoning instead of only Real/Fake output
• Build a complete end-to-end ML-powered web system

---

🧠 Methodology

1. Video upload through frontend
2. Frame extraction based on FPS
3. Face detection and preprocessing
4. CNN-based deep learning inference
5. Confidence score calculation
6. Forgery reason generation
7. Result visualization with key frames

---

🛠️ Tech Stack

Machine Learning

• Python
• PyTorch
• OpenCV
• CNN / EfficientNet

Backend

• Flask
• REST APIs

Frontend

• React.js
• Vite

---

📂 Project Structure

DeepFake_Detection_System/
│
├── backend/
│   ├── app.py
│   ├── model_utils.py
│   └── requirements.txt
│
├── frontend/
│   ├── src/
│   ├── package.json
│   └── vite.config.js
│
├── README.md
└── .gitignore

Note: Datasets, trained models, node_modules, and virtual environments are intentionally excluded.

---

▶️ How to Run

Backend

cd backend
python -m venv venv
venv\Scripts\activate
pip install -r requirements.txt
python app.py

Frontend

cd frontend
npm install
npm run dev

---

📊 Results

• Accurate deepfake video detection
• Frame-level analysis visualization
• Explainable AI-based confidence score
• Clean and interactive user interface

---

🔮 Future Enhancements

• Image-based deepfake detection
• URL-based video analysis
• Downloadable forensic report
• GPU acceleration and cloud deployment

---

📜 License

This project is developed for academic and research purposes. All rights reserved © 2025.

---

📩 Feel free to contact us if you really need this project or require guidance.
Main Project Leads:
📧 mustafeezshaikh88@gmail.com
📧 chutur1408@gmail.com

⭐ If you like this project, give it a star on GitHub ⭐

```