Deepfake Detection & Blockchain Integration

Overview

This project combines AI-powered deepfake detection with blockchain technology to create a secure, verifiable system for identifying manipulated media. The system analyzes images and videos for signs of deepfake manipulation, then stores the results on the Ethereum blockchain for immutable verification.

Key Features

• 🔍 Ensemble Deepfake Detection: Uses multiple AI models (MesoNet, MesoInception, EfficientNet-ViT, Xception) for improved accuracy
• ⛓️ Enhanced Blockchain Integration: Stores comprehensive analysis results on Ethereum including file hash, detection result, confidence score, timestamp, and uploader address
• 🖼️ Multi-format Support: Analyzes both images and videos
• 📊 Confidence Scoring: Provides confidence levels for all detections
• 🚀 Web Interface: User-friendly React frontend with detailed blockchain data display
• 🧠 Training Capabilities: Comprehensive training script for all models

System Architecture

┌─────────────────┐    ┌────────────────────┐    ┌──────────────────┐
│   React Frontend│────│  Flask Backend     │────│  Ethereum Network│
│                 │    │                    │    │                  │
│ • File Upload   │    │ • Deepfake Detection│    │ • Smart Contract │
│ • Results Display│    │ • Blockchain Storage│    │ • Immutable Records│
└─────────────────┘    └────────────────────┘    └──────────────────┘
                                │
                                ▼
                   ┌─────────────────────────┐
                   │  AI Model Ensemble      │
                   │                         │
                   │ • MesoNet               │
                   │ • MesoInception         │
                   │ • EfficientNet-ViT      │
                   │ • Xception              │
                   └─────────────────────────┘

Directory Structure

• deepfake-chain/ - Backend system with deepfake detection and blockchain integration
  • backend/ - Flask API server, AI models, and blockchain integration
  • blockchain/ - Solidity smart contracts and deployment scripts
• deepfake-ui/ - React frontend for user interaction

Prerequisites

• Python 3.8+
• Node.js 14+
• Ethereum wallet (MetaMask recommended)
• Infura account for Ethereum node access
• Docker (optional, for containerized deployment)

Getting Started

Backend Setup:

1. Navigate to the backend directory:

   cd deepfake-chain/backend

2. Install Python dependencies:

   pip install -r requirements.txt

3. Set up environment variables: Create a .env file with the following variables:

   INFURA_PROJECT_ID=your_infura_project_id
   CONTRACT_ADDRESS=deployed_contract_address
   PRIVATE_KEY=your_wallet_private_key
   

4. Run the backend server:

   python app.py

Frontend Setup:

1. Navigate to the frontend directory:

   cd deepfake-ui

2. Install Node dependencies:

   npm install

3. Configure blockchain connection: Update src/config.js with your contract address and Infura project ID
4. Start the development server:

   npm start

Smart Contract Deployment:

1. Navigate to the blockchain directory:

   cd deepfake-chain/blockchain

2. Deploy using Remix IDE or Hardhat:
   • Using Remix: Load DeepfakeAuthentication.sol in Remix and deploy to Sepolia testnet
   • Using script: Modify deploy_contract.py with your credentials and run:

     python deploy_contract.py

Training Models (optional):

To train the ensemble models with your own dataset:

cd deepfake-chain/backend
python train_ensemble.py --data_dir /path/to/dataset

Technology Stack

• Backend: Python, Flask, TensorFlow, PyTorch
• Frontend: React, Tailwind CSS
• Blockchain: Ethereum (Sepolia Testnet), Web3.py, Solidity
• AI Models: MesoNet, EfficientNet, Xception, Vision Transformers

API Endpoints

• POST /detect - Analyze media for deepfakes
• GET /results/{file_hash} - Retrieve analysis results
• POST /store - Store results on blockchain
• GET /blockchain/{file_hash} - Retrieve blockchain record

Security & Privacy

• All media analysis is performed locally
• SHA-256 hashes are used for file identification
• Blockchain records contain only analysis results, not media content
• Private keys are never exposed in the codebase

Contributing

1. Fork the repository
2. Create a feature branch
3. Commit your changes
4. Push to the branch
5. Open a pull request

License

This project is licensed under the MIT License - see the LICENSE file for details.