PixelProof is an AI-powered deepfake detection system that allows users to upload images and receive real-time authenticity analysis using deep learning.
- Frontend: React (Vite, TypeScript)
- Backend: FastAPI (Python)
- AI Model: TensorFlow (MobileNetV2)
- Computer Vision: OpenCV
cd code/backend python -m venv venv venv\Scripts\activate pip install -r requirements.txt python main.py
Backend runs at: http://127.0.0.1:8000/docs
cd code/frontend npm install npm run dev
Frontend runs at: http://localhost:5173
POST /analyze
- Input: Image file
- Output: Fake / Real classification
The deepfake detection model is trained using transfer learning with MobileNetV2
and saved as an .h5 file for inference.
End-to-end pipeline is functional: Frontend → Backend API → AI Model → Result