SM Mirror 2.0 - AI-Powered Professional Makeup Transfer

A cutting-edge AI makeup transfer application that combines Stable-Makeup technology with MediaPipe face detection to deliver professional-quality makeup application results.

🎯 Project Overview

SM Mirror 2.0 is a full-stack application that uses advanced AI to transfer makeup from reference images to source photos while maintaining natural facial features and professional-quality results.

✨ Key Features

• Professional Makeup Transfer: Powered by Stable-Makeup for superior results vs traditional inpainting
• Multi-Region Control: Separate control for lips, eyes, eyebrows, and cheeks
• Real-time Processing: FastAPI backend with async task processing
• Advanced Face Detection: Hybrid MediaPipe + SPIGA detection system
• Progressive Makeup: Multi-step makeup application
• GPU/CPU Support: Optimized for various hardware configurations

🏗️ Architecture

SM_Mirror_2.0/
├── Stable-Makeup/          # Core Stable-Makeup implementation
├── backend/                # FastAPI server & services
├── pytorch_model_2/        # Pre-trained model weights
├── output/                 # Generated results
├── tasks.md               # Development roadmap
└── instructions.md        # Detailed development guide

🚀 Quick Start

Prerequisites

• Python 3.8+
• CUDA-capable GPU (recommended)
• 16GB+ RAM
• 10GB+ disk space for models

1. Environment Setup

# Clone the repository
git clone <your-repo-url>
cd SM_Mirror_2.0

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r backend/requirements_essential.txt

2. Model Setup

Models are automatically detected from:

• pytorch_model_2/ (current setup)
• Stable-Makeup/models/
• Download from Google Drive

3. Start the Backend

cd backend
python main.py

Server starts at http://localhost:8000

4. Test the API

# Health check
curl http://localhost:8000/health

# Test makeup transfer
curl -X POST http://localhost:8000/api/v1/makeup/transfer \
  -F "source_image=@path/to/source.jpg" \
  -F "reference_image=@path/to/makeup_reference.jpg" \
  --output result.jpg

🔧 API Documentation

Health Check

GET /health
GET /api/v1/makeup/health

Makeup Transfer

POST /api/v1/makeup/transfer
Content-Type: multipart/form-data

Fields:
- source_image: Source image file
- reference_image: Makeup reference image file

Response: Processed image with makeup applied

🧠 Technology Stack

AI/ML Components

• Stable-Makeup: Professional makeup transfer model
• MediaPipe: Face landmark detection
• SPIGA: Advanced facial structure analysis
• Stable Diffusion v1.5: Base diffusion model
• PyTorch: Deep learning framework

Backend

• FastAPI: Modern Python web framework
• Uvicorn: ASGI server
• Pillow: Image processing
• NumPy: Numerical computations

Infrastructure

• Docker: Containerization (planned)
• PostgreSQL: Database (planned)
• Redis/Celery: Async task processing (planned)

📊 Performance

• Processing Time: 30-60 seconds per image (GPU)
• Quality: Professional-grade makeup transfer
• Memory Usage: ~6-8GB GPU memory
• Supported Formats: JPG, PNG, WebP

🎨 Makeup Capabilities

Supported Makeup Types

• ✅ Lipstick: Various colors and finishes
• ✅ Eye Makeup: Eyeshadow, eyeliner, mascara
• ✅ Eyebrows: Shape and color enhancement
• ✅ Blush: Cheek contouring and color
• 🔄 Foundation: Skin tone adjustment (planned)
• 🔄 Highlighter: Face highlighting (planned)

Advanced Features

• Multi-region Control: Apply different makeup to different facial areas
• Style Mixing: Combine multiple reference styles
• Intensity Control: Adjust makeup strength
• Progressive Application: Step-by-step makeup building

🔄 Development Phases

Phase 0: PoC ✅

• MediaPipe integration
• Stable-Makeup core functionality
• Basic API endpoints

Phase 1: Walking Skeleton (Current)

• Complete end-to-end pipeline
• Frontend integration
• File upload/download

Phase 2: MVP Features

• Async processing
• Multi-region control
• User interface enhancements

Phase 3: Production Ready

• User authentication
• Multi-step makeup
• Performance optimization
• Deployment

🧪 Testing

# Test Stable-Makeup integration
python backend/test_stable_makeup.py

# Test API endpoints
python backend/test_api_upload.py

# Manual API testing
curl -X POST http://localhost:8000/api/v1/makeup/transfer \
  -F "source_image=@test_images/source.jpg" \
  -F "reference_image=@test_images/reference.jpg" \
  --output result.jpg

🚀 Deployment

Local Development

cd backend
python main.py

Production (Planned)

• GPU-enabled cloud instances (AWS/GCP)
• Docker containerization
• Load balancing with Nginx
• Monitoring and logging

🤝 Contributing

1. Fork the repository
2. Create a feature branch
3. Make your changes
4. Add tests if applicable
5. Submit a pull request

📄 License

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

🙏 Acknowledgments

• Stable-Makeup - Core makeup transfer technology
• MediaPipe - Face detection and landmarks
• SPIGA - Advanced facial analysis
• Hugging Face Diffusers - Diffusion model pipeline

📞 Support

• Check the Issues for common problems
• Review instructions.md for detailed development guidance
• Check tasks.md for development progress

---

🎉 Status: Working!

✅ Backend Integration: Stable-Makeup + FastAPI working
✅ Face Detection: MediaPipe + SPIGA hybrid system
✅ API Endpoints: Upload and process images
✅ Model Loading: Auto-detection of trained models
🔄 Frontend: In development
🔄 Production: Deployment preparation

Ready for makeup magic! 💄✨