ImageRepair AI

An intelligent, self-hostable AI tool to restore and enhance your photos, built with a robust backend for a smooth and transparent processing experience.

Before (Original)	After (Restored)

✨ Key Features

This project is engineered for a superior user experience, giving you more insight and control over the image restoration process.

• Real-Time Processing Status: Track processing progress with status updates and estimated completion times.
• Dual-AI Enhancement Pipeline: Combines the strengths of two powerful AI models: GFPGAN specializes in restoring faces with stunning realism, while Real-ESRGAN upscales and enhances the entire image, including backgrounds.
• User-Centric Interface: A clean and modern UI featuring drag-and-drop, direct URL processing, an interactive before/after comparison slider, and multi-language support (English, 简体中文, 繁體中文, 日本語, 한국어).
• Containerized for Easy Deployment: Includes pre-configured Docker and Docker Compose files for both CPU and GPU environments, ensuring a consistent and reliable setup with a single command.

🚀 Getting Started

The recommended way to run this project is with Docker, which handles all dependencies and configuration automatically.

Prerequisites

• Git
• Docker & Docker Compose
• A compatible NVIDIA GPU with drivers and the NVIDIA Container Toolkit.(Optional：for GPU acceleration):

Installation & Launch with Docker

Remark： Before starting, make sure Docker is started.

CPU Version (works on any system)

git clone https://github.com/diandiancha/ImageRepair-AI.git
cd ImageRepair-AI
docker-compose -f docker-compose.cpu.yml up --build

Access the web interface at: http://localhost:5002

---

GPU Version (for high performance with an NVIDIA GPU)

git clone https://github.com/diandiancha/ImageRepair-AI.git
cd ImageRepair-AI
docker-compose -f docker-compose.gpu.yml up --build

Access the web interface at: http://localhost:5001

GPU Minimum Requirements

• NVIDIA GPU: GTX 1060 6GB or better
• VRAM: 6GB+ for optimal performance
• CUDA: Compatible GPU with CUDA 12.1+ support
• RAM: 8GB system memory (16GB recommended for large images)
• Storage: 5GB free space for models and dependencies

---

Method 2: Manual Python Virtual Environment Setup

This method is for developers who want to run the application outside of Docker.

Remark： This project uses Python 3.10 as the main development environment. The compatibility with other Python versions is unknown. Therefore, if you need to deploy the Python version project yourself, please ensure that the environment is correct.

Step 1: Clone Repository

git clone https://github.com/diandiancha/ImageRepair-AI.git
cd ImageRepair-AI/backend

Step 2: Create and Activate Virtual Environment

python -m venv venv

On Windows:

venv\Scripts\activate

On macOS/Linux:

source venv/bin/activate

Step 3: Install Dependencies

For CPU-only:

pip install torch torchvision torchaudio --index-url https://download.pytorch.org/whl/cpu
pip install -r backend/requirements.txt

For GPU-acceleration (NVIDIA CUDA 12.1 required):

pip install torch==2.1.2+cu121 torchvision==0.16.2+cu121 torchaudio==2.1.2+cu121 --extra-index-url https://download.pytorch.org/whl/cu121
pip install -r backend/requirements.txt

Step 4: Download AI Models

mkdir backend/models
wget -O backend/models/GFPGANv1.4.pth https://github.com/TencentARC/GFPGAN/releases/download/v1.3.0/GFPGANv1.4.pth
wget -O backend/models/RealESRGAN_x4plus.pth https://github.com/xinntao/Real-ESRGAN/releases/download/v0.1.0/RealESRGAN_x4plus.pth

Step 5: Run the Application

python backend/app.py

The application will be available at http://localhost:5000.

🛠️ Technical Details

Project Structure

ImageRepair-AI/
├── backend/
│   ├── app.py                 # Flask backend with task queue and API
│   ├── Dockerfile.cpu         # Docker build definition for CPU
│   ├── Dockerfile.gpu         # Docker build definition for GPU
│   └── requirements.txt       # Python dependencies
├── assets/
│   └── (comparison images)
├── .gitignore
├── docker-compose.cpu.yml     # Docker Compose configuration for CPU
├── docker-compose.gpu.yml     # Docker Compose configuration for GPU
├── index.html                 # Single-page frontend application
├── README.md            
└── LICENSE

API Endpoints

All endpoints are defined in backend/app.py.

Endpoint	Method	Description
/api/repair	POST	Submits an image (base64 or URL) for restoration. Returns a task_id.
/api/status/<task_id>	GET	Checks the status of a submitted task (pending, processing, completed).
/api/cancel/<task_id>	POST	Requests to cancel a pending or processing task.
/api/health	GET	Provides a health check of the service, including model status and queue size.

🤝 Contributing

Contributions are welcome! Please follow these steps:

1. Fork the repository
2. Create a feature branch (git checkout -b feature/amazing-feature)
3. Make your changes
4. Test thoroughly
5. Submit a pull request

📄 License

This project is open-source and available under the MIT License.

🙏 Acknowledgments

This project stands on the shoulders of giants. My heartfelt thanks go to the creators of these incredible open-source models:

• GFPGAN
• Real-ESRGAN
• BasicSR