🐍 A FastAPI service that removes backgrounds from images using deep learning models, optimized for CPU and low-memory environments.

• High-quality background removal using CarveKit
• FastAPI-based REST API
• Docker support
• CPU-optimized processing
• Batch processing support
• Easy integration

This API uses CarveKit's Tracer-B7 model which provides:

• 90% accuracy (mean F1-Score, DUTS-TE)
• Optimized for general objects (people, animals, products etc.)
• CPU-friendly processing
• High-quality edge detection

Note: This API currently uses Tracer-B7 as it provides the best general-purpose performance.

• Python 3.10+
• Docker (optional)

1. Clone the repository

git clone <repository-url>
cd image-processing-cpu

2. Create virtual environment

python -m venv .venv
source .venv/bin/activate  # Linux/Mac
# or
.venv\Scripts\activate  # Windows

3. Install dependencies

pip install -r requirements.txt

4. Setup environment variables

cp .env.example .env

5. Run the application

python main.py

1. Build the image

docker build -t background-remover .

2. Run the container

docker run -p 8000:8000 background-remover

GET /

POST /remove-bg
Content-Type: multipart/form-data

file: <image-file>

Example using curl:

curl -X POST -F "[email protected]" http://localhost:8000/remove-bg

The service uses CarveKit with the following optimized settings:

• Segmentation Network: Tracer-B7
• Segmentation Mask Size: 640
• Trimap Parameters:
  • Dilation: 30
  • Erosion: 5
• Post-processing: FBA Matting
• Device: CPU optimized

.
├── app/
│   ├── api/
│   │   ├── __init__.py
│   │   └── routes.py
│   ├── core/
│   │   ├── __init__.py
│   │   └── config.py
│   ├── services/
│   │   ├── __init__.py
│   │   └── background_remover.py
│   └── utils/
│       ├── __init__.py
│       └── file_handler.py
├── script/
│   └── download-weights
├── .env.example
├── .gitignore
├── Dockerfile
├── README.md
├── RemoveBackground.py
├── main.py
└── requirements.txt

1. Create and activate virtual environment
2. Install dependencies
3. Copy .env.example to .env and adjust as needed
4. Run the application in development mode:

uvicorn main:app --reload

For production deployment, it's recommended to:

1. Use Docker
2. Set appropriate environment variables
3. Use a reverse proxy (e.g., Nginx)
4. Implement proper security measures

This project uses CarveKit for background removal.

MIT License