danielgatis/rembg

What it solves

Rembg provides a simple way to remove backgrounds from images automatically. It eliminates the need for manual masking or complex image editing software by providing a tool that can be used as a command-line interface, a Python library, or a hosted server.

How it works

The tool uses pre-trained ONNX models to perform image segmentation and background removal. It supports a wide variety of models for different use cases, including general-purpose removal, human segmentation, cloth parsing, and anime characters. It can run locally on CPUs, NVIDIA GPUs (via CUDA), or AMD GPUs (via ROCm), and also offers a cloud-based API option via withoutBG.

Who it’s for

Developers and users who need to automate background removal for single images, batch process entire folders, or integrate background removal into their own applications via a Python API or HTTP server.

Highlights

• Multiple Interfaces: Available as a CLI, Python library, HTTP server, and Docker container.
• wget-style Batch Processing: Supports processing entire folders with a "watch mode" to automatically handle new or changed files.
• Hardware Acceleration: Optimized for CPU, NVIDIA CUDA, and AMD ROCm backends.
• Extensive Model Library: Includes specialized models for humans, clothing, anime, and general-purpose salient object detection.
• Video Support: Can process RGB24 pixel binary streams, making it compatible with tools like FFmpeg for video background removal.