1038lab/ComfyUI-RMBG

What it solves

ComfyUI-RMBG provides a comprehensive set of tools for removing image backgrounds and performing precise object segmentation. It addresses the difficulty of isolating specific elements—such as people, clothing, fashion accessories, or facial features—from an image with high accuracy and clean edges.

How it works

The project implements these capabilities as custom nodes for ComfyUI, integrating a wide variety of specialized AI models. It uses models like RMBG-2.0, BiRefNet, and BEN2 for general background removal, and SAM (Segment Anything Model) and SAM2 for text-prompted segmentation. For specialized tasks, it includes dedicated models for face parsing (segmenting 19 facial categories) and fashion/clothing segmentation (up to 18 categories).

Who it’s for

This tool is designed for digital artists, AI image generators, and designers using ComfyUI who need professional-grade masking and segmentation to refine their compositions or perform real-time background replacement.

Highlights

• Diverse Model Support: Integrates multiple state-of-the-art models including RMBG-2.0, INSPYRENET, BEN, BEN2, BiRefNet, Lucida, and SDMatte.
• Advanced Segmentation: Supports text-prompted object detection using natural language or tags via SAM, SAM2, and GroundingDINO.
• Specialized Parsing: Dedicated nodes for precise face and clothing segmentation.
• Couture-Ready: Includes specific models for fashion and accessories segmentation.
• Edge Refinement: Features tools for mask blurring, boundary offsets, and foreground refinement for better transparency handling.