Stable Diffusion WebUI Extension
3.5k 2026-05-06
continue-revolution/sd-webui-segment-anything
An extension for Stable Diffusion WebUI that integrates Segment Anything (SAM) and GroundingDINO to enhance image inpainting, segmentation, and matting.
Core Features
Generate precise segmentation masks via image clicks or text prompts.
Enhance Stable Diffusion and ControlNet inpainting with SAM-generated masks.
Automate image matting and facilitate LoRA/LyCORIS training set creation.
Support for advanced SAM models like SAM-HQ and MobileSAM, including CPU inference.
Integrates with ControlNet for semantic segmentation and image layout generation.
Detailed Introduction
This project is an essential extension for AUTOMATIC1111 Stable Diffusion WebUI, integrating advanced segmentation models like Segment Anything (SAM) and GroundingDINO. It significantly enhances AI image generation workflows by providing precise object selection and masking capabilities. Users can leverage its features for superior inpainting, semantic segmentation, automated image matting, and efficient creation of datasets for LoRA/LyCORIS training, streamlining complex image manipulation tasks within the Stable Diffusion ecosystem.