Language Segment-Anything is an open-source project that uses text prompts and instance segmentation to create masks for specific objects in images. It's an easy-to-use for object detection and image segmentation that's built on the recently released SAM model by Meta AI and the GroundingDINO detection model.
- Add Stable diffusion inpainting.
- Add Lama-cleaner to remove segmented area.
- Zero-shot text-to-bounding box approach for object detection.
- GroundingDINO detection model integration with the latest SAM model.
- Deployment using Gradio on HuggingFace.
- Customizable text prompts for object detection (using GroundingDINO) and segmentation (using SAM from Meta AI).