Segment Anything at Home

This project is fork of Segment Anything WebUI which uses Segment Anything Model by Meta. The UI is based on Gradio.

Changes from the original version

• No bells and whistles: this version is stripped from video segmentation and text prompt segmentation. Focusing only on click-based interactive segmentation.
• Performance improvements: faster use of SAM predictor using these tricks:
  • The model is loaded only ONCE per session reducing unnecessary I/O (this reduces annotation latency by ~2x.)
  • SAM embeddings are calculated only ONCE per image (this reduces annotation latency by ~10x.)
• Cleaner codebase: the codebase is simplified and cleaned up. The codebase is also made more modular and easier to understand.

Usage

Following usage is running on your computer.

• Install Segment Anything（more details about install Segment Anything:

  $ pip install git+https://github.com/facebookresearch/segment-anything.git

• git clone this repository：

  $ git clone https://github.com/imadtoubal/sam-home.git
  $ cd sam-home

• Make a new folder named checkpoints under this project，and put the downloaded weights files in checkpoints。You can download the weights using following URLs：

  • vit_h: ViT-H SAM model
  • vit_l: ViT-L SAM model
  • vit_b: ViT-B SAM model

• Run：

  python app.py
  

Note： Default model is vit_b，the demo can run on CPU. Default device is cpu。

Reference

• Original repository: Segment Anything WebUI
• Thanks to the wonderful work Segment Anything and OWL-ViT