liliu-avril/Awesome-Segment-Anything

This repository is for the first comprehensive survey on Meta AI's Segment Anything Model (SAM).

What it solves

This project provides a centralized, curated resource for understanding the Segment Anything Model (SAM) and its evolution. It addresses the challenge of keeping track of the rapidly expanding ecosystem of foundation models for image and video segmentation by organizing academic papers, open-source projects, and technical surveys in one place.

How it works

The repository acts as a comprehensive index and knowledge base. It categorizes research into seminal papers (the core SAM models) and follow-up works, while maintaining a list of related open-source projects. It also hosts a systematic survey that analyzes the background, terminology, and practical applications of SAM across various fields, including medical imaging and remote sensing.

Who it’s for

This project is for researchers, computer vision engineers, and students who want to stay updated on the latest developments in foundation models for segmentation. It is particularly useful for those looking to understand the impact of SAM on vision tasks or seeking specific implementations for their own projects.

Highlights

  • Maintains the first comprehensive survey on the Segment Anything Model.
  • Provides a continuously updated list of academic papers and open-source projects.
  • Tracks the evolution of SAM, including SAM 2, SAM 3, SAM 3D, and SAM Audio.
  • Encourages community contribution through pull requests to keep the resource current.

Related

  • Project
    Linfeng-Tang/Image-Fusion

    A comprehensive collection of research papers and code implementations for image fusion, covering multi-modal, digital photography, and remote sensing image combination techniques.

  • Project
    xemle/home-gallery

    A self-hosted web gallery for personal media that uses AI for face recognition and image discovery while keeping data private and local.

  • Project
    raspberrypi/picamera2

    Picamera2 is a Python interface for the Raspberry Pi camera system that replaces the legacy stack with a modern, libcamera-based solution.

  • Project
    harvardnlp/im2markup

    im2markup is a deep learning system that converts images of math formulas or web pages into their original LaTeX or HTML markup, using a CNN with attention-based neural machine translation. It includes a dataset, pretrained models, and full training/evaluation scripts.