kadirnar/segment-anything-video

MetaSeg: Packaged version of the Segment Anything repository

What it solves

MetaSeg is a packaged version of the Segment Anything Model (SAM), designed to make the powerful image and video segmentation model more accessible and easier to install via pip. It simplifies the process of applying SAM to various media types and integrates with other computer vision tools.

How it works

The project provides high-level wrappers (SegAutoMaskPredictor and SegManualMaskPredictor) that allow users to perform automatic or manual mask generation. It supports different model types (ViT-B, ViT-L, and ViT-H) and can be applied to both static images and video files.

Additionally, it integrates with:

  • SAHI (Slicing Aided Hyper Inference): To improve detection of small objects by slicing images into patches.
  • FalAI: To enable running segmentation tasks on cloud GPUs.
  • External Detectors: It can use bounding boxes from models like YOLOv5/8, Detectron2, MMDetection, and Torchvision to guide the segmentation.

Who it’s for

Developers and researchers who want to use the Segment Anything Model without the complexity of manual installation or setup, and those who need to integrate segmentation into video workflows or cloud-based pipelines.

Highlights

  • Easy Installation: Installable via pip install metaseg.
  • Video Support: Built-in capabilities for segmenting objects in video files.
  • ** uma l-box/point selection**: Support for manual selection using points and bounding boxes.
  • Tooling Integration: Native support for SAHI and FalAI cloud GPU services.
  • Multi-Model Support: Compatible with various object detection models for automated masking.

Related

  • Project
    opengeos/segment-geospatial

    A Python package that adapts the Segment Anything Model (SAM) for geospatial data, enabling easy segmentation of satellite imagery using text, points, or boxes.

  • Project
    yatengLG/ISAT_with_segment_anything

    An interactive semi-automatic image segmentation annotation tool that uses the Segment Anything Model (SAM) to accelerate the creation of labeled datasets.

  • Project
    tianrun-chen/SAM-Adapter-PyTorch

    A framework for adapting the Segment Anything Model (SAM) to improve performance in challenging scenes like camouflage, shadows, and medical imaging.

  • Project
    computational-cell-analytics/micro-sam

    micro_sam is a napari‑based library that adapts Meta’s Segment Anything model for microscopy. It provides interactive 2‑D/3‑D segmentation and tracking with a few clicks, includes microscopy‑fine‑tuned SAM weights, and offers optional automatic prompt generation and lightweight fine‑tuning tools.

  • Project
    facebookresearch/sam3

    SAM 3 (Segment Anything with Concepts) is Meta’s 848 M‑parameter foundation model that lets you prompt an image or video with free‑form text (or visual exemplars) and receive masks, boxes, and scores for *all* matching objects. It combines a DETR‑style detector and a SAM 2‑style tracker, introduces a presence token for fine‑grained prompt discrimination, and is trained on >4 M auto‑annotated concepts. The repo provides installation steps, example notebooks, and a new SA‑CO benchmark (270 K concepts) for evaluation.