# bowang-lab/MedSAM

Segment Anything in Medical Images

Repository: https://github.com/bowang-lab/MedSAM
Canonical: https://ross.abutalabs.com/products/medsam
Homepage: https://www.nature.com/articles/s41467-024-44824-z
Language: Jupyter Notebook
License: Apache-2.0
License Family: permissive
Last push: 2025-05-07T13:36:16+00:00

## Health v2 (maintenance only)
Score: 29/100 (v2, computed 2026-09-03T02:20:16.233290+00:00)
- activity 20, release rhythm 8, longevity 87
- inputs: {"age_days": 1229, "days_push": 483, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: none
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 4379, forks 597 (observed 2026-08-28T04:08:46.943491+00:00)

## What it is
MedSAM is a fine-tuned Segment Anything Model (SAM) foundation model for universal medical image segmentation, trained on over 1.5 million image-mask pairs across 10 imaging modalities and 30+ cancer types. It provides CLI, Jupyter notebook, and PyQt GUI interfaces for promptable segmentation of medical images via bounding boxes.

## Use cases
- segment tumors and organs in CT and MRI scans
- run promptable segmentation on medical images with bounding boxes
- fine-tune a segmentation model on my own medical imaging dataset
- preprocess CT/MR scans for segmentation model training
- interactively segment medical images through a desktop GUI

## When to choose
- you need a general-purpose medical image segmentation model that generalizes across modalities and organs
- you want a published, well-validated foundation model with checkpoints and training pipelines
- you prefer prompt-based (bounding box) segmentation instead of training task-specific models

## When to avoid
- you need 3D volumetric or video segmentation natively (use MedSAM2 instead)
- you need fully automatic segmentation without user-provided prompts
- you lack a GPU or cannot download large model checkpoints

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, image-processing, computer-vision, cli, gui
- domain: healthcare, deep-learning, artificial-intelligence, image-processing
- platform: python, cross-platform
- tags: medical-imaging, image-segmentation, segment-anything, foundation-model, sam, ct, mri, pytorch, gpu

## Member repositories
- bowang-lab/MedSAM (main) score 29

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:08:46.943491+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:21:18.482581+00:00, confidence not recorded.
  - readme: https://github.com/bowang-lab/MedSAM (fetched 2026-08-28T04:08:46.943491+00:00, sha 09909f076728)
  - homepage: https://www.nature.com/articles/s41467-024-44824-z (fetched 2026-08-29T09:09:46.739837+00:00, sha de51146e6529)
  - site_page: https://www.nature.com/openresearch/about-open-access/information-for-institutions (fetched 2026-08-29T09:09:46.753627+00:00, sha 9631b227fce6)
  - site_page: https://www.nature.com/npg_/company_info/index.html (fetched 2026-08-29T09:09:46.750342+00:00, sha 02db2231ae28)
  - site_page: https://www.nature.com/npg_/press_room/press_releases.html (fetched 2026-08-29T09:09:46.752105+00:00, sha 8d21d099acec)
- Data as of 2026-08-30T08:39:29.467469+00:00.