# ImprintLab/MedSegDiff

Using Diffusion Models to Segment/Reconstruct Organs from Medical Images [AAAI Most influential Paper]

Repository: https://github.com/ImprintLab/MedSegDiff
Canonical: https://ross.abutalabs.com/products/medsegdiff
Language: Python
License: MIT
License Family: permissive
Topics: artificial-intelligence, deep-learning, denoising-diffusion, image-segmentation, medical-imaging, segmentation
Last push: 2025-09-10T15:49:20+00:00

## Health v2 (maintenance only)
Score: 50/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 41, release rhythm 35, longevity 99
- inputs: {"age_days": 1397, "days_push": 357, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 1363, forks 202 (observed 2026-08-28T04:04:30.354546+00:00)

## What it is
MedSegDiff is a diffusion probabilistic model framework for segmenting and reconstructing organs and tissues from medical images, with a transformer-based V2 variant. It provides training and sampling scripts for datasets like BraTS and ISIC.

## Use cases
- segment organs from MRI or CT scans
- apply diffusion models to medical image segmentation
- train a segmentation model on BraTS brain tumor data
- reconstruct tissue masks from medical images
- compare transformer-based diffusion segmentation models
- run fast diffusion sampling with DPM-Solver

## When to choose
- you need state-of-the-art diffusion-based medical image segmentation
- you want a research-backed model with published papers (MIDL 2023, AAAI 2024)
- you work with 2D or 3D medical imaging datasets like BraTS or ISIC

## When to avoid
- you need a production-ready clinical imaging tool with a GUI
- you want lightweight real-time segmentation on CPU
- you need non-medical general image segmentation out of the box

## Facets
- artifact type: library
- maturity: active
- function: machine-learning, deep-learning, image-processing, computer-vision
- domain: deep-learning, artificial-intelligence, image-processing
- platform: python
- tags: diffusion-models, medical-imaging, image-segmentation, denoising-diffusion, transformer, research-code, gpu, linux

## Member repositories
- ImprintLab/MedSegDiff (main) score 50

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:04:30.354546+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 2026-08-30T04:41:23.616345+00:00, confidence not recorded.
  - readme: https://github.com/ImprintLab/MedSegDiff (fetched 2026-08-28T04:04:30.354546+00:00, sha cb3e49442d11)
- Data as of 2026-08-30T08:39:29.467469+00:00.