Segment Anything for Pathology
Applying foundation segmentation models to histology.
Its practical value in pathology today is speeding up manual annotation, not replacing it.
- Cost
- Free
- Hardware
- Needs a consumer GPU
- Offline
- Works without a connection
- Scanner
- Not required
- Your data
- Stays on your machine
- Licence
- Apache-2.0
- Maturity
- research
- From
- Meta AI
What it is
General-purpose promptable segmentation model that has been widely adapted for histology, most usefully as an interactive annotation accelerator rather than as an unattended segmenter.
Caveats
Not trained on pathology. Performance on nuclei is well below purpose-built models; treat pathology-specific fine-tunes separately.
Licence notes
Check the repository before any reuse.
Activity: 54,633 stars · last commit 2024-09-18 — refreshed nightly
Related
Catalogued 2026-08-02, last checked 2026-08-02. View the source record · Report an error