What it does
Cellpose provides cellular image segmentation through a graphical interface, model processing and manual region editing.
Cellpose’s GUI accepts common image files, generates masks and lets you correct regions against the source. Save explicitly after running a model or drawing 3D masks: those actions do not trigger its usual ROI autosave. The _seg.npy output stores masks and related information, but the current documentation says it no longer includes the source image, so retain that separately. PNG masks and ImageJ ROI archives provide exchange routes. Check the input layout for multi-Z, multi-channel TIFFs and use the documented Z-stack GUI mode. The repository identifies BSD-3-Clause code and separately notes CC-BY-NC training data; those are distinct licensing statements.
Strengths
- Model-based and manual segmentation
- Dedicated interface mode for multi-Z data
Watch out for
- Model and 3D segmentation results require an explicit save
Best for
- Creating masks for cellular image analysis
Sources
- cellpose — cellpose 4.2.1-1-ga54cb48 documentation (Oct 26, 2026)
- GUI — cellpose 4.2.1-1-ga54cb48 documentation (Oct 26, 2026)
- Cellpose saved outputs (Oct 26, 2026)
- Cellpose code and training-data scope (Oct 26, 2026)