
StarDist
Star-convex object detection for nuclei segmentation
StarDist represents objects as star-convex polygons, making it especially strong at separating touching nuclei in dense images. It ships pretrained models and runs in Python, Fiji, QuPath and napari.
Highlights
- Accurate instance segmentation of crowded nuclei in 2D and 3D.
- Pretrained models for fluorescence and H&E images.
- Front-ends for napari, Fiji and QuPath.