File(s) stored somewhere else

Please note: Linked content is NOT stored on Janelia Research Campus and we can't guarantee its availability, quality, security or accept any liability.

Ground-truth segmentation data used for training Cellpose

posted on 10.12.2020, 19:33 by Carsen Stringer, Michalis Michaelos, Marius Pachitariu
This dataset primarily contains images of cells from a large variety of fluorescent markers and image modalities, as well as ground truth segmentations. The dataset also contains a few images of non-cell biological structures and other natural images that can be naturally segmented into repetitive structures. In some cases the nucleus is labelled in a secondary blue channel.

The dataset was used to train the Cellpose segmentation algorithm, and can be accessed at

For a complete description of the dataset, see the related paper: Cellpose: a generalist algorithm for cellular segmentation
Carsen Stringer, Tim Wang, Michalis Michaelos, Marius Pachitariu
bioRxiv 2020.02.02.931238; doi:


Howard Hughes Medical Institute



Janelia Research Campus