Training and testing images from TissueNet used in human-in-the-loop performance quantifications for Cellpose 2.0 (https://www.biorxiv.org/content/10.1101/2022.04.01.486764v1).
Labels are saved as _seg.npy files (see notebook for more details on how to use them for training: https://colab.research.google.com/github/MouseLand/cellpose/blob/main/notebooks/run_cellpose_2.ipynb).
See the license for the TissueNet images shared, please use their link for downloading the full dataset: https://datasets.deepcell.org/. And if you use TissueNet in your work cite their paper: https://www.nature.com/articles/s41587-021-01094-0.
Funding
Howard Hughes Medical Institute, Janelia Research Campus