These Jupyter notebooks allow to use 5 different deep learning architectures (Mask R-CNN, Cellpose, Stardist, Inception-V3 and U-Net) to segment nuclei. These approaches have been tuned to be efficient in a complex tissue, the mouse intestinal epithelium, with a small dataset.
Marker identification is also performed using an Inception-V3 architecture.
Video tutorials are available for:
Software installation for Windows
Software installation for macOS
Codes downloading and virtual environment creation
Training U-Net
Running U-Net
Training and running Stardist
Training and running Cellpose
Training and running Mask R-CNN
Training and running Inception-V3
Please cite our paper if you use our method:
Thierry Pécot, Maria C. Cuitiño, Roger H. Johnson, Cynthia Timmers, Gustavo Leone (2022): Deep learning tools and modeling to estimate the temporal expression of cell cycle proteins from 2D still images