Baseline model and tutorial¶
We provide step-by-step tutorials, including
- Preprocessing: intensity normalization
- Training: U-Net, ViT+U-Net, Swin Transformer + U-Net
- Inference
- Build docker and make submissions
https://github.com/JunMa11/NeurIPS-CellSeg
The following open-sourced methods are highly recommended!
- Cellpose: https://github.com/MouseLand/cellpose
- Mesmer: https://github.com/vanvalenlab/deepcell-tf
- Stardist: https://github.com/stardist/stardist
- Omnipose: https://github.com/kevinjohncutler/omnipose
This nuclear segmentation summary paper is also very useful.
https://www.kaggle.com/competitions/data-science-bowl-2018/discussion/117489