Robust virtual staining of landmark organelles with Cytoland

Model: VSCyto2D

Input: label-free image (e.g., QPI or phase contrast).

Output: Virtual staining of nucleus and membrane.

Note: The model works well with QPI, and sometimes generalizes to phase contrast and DIC.
It was trained primarily on HEK293T, BJ5, and A549 cells imaged at 20x.
We continue to diagnose and improve generalization

Check out our paper: Liu et al., Robust virtual staining of landmark organelles with Cytoland

For training your own model and analyzing large amounts of data, use our GitHub repository.

0.01 5
Examples
Upload Image Cell Name Imaging Modality Adjust Gamma Invert Image Rescaling image factor References
  • 1. Liu et al., Robust virtual staining of landmark organelles with Cytoland
  • 2. Edlund et. al. LIVECEll-A large-scale dataset for label-free live cell segmentation
  • 3. Maska et. al.,The cell tracking challenge: 10 years of objective benchmarking
  • 4. Guo et. al., Revealing architectural order with quantitative label-free imaging and deep learning