Skip to content
This repository has been archived by the owner on Aug 21, 2024. It is now read-only.

U-Net model for glomeruli segmentation in Whole Slide Images

License

Notifications You must be signed in to change notification settings

TheJacksonLaboratory/unet-glomeruli

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 
 
 
 
 
 
 
 
 

Repository files navigation

U-Net for Glomeruli segmentation in H&E WSI

U-Net model for glomeruli segmentation in H&E stained Whole Slide Images.

Usage

Segment glomeruli in a kidney H&E stained WSI using a pre-trained U-Net model. This generates a glomeruli segmentation mask that is stored in zarr format. The segmentation mask is stored inside a group called class.

python segment.py -m /path/to/checkpoint -i /path/to/zarr/files -o /ouput/directory

If this is run on a machine with GPUs, the size of the processed chunks can be modified to make the segmentation more efficient. This is limited by the GPU's memory.

python segment.py -m /path/to/checkpoint -i /path/to/zarr/files -o /ouput/directory -cs 2048

By default, only the class (Glomeruli/Background) are stored in the output file. The option -sp can be used to store the prediction probabilities. These will be stored in a separate group called probs.

python segment.py -m /path/to/checkpoint -i /path/to/zarr/files -o /ouput/directory -sp

About

U-Net model for glomeruli segmentation in Whole Slide Images

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages