Skip to content

fafsari/Multi-Compartment-Segmentation

 
 

Repository files navigation

Multi Compartment Segmentation

Multi compartment segmentation is a detectron2 based panoptic segmentation model that is implemented as a pipeline in Digital Slide Archive (DSA) developed and maintained by CMI lab at the University of Florida. This DSA plugin allows users to segment a kidney Whole Slide Image (WSI) into 6 compartments, namely: cortical interstitium, medullary interstitium, non-sclerotic glomerulus, sclerotic glomerulus, tubule, and artery/arteriole.

Quick Start Guide

This DSA plugin allows users to run image analysis jobs through the HistomicsUI by the following steps:

  1. Create an account at DSA and log in with your credentials.
  2. Upload a WSI in a folder either unders Collections tab or in your User directory
  3. Open the image in HistomicsUI by clicking the arrow icon:

  1. From Analyses tab on the top right in HistomicsUI select: sarderlab/ComPRePS/segmentation/MultiCompartmentSegment

  1. Populate the fields in the tab that appears on the left side of the page: Select the Base Directory as the directory where your WSI is located and select the Model File from the Collections/models/segmentation_models/Multi_compartment_model with an available pre train model and click Sumbit botton.

  1. Once the job sucessfuly runs, the segmented compartments will appear on top of the image. You can visuazlie them all or one-by-one by selecting the layers you wish to view from Annotations/Other tab on the right.

See Also

Correlating Deep Learning-Based Automated Reference Kidney Histomorphometry with Patient Demographics and Creatinine by Lucarelli N. et. al.

Contact

Sayat Mimar (SWE): [email protected]
Pinaki Sarder (PI): [email protected]

About

For hpg prediction

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 98.3%
  • Dockerfile 1.2%
  • Other 0.5%