This repository contains computational pipelines to perform Deep Learning-based classification on Whole Slide Images (WSIs).
Currently, this codebase is structured as follows -
[1] CPATH - Deep Learning workflows for WSI classification/feature extraction using "supervised" as well as "self-supervised" learning.
[2] Czi_processing - Working with WSIs in ".czi" format.
[3] HnE_normalization - Stain normalization protocol.
[4] Testing_other_repos - Adapting latest SOTA such as HIPT (https://github.com/mahmoodlab/HIPT) and SAM (https://github.com/facebookresearch/segment-anything).