Skip to content

Implementation of the methods for modelling stain variation in a given dataset described in "StainStyleSampler: Clustering-based sampling of whole slide image appearances"

Notifications You must be signed in to change notification settings

patologiivest/StainStyleSampler

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

10 Commits
 
 

Repository files navigation

StainStyleSampler

[Under construction]

Overview

The StainStyleSampler is a Python tool designed for extracting color and stain features from histological images. It leverages various color conversion techniques, stain deconvolution, and clustering methods to build feature representations, generate 2D embeddings (via PCA or UMAP), create histograms, and select representative image references. The project integrates multiple libraries including HistomicsTK for specialized color deconvolution and tissue detection.

Features

  • Features extraction:
    • Extracts image features using different color spaces (lab, rgb, hsi).
    • Supports stain deconvolution and splitting of stain features.
    • Support for automatic clustering algorithms.
  • Features visualization:
    • Generates 2D embeddings using PCA or UMAP.
  • Reference Selection:
    • Supports various reference selection methods including random, representative, grouped, and density-based approaches.

Dependencies

About

Implementation of the methods for modelling stain variation in a given dataset described in "StainStyleSampler: Clustering-based sampling of whole slide image appearances"

Topics

Resources

Stars

Watchers

Forks