MONAI
MONAI copied to clipboard
Develop a dataset for global stain extraction
Stain normalization requires a stain matrix that are particular to an image or a set of image acquired under the same condition and settings. In MONAI pathology, we have this functionality to extract the stain matrix from a patch of a whole slide image (WSI) with deconvolution; however, to make this process robust, we need to have pipeline to perform it on all the patches of one WSI (or several WSIs), which requires a specific dataset that handles loading and feeding all patches into this transform. This dataset shares a lot of similarities with a general patch-based inference dataset so it would be better to design and implement such inference dataset with having this application in mind.