WSI preprocessing

[Irina-Zhang]

Hi, Thanks for sharing this inspiring work :)

Could you please share the steps and criteria in the preprocessing in this project? For example, the preset parameters used for CLAM preprocessing, and patch selection criteria (e.g. to exclude artefacts and patches without enough tissue, etc.). Or other preprocessing steps such as stain normalisation.

Many thanks!

[kmuw]

Dear Irina,
I've had the same issues and used create_patches_fp.py and extract_features_fp.pyof the CLAM repository. There were some issues regarding the csv files listing WSIs used for the sample/test split which needed fixing (e.g. Cluster file missing). If you are still interested I could write up my steps.

cheers
Christopher

[pniedziela]

Hi @kmuw,

I would appreciate if you could describe it.

[clemsgrs]

in case you're interested, I adapted CLAM preprocessing pipeline and add some nice improvements (e.g. multi-cpu preprocessing) : https://github.com/clemsgrs/hs2p

[irina-zys]

Dear Irina, I've had the same issues and used create_patches_fp.py and extract_features_fp.pyof the CLAM repository. There were some issues regarding the csv files listing WSIs used for the sample/test split which needed fixing (e.g. Cluster file missing). If you are still interested I could write up my steps.

cheers Christopher

Thank you very much! Could you please share more?

[irina-zys]

in case you're interested, I adapted CLAM preprocessing pipeline and add some nice improvements (e.g. multi-cpu preprocessing) : https://github.com/clemsgrs/hs2p

Many thanks for sharing! This is very helpful :)