Hi Sam, sorry I am not totally clear about your question, are you saying you already have the tiles (presumably stored in some standard image format like .jpg or .png) and are wondering how you might generate h5/pt files used as inputs to training a model?
Max
My apologies for the delay in answering this, I had not seen your response. Yes your question is correct, I have already generated the tiles for my cohort in JPG format, and I am wondering how to use these as input. This is necessary as the slide images and the annotations are in pretty strange formats at present.
Hi Sam, sorry I am not totally clear about your question, are you saying you already have the tiles (presumably stored in some standard image format like .jpg or .png) and are wondering how you might generate h5/pt files used as inputs to training a model? Max
Originally posted by @fedshyvana in https://github.com/mahmoodlab/CLAM/issues/22#issuecomment-733254901
My apologies for the delay in answering this, I had not seen your response. Yes your question is correct, I have already generated the tiles for my cohort in JPG format, and I am wondering how to use these as input. This is necessary as the slide images and the annotations are in pretty strange formats at present.