search

NKI-AI / ahcore-old

Ahcore are the AI for Oncology histopathology core models
Apache License 2.0
0 stars 1 forks source link

Enable Macenko Normalisation for P1000 images #23

Closed sinberlin2 closed 1 year ago

sinberlin2 commented 1 year ago

Describe the bug The P1000 images have an offset that needs to be taken into consideration. Otherwise the normalisation will occur with stain vectors that are largely based on white background.

To Reproduce Run macenko with any P1000 images and visualise output.

Expected behavior The slide_image needs to be offset to compute the stain vectors across the wsi area. Environment dlup version: How installed: Python version: Operating System:

Additional context Add any other context about the problem here.

AjeyPaiK commented 1 year ago

Hi @sinberlin2 , thanks for raising the issue.

I'm trying to understand it. Do you mean to say that stain vectors are hard to calculate over the WSI for P1000 images since there is a lot of background and very little tissue region in these images?

sinberlin2 commented 1 year ago

Hello Ajey, I tried to add more details in the PR. You can also check the lines I have added in the code for further clarification. It is just 3 lines basically

AjeyPaiK commented 1 year ago

Awesome, I'll check it and also test it. Thanks Shannon! :)