mahmoodlab / PANTHER

Morphological Prototyping for Unsupervised Slide Representation Learning in Computational Pathology - CVPR 2024
Other
91 stars 11 forks source link

Doubt regarding PANDA dataset #19

Closed Rukhmini closed 2 weeks ago

Rukhmini commented 4 weeks ago

Hello,

I am facing some problems hndling the PANDA dataset. How did you process the PANDA dataset? I see that in the splits section of the code for PANDA data train, val and test set have image_fn column which has the filename like this "024ed1244a6d817358cedaea3783bbde_x1672_y3836_w20587_h13031_crop.h5". Did you crop the tissue region from the whole slide images and then extracted the patches from the cropped region? TIA.

andrewsong90 commented 2 weeks ago

Hi @Rukhmini

Feel free to ignore image_fn this was an internally generated column from the tissue preprocessing code. We used standard CLAM pipeline for segmenting and patching.

Best,