search

mahmoodlab / CLAM

Data-efficient and weakly supervised computational pathology on whole slide images - Nature Biomedical Engineering
http://clam.mahmoodlab.org
GNU General Public License v3.0
1.11k stars 362 forks source link

PAIP 2019 datasets #33

Closed codingcode111 closed 3 years ago

codingcode111 commented 3 years ago

Hi, Is it possible to use your work on PAIP 2019 Liver cancer Segmentation Challenge datasets. This challenge is a regional segmentation https://paip2019.grand-challenge.org/Dataset/ we need first to generate patches and classify wether the patches (viable or whole) based on the ROI for viable for each slide, which has been provided in the csv file, and then we do the segmentation task. If you think your work can be used on my datasets, can you kindly explain how and provide more details, I’m new to this task and your reply and suggestion would be highly appreciated.

best regards

fedshyvana commented 3 years ago

Hi, the repository is meant for weakly-supervised classification (without detailed patch-level annotations) therefore I don't think it's the right tool for the challenge.