4 ms·
Each image can be multiple gigabytes so it could be computationally difficult. Also those are probably all the lung cancer slides that TCGA has (edit: there see
by yread 10y ago
Each image can be multiple gigabytes so it could be computationally difficult. Also those are probably all the lung cancer slides that TCGA has (edit: there seems to be 3500+ lung cancer slides at [1] so I was wrong. Maybe they're not all H&E stained)
[1] http://cancer.digitalslidearchive.net/ http://cancer.digitalslidearchive.net/
- acveilleux 10y agoYou also need to retain a significant set of slides the classifier wasn't trained with so you can verify that it does work for data outside the training set.