3 ms·
I am surprised they only used ~2000 samples. That's a very tiny amount of data to train a decently sized CNN, even with data augmentation. I presume using a p
by Fede_V 10y ago
I am surprised they only used ~2000 samples. That's a very tiny amount of data to train a decently sized CNN, even with data augmentation.
I presume using a pretrained net probably won't help you since the features found in pathology slides are so unlike those of natural images.
- yread 10y agoEach 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.
- dave_sullivan 10y agoI've done a bit with deep learning and medical images (brain, lung, skin). One thing that surprised me, but kind of makes sense, is the net seemed to perform well without a ton of training data. I think it might actually be easier for the model because the images are much more similar than in a "natural image" setting where there are more degrees of freedom in the subject itself.
- shoyer 10y agoSomewhat surprisingly, the paper doesn't actually use deep learning.
- dandermotj 10y agoEveryone is missing this. They used simple random forests and survival models with basic k-fold cross validation. There's no breakthrough in ML here!
- drpgq 10y agoPattern recognition has been getting results for decades now on various problems and for some of those problems deep learning doesn't help that much.
- frozenport 10y agoSo, making shit up is standard for this field of biomedical imaging.