Closed xwjBupt closed 3 years ago
Hi,
Focal loss would be fine, since it is point-wise loss, and each sample is considered as a binary classification problem. As for other losses, I am afraid that they are not designed for 3d operations, but you might try to reform your problem and see if the 3d loss can be expressed in a 2d form. Good luck !!
I am closing this. You can leave messages if you need further discussions.
Hi, thanks for your outstanding work!
Can your losses, for example focal loss, be used on multi-label 3D medical image segmentation task?
thank you so much! Look forward of your reply~