SSL4MIS
SSL4MIS copied to clipboard
HiLab-git
Semi Supervised Learning for Medical Image Segmentation, a collection of literature reviews and code implementations.
Hi, In the code: ```python loss1 = 0.5 * (ce_loss(outputs1[:args.labeled_bs], label_batch[:][:args.labeled_bs].long()) + dice_loss( outputs_soft1[:args.labeled_bs], label_batch[:args.labeled_bs].unsqueeze(1))) loss2 = 0.5 * (ce_loss(outputs2[:args.labeled_bs], label_batch[:][:args.labeled_bs].long()) + dice_loss( outputs_soft2[:args.labeled_bs], label_batch[:args.labeled_bs].unsqueeze(1))) ``` for ce_loss, it fetch...
Hi, I would like to ask whether Fixmatch algorithm supports 3D medical image dataset. As in the code, there seems to be lack of a 3D input version. Is it...
Thank you for your work, How can I download datasets from Baidu with American phone number???
Thank you for your great work! However, I do have a problem concerning train_regularized_dropout_2D.py, which is an implementation of R-Drop. From what I know about R-Drop, it constrains two sub-models,...
Hi, Thanks for your great work! In the paper "Semi-Supervised Medical Image Segmentation via Cross Teaching between CNN and Transformer", you mentioned that "we used the **latest** checkpoint for testing...
metric
Thank you so much for your work. Why are my test.py results and the results given in the papers far different on HD95? metric: [array([ 0.85472167, 1.11961011, 22.76351007]), array([ 0.83855158,...
Can you please tell me how you handle the files, I want to use some other dataset. Or how should I read the h5 file.
Thank you for the awesome repository! I've noticed `torch.nn.CrossEntropyLoss` is used for the cross entropy loss and a custom loss from `utils.losses` is used for the Dice loss, as used...
Note able to get complete ACDC data hence the training doesnot run fully?
