Hello,thanks for sharing your great job but I can’t reproduce your paper. I guess it may be caused by the following:
Dataset release from PnP-AdaNet doesn’t match with Configuration of dataset.py. After I change the function sort_vol_slice(), it runs. But it still confuse me a lot.
The module DCAM base on CycleGAN synthesis source-like domain image and target-like domain image can not improve the performance of evaluation.
Why you choose tested on fake MR instead of ct.npz?
Would you suggest any ways to inspect why I'm getting bad results? Looking froward to your reply.
Hello,thanks for sharing your great job but I can’t reproduce your paper. I guess it may be caused by the following: