Qiufu Li
Qiufu Li
> leaky relu is in batchnorm2d Thanks for your reply. I have found that. That is a strange way of coding.
I have updated the download link and password for the pre-training file.
Thank you for your attentions to our works! The code of WResNet_A and WResNet_C have been uploaded.
All of them apply DWT to down-sample the internal data in deep networks. In the down-sampling, WResNet only keeps the low-frequency component output by DWT, and abandon the three high-frequency...
I guess that you didn't add "data.cuda()" for your input data.
Are you processing one-channel images? You should set the "in_channel = 1" in the first convolutional layer of VGG19, which can be implemented by replacing the line 100 in "SegNet.py"...
> Thank you so much for the help!Also looking at the folder of the DWT-IDWT I have a question:What are the difference between the DWT_IDWT_layer and the DWT_IDWT_CS? I have...
Yes. The wavelet.rec and wavelet.dec can be reversed in DWT and IDWT.
of course. You should reset some parameters, such as 'num_classes', 'in_channel', etc., before running the code on your data. But there is no information about your own dataset. So I...
Are you trying to classify the image pixels into foreground (damage) and background? I believe that wavelet is useful to detect the fine damage like scratches, etc., and I suggest...