search

Yanfeng-Zhou / XNet

[ICCV2023] XNet: Wavelet-Based Low and High Frequency Merging Networks for Semi- and Supervised Semantic Segmentation of Biomedical Images
MIT License
190 stars 9 forks source link

Wavelet transform #4

Open Ystartff opened 1 year ago

Ystartff commented 1 year ago

HI! Congratulations on your successful submission,I checked your xnet.py and found that there are no wavelets in it, but I found that they are provided in your wavelet2D.py, is your wavelet a pre-processing method? After processing, save as two paths and enter them separately? If I misunderstand, would you please provide the corresponding code

Yanfeng-Zhou commented 1 year ago

You are right. Our model uses low-frequency and high-frequency images produced by wavelet transform as input. So you can prepare it before training!

Ystartff commented 1 year ago

I've noticed that using your wwavelet2D causes the image width to be halved, is that how your input works?

Yanfeng-Zhou commented 1 year ago

please see this issue #2