search

JimmyChame / SADNet

Pytorch code for "Spatial-Adaptive Network for Single Image Denoising"
128 stars 12 forks source link

Why RSABs are not used at encoder? #3

Closed mosquitobite closed 4 years ago

mosquitobite commented 4 years ago

Hi, from ablation study in your paper, RSAB without offset transfer seems to achieve the most denoising performance. Since RSAB without offset transfer can be used at encoder, why not do that? Intuitively, performance could be much better.

JimmyChame commented 4 years ago

Thanks for your advice. Offset transfer and context block assist RSAB in better locating feature offsets. RSAB with offset transfer achieves the best performance. We calculate the offset maps from coarse to fine in the multiscale structure.