The-Learning-And-Vision-Atelier-LAVA / DASR

[CVPR 2021] Unsupervised Degradation Representation Learning for Blind Super-Resolution
MIT License
388 stars 50 forks source link

I wonder why crop HR image too? #28

Open choieq opened 3 years ago

choieq commented 3 years ago

Hi, Thank you for your work and sharing code.

I would like ask to about multiscalesrdata.py code. I don't understand why you crop HR images in line 157. Can't we just crop the low resolution image for degradation learning?

Thank you!

LongguangWang commented 3 years ago

Hi @choieq, thanks for your interest in our work. In our implementation, we generate LR images online in the dataloader using random degradations (see here). So we crop HR image patches in multiscalesrdata.py. Of course we can produce LR images in advance and just crop LR images for degradation learning. However, this will take up lots of space on the hard device.