search

XPixelGroup / BasicSR

Open Source Image and Video Restoration Toolbox for Super-resolution, Denoise, Deblurring, etc. Currently, it includes EDSR, RCAN, SRResNet, SRGAN, ESRGAN, EDVR, BasicVSR, SwinIR, ECBSR, etc. Also support StyleGAN2, DFDNet.
https://basicsr.readthedocs.io/en/latest/
Apache License 2.0
6.51k stars 1.15k forks source link

how to directly input my own low-resolution image (without high-resolution downsampling) #260

Open zymize opened 3 years ago

zymize commented 3 years ago

Sorry to disturb you.I have one question.

How to test my own pictures using this code in tensorflow or pytorch? I want to restore my low resolution image to reconstruction ×2 resolution. In my opinion, the test is always to downsample the high-resolution test set to 2 times, and then reconstruct it. I don't know how to directly input my own low-resolution image (without high-resolution downsampling),and then reconstruct it to twice its original resolution?

Looking forward to your reply. Thank you very much.

xinntao commented 3 years ago

Hi @zymize , sorry for the late reply.

BasicSR supports upsampling with only LR inputs.

You can have a look at this option file. options/test/SRResNet_SRGAN/test_MSRResNet_x4_woGT.yml

It uses the SingleImageDataset ~