Giters

XLearning-SCU / 2022-CVPR-AirNet

PyTorch implementation for All-In-One Image Restoration for Unknown Corruption (AirNet) (CVPR 2022)

Geek Repo:Geek Repo

Github PK Tool:Github PK Tool

Use both custom clean photos and degraded photos?

corranmac opened this issue · comments

commented

Hi there, your work looks fantastic.

I was wondering if theres a way to train the model on a custom folder of clean and noisy photos?

Thanks!

commented

Thanks for your attention.
If you have paired clean and noisy images, you could modify the code of TrainDataset in the utils/dataset_utils.py to train the model on a custom folder. If not, it may be hard to train the model.