search

zejinwang / Blind2Unblind

This is an official implementation of the CVPR2022 paper "Blind2Unblind: Self-Supervised Image Denoising with Visible Blind Spots".
https://arxiv.org/abs/2203.06967
122 stars 14 forks source link

Grayscale synthetic noise #17

Closed pchizhov closed 1 year ago

pchizhov commented 1 year ago

There seems to be no code for training/testing on grayscale data with synthetic Gaussian noise (BSD400, BSD68, Set12, and Urban100), the results for which are presented in Table 2. The weights for these models also are not present.

Also, it is not clear from the paper, how exactly you use Urban100 data, because it is originally sRGB and in pairs of high- and low-resolution images.

It would be great if you could comment on that.

zejinwang commented 1 year ago

Please refer to Recorrupted-to-Recorrupted for more details. Actually, Blind2Unblind can perform much better on tiny grayscale images. The initial learning rate should be set to 1e-3 on small train sets.