Blind2Unblind icon indicating copy to clipboard operation
Blind2Unblind copied to clipboard

Datasets for synthetic noise

Open sherbret opened this issue 7 months ago • 0 comments

Hi,

Congratulations for the great work !

I was wondering: what images did you use precisely for the evaluation on BSD300 ? For what I understand it is a dataset composed of 100 images (contrary to its name). Could you please provide a list of the test images ?

Also for the Set14 dataset, two images are grayscale in this dataset. Did you remove them to create a dataset composed of 12 images ?

Thank you very much,

Sébastien Herbreteau

sherbret avatar Jul 23 '24 09:07 sherbret