Skip to content

Question #1

@tolotolo10

Description

@tolotolo10

Hello,

I am trying to use your code for .raw images since I am working on Raw Image Denoising. I load my .raw images and shuffle pixels to get a 4 color channel image 'bggr' and I made the necessary changes so that the architecture accepts 4 color channels from the input images. Also i decreased the number of patches from 200 in your default patch generation script to just 10 patches since I already have 5000 .raw images so that becomes 5000 image patches in total. But do you think your model will work this way or you are not sure about it or you know that the models is only for gray and rgb images and it won't work this way?

For me i made the changes and the model runs but the inital loss is really poor and the PSNR too (it is still running!)
Any tips would be appreciated. Thank you!

Metadata

Metadata

Assignees

No one assigned

    Labels

    No labels
    No labels

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions