Open ckldan520 opened 5 years ago
Thanks for your good work. Will you consider supporting color image denoising in the future?
I definitely will support color images. I am really busy now. Can you wait for about 1 or 2 months?
Thank you very much!
Thanks for your work! Can you write more comments to your code ?
Why does the line 25 of bm3d_1st_step.py require that the height and width of img_noisy be equal? This will require the height and width of the original input image to be equal. For example, if I input a 256 * 320 image, there will be some bugs in the program.
@ckldan520 Yeah, you are right. This is a problem. Cause We only need square pictures for experiments. You can either wait for me to fix this or fix it by yourself and pull request. I think I will fix this in one month.
@ckldan520 I fixed the bug that the image height must be equal to the width, images with different widths and heights work fine now.
@ckldan520 I fixed the bug that the image height must be equal to the width, images with different widths and heights work fine now.
Thank you very much, it works fine now
Thank you for your excellent work. Will you continue to change BM3D to support color images?
Thank you for your excellent work. Will you continue to change BM3D to support color images?
I don't think I have time to read the paper. I just read some other codes of BM3D, and I believe it only involves two color transforms. But I must verify the code before releasing it. The verification workload is huge. And I have always been trying to refactor this code, and it also takes time.
Thank you for your excellent work. Will you continue to change BM3D to support color images?
https://webpages.tuni.fi/foi/GCF-BM3D/ This is a python implementation that encapsulates CPP. The color transform is not encapsulated in CPP. You can refer to this one.
Thank you for your excellent work. Will you continue to change BM3D to support color images?
https://webpages.tuni.fi/foi/GCF-BM3D/ This is a python implementation that encapsulates CPP. The color transform is not encapsulated in CPP. You can refer to this one.
Thank you very much.
Thanks for your good work. Will you consider supporting color image denoising in the future?