search

Reyhanehne / CVF-SID_PyTorch

Official implementation of the paper "CVF-SID: Cyclic multi-Variate Function for Self-Supervised Image Denoising by Disentangling Noise from Image" (CVPR 2022)
70 stars 11 forks source link

Why we need def padr_crop(img) ? #4

Open minhoooo1 opened 2 years ago

minhoooo1 commented 2 years ago

https://github.com/Reyhanehne/CVF-SID_PyTorch/blob/c89242a4ef5cc98481404192392a6f76e0fb0ec9/src/trainer/trainer.py#L25

hale1a1 commented 1 year ago

I have the same problem, have you solved it?

myavartanoo commented 1 year ago

Thank you for your question. To reduce the padding issue, we first add a padding to an input image and then crop the generated images to have the same size as the original image. Please note that, we use the "reflect" padding to follow the same noise distribution as the original image for padded pixels. Using other padding may change the noise distribution which can effect on the final result.

minhoooo1 commented 1 year ago

Thank you for your reply

------------------ 原始邮件 ------------------ 发件人: "Mohsen @.>; 发送时间: 2022年12月8日(星期四) 下午3:27 收件人: @.>; 抄送: @.>; @.>; 主题: Re: [Reyhanehne/CVF-SID_PyTorch] Why we need def padr_crop(img) ? (Issue #4)

Thank you for your question. To reduce the padding issue, we first add a padding to an input image and then crop the generated images to have the same size as the original image. Please note that, we use the "reflect" padding to follow the same noise distribution as the original image for padded pixels. Using other padding will change the noise distribution which can effect on the final result.

— Reply to this email directly, view it on GitHub, or unsubscribe. You are receiving this because you authored the thread.Message ID: @.***>