search

akshaydudhane16 / Burstormer

SOTA for Burst Super-resolution, Low-light Burst Image Enhancement, Burst Image De-noising
111 stars 5 forks source link

Question for training of Burst De-noising.(loss function, data generation and training tricks) #7

Open hanhhhhh opened 1 year ago

hanhhhhh commented 1 year ago

Hello,

We noticed that you mentioned in the article that all training utilized L1 Loss. However, in the Burst Denoising section, you referred to the experimental setups of KPN and BPN, where KPN employed the Basic Loss (L2 Loss + gradient loss). We are curious about the Loss function employed in your Burst Denoising.

Furthermore, in Burst Denoising, when converting Open Image data, apart from what was mentioned in KPN, are there any additional steps or specific considerations? During the training phase of Burst Denoising, were there any other training techniques applied? Currently, we are unable to replicate the training accuracy.

We eagerly await your response!

cvkaiming commented 1 year ago

Hi, did you get the results you got from the BurstSR Track2 part of the training up to the PSNR in the article, thanks!

hanhhhhh commented 1 year ago

Sorry, I've only tried the denoising part of the code so far.

------------------ 原始邮件 ------------------ 发件人: "akshaydudhane16/Burstormer" @.>; 发送时间: 2023年8月9日(星期三) 上午9:09 @.>; @.**@.>; 主题: Re: [akshaydudhane16/Burstormer] Question for training of Burst De-noising.(loss function, data generation and training tricks) (Issue #7)

Hi, did you get the results you got from the BurstSR Track2 part of the training up to the PSNR in the article, thanks!

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

GopiRajuMatta commented 10 months ago

Hello @akshaydudhane16!

Thank you for releasing the code for testing.

Can you also please update the training code for de-noising.

Thank you Gopi

Egkang-Luis commented 9 months ago

@hanhhhhh Hi hanhhhhh. I think you success making training code for burst denoising task.

I would like to make training code for dataset, but I am curious something.

If you don't mind, could you let me know as below?

  1. Open image dataset : The training set volume of open image dataset is over 500GB. Did you used whole dataset or specific images only?
  2. If you can share the code for dataset, it will be helpful to me.

My email address is eunggukang@gmail.com

Thanks. Luis Kang.

YohanJRoh commented 5 months ago

@hanhhhhh @Egkang-Luis Egkang-Luis @akshaydudhane16 Dear all,

Could you please share your training code for Color burst denoising burstormer as well as the creation code for dataset, etc?

My email address is yohan.roh@samsung.com

I would greatly appreciate it if you could share it with me.

Thank you in advance.

Best regards and respect, Yohan