search

Wenchao-Du / LIR-for-Unsupervised-IR

This is an implementation for the CVPR2020 paper "Learning Invariant Representation for Unsupervised Image Restoration"
https://arxiv.org/pdf/2003.12769.pdf
98 stars 21 forks source link

About the source and placement of the data set #13

Open haolin512900 opened 2 years ago

haolin512900 commented 2 years ago

Does the clean data set come from the VOC data set? So do you need to search for noise data by yourself? Where is the noise data set used by the author? Is the clean data set put into Celeba_A, and then the noise data set into Celeba_B?

Wenchao-Du commented 2 years ago

Voc data contains lots of clean images, which could be used to construct unpaired clean and noisy data easily. In fact, you could use any other dataset to train your tasks.

haolin512900 commented 2 years ago

thank you for your answer It means that you can train with the voc data set, right, and then divide the voc into a noise data set and a clear data set into Celeb_A and Celeb_B, right? OK

------------------ 原始邮件 ------------------ 发件人: "Wenchao-Du/LIR-for-Unsupervised-IR" @.>; 发送时间: 2021年9月26日(星期天) 上午9:38 @.>; @.**@.>; 主题: Re: [Wenchao-Du/LIR-for-Unsupervised-IR] About the source and placement of the data set (#13)

Voc data contains lots of clean images, which could be used to construct unpaired clean and noisy data easily. In fact, you could use any other dataset to train your tasks.

— You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

s0966066980 commented 2 years ago

Will Celeb_TA and Celeb_TB affect the training result?

wyhhhhhhhh commented 3 months ago

thank you for your answer It means that you can train with the voc data set, right, and then divide the voc into a noise data set and a clear data set into Celeb_A and Celeb_B, right? OK … ------------------ 原始邮件 ------------------ 发件人: "Wenchao-Du/LIR-for-Unsupervised-IR" @.>; 发送时间: 2021年9月26日(星期天) 上午9:38 @.>; @.**@.>; 主题: Re: [Wenchao-Du/LIR-for-Unsupervised-IR] About the source and placement of the data set (#13) Voc data contains lots of clean images, which could be used to construct unpaired clean and noisy data easily. In fact, you could use any other dataset to train your tasks. — You are receiving this because you authored the thread. Reply to this email directly, view it on GitHub, or unsubscribe. Triage notifications on the go with GitHub Mobile for iOS or Android.

Hello, I would like to ask if you have reproduced this paper, and whether the clear image and the noisy image correspond to Celeb_A and Celeb_B respectively.