li-chongyi / lighting-the-darkness-in-the-deep-learning-era-open Goto Github PK

NPE数据集我下载下来数了一下一共有85张图像，但是贵作者提供的数据显示NPE数据集图像只有84张？由于本人是个新手，所以想让作者帮忙看一哈，是不是我哪儿弄错了

I think this platform makes people more convenient to implement using different kinds of method quickly, I want to ask a question of website building,Actually it is not about low light enhancement.

I wonder
how to build this user-friendly website (using API? or something, I am learning website building with HTTP CSS JS language)
how to deal with those uploaded images? save to local drive or directly delete it? (what if there are lots of images uploaded at the same time, then how to keep this website working instead of crashing )

How to lighten black and white images with dark field noise
such as this

你好呀大佬，为啥平台现在无法进入了呢

Hi, thank you for the great work of the LLIE-Platform! It is extremely helpful.

However, I found the resolution of the output image has been downsampled for all methods. Is it possible to output the original resolution as the input?

现在平台好像无法进入了，要登录南开的VPN才可以

您好，现在平台又无法进入了，需要登陆校内VPN。

Hi, Do you know of any indicators that can evaluate the difference in illumination between two images？

Dear professor:

Thank you for providing us the excellent paper as a reference.
The table 5 in your paper published on PAMI, the results of DRBN (PSNR, SSIM, LPIPS) are 15.125, 0.472, 0.316, respectively, but in official paper(From fidelity to perceptual quality: A semi-supervised approach for low-light image enhancement, web address: https://openaccess.thecvf.com/content_CVPR_2020/papers/Yang_From_Fidelity_to_Perceptual_Quality_A_Semi-Supervised_Approach_for_Low-Light_CVPR_2020_paper.pdf）,the results are : 20.13, 0.82, 0.16).

So, the values is wrong in your paper?
Best regards,
Chris

In addition to the datasets mentioned by the author, are there any other datasets with no more than 1k images?

现在平台无法访问了，请问什么时间再次开放

麻烦大佬检查一下，目前没有办法上传图像，只能点击已有的图像进行增强，谢谢！

请问测试平台什么时候可以开放呢，谢谢

Hi Prof Li,

Thanks for your extraordinary job on low-light enhancement! When will the helpful platform be re-opened?
Looking forward to your reply!

Best wishes,

辛苦大佬有空看看有没有什么外网可以访问的方法，谢谢了

Hi, 请问大佬，最近好像平台无法上传图片进行增强了。

Hi! Thanks for the great survey!

I found the LOE (Lightness Order Error) metric is not available now. It seems that I do not have the corresponding permission. Could you please provide it? Or is there any other way I can get it?

请问LOE评测指标matlab代码怎么拒绝访问？有无参考的LOE的具体代码？Python版本和Matlab版本的都可以。

We seem to be restricted from accessing the website now.

能解决一下吗大佬

Thanks for your amazing works.

I'd like to know more about the code you used for measuring the running time.

If it is not convenient for you to release the time testing code:

1. What is the input for each model? Is it 1200*900*3 as in ZeroDCE++?
2. How did you measure the training time for each model? Did you use torch.cuda.Event for time testing, or some other methods?
3. Did you use any other ways to ensure the fairness of the comparison?

I believe that posting these would be a great help for future works.

Looking forward to your reply.

大佬，国庆前能开放服务器吗？

Hi ,
Thanks for your work. The platform covers 13 popular deep learning-based LLIE methods, where the results of any inputs can be produced through a user-friendly web interface. So what training sets were these 13 methods trained with? Did you retrain them?
Thank you for your reply.

现在平台无法访问，需要学校WebVPN，请问什么时候开放呢？