junlinhan / bid Goto Github PK

你好，我想请问您在Task2中，为什么在叠加雨、雾、雪、雨滴等图层时，要将图像先加上一再除二？我实验后似乎不能得到与论文中类似的图像，但如果不采用这一步骤则可以得到类似的图像。

Hi @JunlinHan,

When performing the test with the following command:

python test.py --dataroot ./datasets/raina --name task2a --model raina --dataset_mode raina --test_input B

--test_input B means I'm only testing with the rain noise cases? Use only the weight of that specific case?

(I'm sorry for this primary question, I would really like to understand about your project)

Thank you!

I apologise to ask in this way here,

I am working on a 360 film for an exhibition which has a lot of reflections in the glass of the opposite of the room which of course also part of the 360 recording.

I got your fascinating code running but i am bit confused how to proceed with my own image data.

Do You think Your Project is cape able to take glass reflections away?

Do I need to train my own model for it?

Can the jointremoval model train models of arbitrary resolution?
What parameters need to be configured?

Hi, authors! I am writing a survey paper on image deraining and am very interested in your BID dataset. Could you tell me how many training and testing images are included in the BID dataset?

Hi! Nice work!
To change the resolution on test dataset, I follow your advice to change "--preprocess" option to "none". But I meet some errors when processing the input image.

I think the possible reason is the input image and rain-steak&snow have different resolutions. So could you give me some advice on how to deal with this situation and get the similar visual results as illustrated in your paper. Thanks a lot!

In the test stage, the size of the output image is 256*256, how to set the size of the output image to the original size?

Hello！Thank you for your excellent work！
I have a question about the test data：
About Task II A. May I ask what is the size of the test data displayed in the paper?

抱歉 请问您的代码可以进行单张图片的阴影去除么 比如只想用一张图片去测试 貌似也要建立十几个文件夹

Hello @JunlinHan!
What CUDA version did you use?