Cui Ziteng

Topic: Computational Photographt, Low-light Object Detection

It seem you haven't load pre-train model or just train on 1 GPUs and didn't adjust learning rate. Here is my results: +---------------+-------+-------+ | Class | IoU | Acc |...

底层视觉通常不用bn

Hi! Thank you, this paper is still under-review, I'll release the code as soon as possible~

Segmentation code uploaded, and our paper also renew the segmentation results in https://arxiv.org/abs/2205.14871 , please check~

I get your problem, let me write a code for demo.py.

Hello, the Reason you could not use the original data-loader is that, that data-loader only used for the exposure correction dataset in paper. I've add a new file, you could...

Hello, If you only have high light image, I think you could translate it to it's pair low-light data by using some function, such as gamma correction (A*(x**gamma)), and the...

I got no idea about that, so you can use the simple way to generate low-light data.

Thanks for the explanation, I'll follow that setting.