<Yun>

Hi, @tyshiwo Thanks for your papar and code, really like it. There is one thing which I don't really understand: if we use BN layer in recursive model, means we...

为什么只用YCBCR中的Y进行训练？

1

我有一个疑问，那就是我看了你很多的implementation，你都是用YCBCR的Y来训练。而且你计算PSNR也都是用一个Y channel来计算的。 那么现在我有一张图，没有GT做比对，你只用Y来行进2倍或者4倍扩大，那么剩下的两个CBCR怎么办呢？双三次扩大？那么边缘对不齐啊。按照你这个思路，最后怎么才能得到RGB图呢？ 为什么不直接RGB丢进去得到一个RGB的输出？

猪猪们，这是你们想要的东西吗？

8

这样的TNet？这样的ANet?   别想着白嫖了， 自己动手吧。 这个repo的作者已经给你们平了90%的路了，兄弟们。

Tricks also apply for ResNets?

2

Hi, Thanks for your implementation and repo. I tested on ResNet for CIFAR100 and seems Label Smoothing, No-bias-decay does not improve the result. BTW, I disable all the image augmentation...

Hi, again, thanks for your work and paper! During my search of the model, I found some of them do not have pre-calculated saliency map, like PiCANet, DSS(no DUTSTE), etc....

rgb to ycbcr formula coefficient problem.

2

hi, I checked your image transformation part in file **utils.py** the formula for rgb to ycbcr image is a little different from Wikipedia: yours: y = 16. + (**64.738** *...

Hi, if I understand the paper correctly, we input a small image and output a large image with good resolution. So I read through your code and I dont know...

why code for dataloader is this complicated???

3

first thanks for your code. but TBH, why the code for dataloader is so complex and so hard to read???

Hi Weijun, in your Emeasure.m, there is an error: https://github.com/weijun88/F3Net/blob/eecace3adf1e8946b571a4f4397681252f9dc1b8/eval/Emeasure.m#L22 no need to *2, and as a result, your Emeasure score is 0.02 lower than the correct value.

a bit confused about the theory

3

Hi, I understand you try to encode GT into high-dimension features using a model F_gt, but I just confused about the formula for GT encoder in the paper: **argmin(BCE(F_gt(theta, GT),...