Why are the features extracted by VGG not used as loss terms

[shi991027]

Thank you very much for your work, but I found that in the training code for the Light-Effects Suppression section,ENHANCENET.py, it does not use VGG to extract features from the original and generated images and use this as a loss term, which is mentioned in the paper, why is this?

[jinyeying]

Thank you very much for your work, but I found that in the training code for the Light-Effects Suppression section,ENHANCENET.py, it does not use VGG to extract features from the original and generated images and use this as a loss term, which is mentioned in the paper, why is this?

refer to PerceptualLossVgg16ExDark(nn.Module) in demo.py