Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Margin Loss #2

Open
lovekittynine opened this issue Nov 11, 2021 · 1 comment
Open

Margin Loss #2

lovekittynine opened this issue Nov 11, 2021 · 1 comment

Comments

@lovekittynine
Copy link

Hi, thanks for your great work! I have a question about Margin loss! Duration the training process of Identity Representation, margin loss is used to prevent the obtained mean face far from face image. But, in the training process of Expression Representation, why not use margin loss for neural face? In my opinion, both process are somewhat similar.

@JiaRenChang
Copy link
Owner

Hi, @lovekittynine :
We don't use margin loss in training expression representation because that the input faces contain various poses and expression.
In contrast, the head pose of input face in training identity representation are always front.
The margin loss constrains the generated mean face which to also be front.

Jia-Ren

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

2 participants