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

Does pretrained model can segment the lens of eyeglasses too? #10

Open
phuongpropk05 opened this issue Oct 11, 2024 · 1 comment
Open

Comments

@phuongpropk05
Copy link

I want to remove the léns of sunglasses too, how can i do that. I hope you can help me soon

@StoryMY
Copy link
Owner

StoryMY commented Oct 11, 2024

Our work is not good at removing the lenses of eyeglasses (see Limitations section in our paper for more details).

One way to compel the networks is to replace our glass mask with the complete glass mask (e.g., detected by face-parsing), which includes the lenses region. However, I guess the final result would be really bad as it is difficult to recover eye region under such serious occlusion.

I recommend to use generative models to solve this problem, but I am not sure which work is the SOTA. Maybe you can conduct some research on GAN and Diffussion Model. In my impression, many related works are good at inpainting task. I hope these works can help you.

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