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

How to get training data? Why is the betas parameter is 200? #41

Open
Timkeeper2018 opened this issue Jun 6, 2022 · 2 comments
Open

Comments

@Timkeeper2018
Copy link

As you mentioned in your paper, the training data is SMPL-X fit of exPose. But the number of betas parameters given by exPose is 10, but the number of betas parameters given in your method is 200, I was wondering how you guys Get the SMPL-X dataset with betas parameter of 200? (I see that most of the methods set the number of betas parameter of SMPL-X to 10)

@Timkeeper2018 Timkeeper2018 changed the title How to get training data? Why is the betas parameter 200? How to get training data? Why is the betas parameter is 200? Jun 6, 2022
@MichaelJBlack
Copy link
Collaborator

Indeed, most people use 10, but this gives quite a coarse shape. If one is fitting to 2D joint locations, 10 is enough because there is very little information about body shape in a few 2D joints. The AMASS dataset uses 16. We make 300 available for download. People can choose how many they need for their application -- with 300, you can get very high-fidelity body shape but it is hard to reliably estimate 300 from images.

@Timkeeper2018
Copy link
Author

Wait, does that mean you have a training dataset with betas of 300? However, I did not find the download option of the relevant training dataset on your project page, so kind of you if you can share the training dataset. If it is not convenient for public disclosure, this is my email: [email protected].
Thanks a lot

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