search

EvelynFan / FaceFormer

[CVPR 2022] FaceFormer: Speech-Driven 3D Facial Animation with Transformers
MIT License
755 stars 134 forks source link

Lip-sync Evaluation #92

Open EESJGong opened 11 months ago

EESJGong commented 11 months ago

Thanks to the author for the code. But I have doubts about the way the Lip-sync Evaluation metrics are implemented in the paper. Is there any specific implementation code for this metric?

ShengnanGui commented 11 months ago

Thanks to the author for the code. But I have doubts about the way the Lip-sync Evaluation metrics are implemented in the paper. Is there any specific implementation code for this metric?

Hey bro, you can see https://github.com/EvelynFan/FaceFormer/issues/14; But I still have a question, which vertices are the lips among the 5023 vertices in the VOCASET dataset? How do I calculate Lip-sync if I don’t know this?

wangryhen commented 6 months ago

Thanks to the author for the code. But I have doubts about the way the Lip-sync Evaluation metrics are implemented in the paper. Is there any specific implementation code for this metric?

Hey bro, you can see #14; But I still have a question, which vertices are the lips among the 5023 vertices in the VOCASET dataset? How do I calculate Lip-sync if I don’t know this?

Hi, have you known how to solve the problem?(which vertices are the lips among the 5023 vertices in the VOCASET dataset? How do I calculate Lip-sync if I don’t know this?). Thank you!

Tankaishen commented 3 months ago

Thanks to the author for the code. But I have doubts about the way the Lip-sync Evaluation metrics are implemented in the paper. Is there any specific implementation code for this metric?

Hey bro, you can see #14; But I still have a question, which vertices are the lips among the 5023 vertices in the VOCASET dataset? How do I calculate Lip-sync if I don’t know this?

Hi, have you known how to solve the problem?(which vertices are the lips among the 5023 vertices in the VOCASET dataset? How do I calculate Lip-sync if I don’t know this?). Thank you!

You can refer to the CodeTalker: https://github.com/Doubiiu/CodeTalker/issues/64