How to get the pose boundary?

genforce / interfacegan

[CVPR 2020] Interpreting the Latent Space of GANs for Semantic Face Editing

https://genforce.github.io/interfacegan/

MIT License

1.51k stars 281 forks source link

How to get the pose boundary? #29

Closed BERYLSHEEP closed 4 years ago

BERYLSHEEP commented 4 years ago

Hello! I am confused about how to get the pose boundary. The paper said that the auxiliary attribute prediction model predicts the 5-point facial landmarks. I find the label about the facial landmarks is coordinate in the CelebFace Dataset. The paper said that turning right is the corresponding positive direction. How did you turn the coordinate into binary code that represents the direction?

ShenYujun commented 4 years ago

I think there should be many feasible ways. For example, you can compute the midpoint of two eyes, and see where it locates in the image. Taking the middle vertical line as a reference, you can get a binary label :). Also, you can also try to estimate the yaw pose based on the five landmarks with some basic geometry and use angle 0 for separation. Or, you can try something else. Good luck!