I am currently working on a project that trying to reconstruct 3D human models from skeleton information. It seems the ZED SDK’s body tracking module can do this. However, the key points in the human face are very small, with only 5: Body Tracking Overview | Stereolabs
In my project, I need a detailed face after reconstruction, with expressions. Thus, I need very more key points in the human face.
Is there any method/model that can extract more key points from the human face?