Research on Face Pose Estimation Method for Virtual Try-on System

Abstract

This paper proposes a 3D face pose estimation algorithm based on feature points, which realizes the angle estimation of human faces in any pose. It can accurately estimate within 30 degrees on 6 degrees of freedom, and provides the algorithm basis for a subsequent real 3D try-on.