Structure-Aware Editable Morphable Model for 3D Facial Detail Animation and Manipulation

Morphable models are essential for the statistical modeling of 3D faces. Previous works on morphable mostly focus large-scale facial geometry but ignore details. This paper augments in representing details by learning a Structure-aware Editable Model (SEMM). SEMM introduces detail structure representation based distance field wrinkle lines, jointly modeled with displacements to establish better correspondences and enable intuitive manipulation structure. Besides, two transformation modules translate expression blendshape weights age values into changes latent space, allowing effective semantic editing while maintaining identity. Extensive experiments demonstrate that proposed model compactly represents details, outperforms previous methods animation qualitatively quantitatively, achieves line Code available at https://github.com/gerwang/facial-detail-manipulation .