Skip to content

← ArchivePaper2023

SoftDECA: Computationally Efficient Physics-Based Facial Animations

Nicolas Wagner, Mario Botsch, Ulrich Schwanecke

MIGAcademic14 citesFacialMuscles

Computationally efficient physics-based facial animation method integrating soft tissue dynamics with the DECA morphable face model.

Abstract

Facial animation on computationally weak systems is still mostly dependent on linear blendshape models. However, these models suffer from typical artifacts such as loss of volume, self-collisions, or erroneous soft tissue elasticity. In addition, while extensive effort is required to personalize blendshapes, there are limited options to simulate or manipulate physical and anatomical properties once a model has been crafted. Finally, second-order dynamics can only be represented to a limited extent. For decades, physics-based facial animation has been investigated as an alternative to linear blendshapes but is still cumbersome to deploy and results in high computational cost at runtime. We propose SoftDECA, an approach that provides the benefits of physics-based simulation while being as effortless and fast to use as linear blendshapes. SoftDECA is a novel hypernetwork that efficiently approximates a FEM-based facial simulation while generalizing over the comprehensive DECA model of human identities, facial expressions, and a wide range of material properties that can be locally adjusted without re-training. Along with SoftDECA, we introduce a pipeline for creating the needed high-resolution training data. Part of this pipeline is a novel layered head model that densely positions the biomechanical anatomy within a skin surface while avoiding self-intersections.

How to read this

Category
Method: efficient physics-based facial animation
Contributions
  • A hypernetwork (SoftDECA) that approximates FEM-based facial simulation at speeds comparable to linear blendshapes
  • Generalization over the DECA model of identities, expressions, and a range of materials, with locally adjustable properties without retraining
  • A pipeline for creating the high-resolution data the approach needs
Context
Builds on the DECA morphable face model (Feng et al.) to bring physics-based soft-tissue behavior to blendshape-speed runtimes.Builds on: Learning an Animatable Detailed 3D Face Model from In-The-Wild Images
Correctness
It assumes a learned hypernetwork can faithfully stand in for FEM simulation across identities and materials; presented as addressing blendshape artifacts (volume loss, self-collision, limited dynamics), though fidelity to true FEM and behavior outside the training distribution warrant scrutiny.
Clarity
Moderately technical; a first pass conveys the blendshape-versus-physics motivation and the hypernetwork idea, a second pass clarifies the FEM approximation and data pipeline.
How to read it
Focus on what the hypernetwork approximates and which blendshape artifacts it removes; a second pass on the FEM setup and material parameterization pays off if you build facial rigs.

Builds on

Built upon by

Nothing yet.

Related work

Keywords

This page summarises the entry and links to its original source. The archive never hosts or redistributes the publication itself.Show it in the full archive list →