Skip to content

← ArchivePaper1987

A Muscle Model for Animating Three-Dimensional Facial Expression

Keith Waters

SIGGRAPHAcademic14 descendantsFacialMuscles

Parameterized face muscle model using linear and sphincter muscle types to animate facial expressions, influencing decades of facial animation research.

Abstract

Develops a parameterized muscle model for three-dimensional facial animation that controls facial expressions through muscle vectors rather than hard-wired actions. Uses action units from the Facial Action Coding System to define muscle operations with parameters for zone of influence, falloff radius, and spring constants. The model separates linear muscles that pull from sphincter muscles that squeeze, enabling animation of diverse facial types without topology-specific constraints.

How to read this

Category
Method: a parameterized muscle model for facial animation
Contributions
  • A parameterized muscle model that controls expressions through muscle vectors rather than hard-wired actions
  • Two muscle types, linear muscles that pull and sphincter muscles that squeeze, defined via FACS action units
  • Parameters for zone of influence, falloff radius, and spring constants that generalize across facial types without topology-specific constraints
Context
Builds directly on Platt and Badler's muscle-based facial animation (Animating Facial Expressions, 1981) and continues the FACS-driven line of facial deformation.Builds on: Animating Facial Expressions
Correctness
Demonstrated to animate diverse face shapes by abstracting muscles as parameterized vectors; a reader should note it is a procedural, geometry-driven approximation of muscle action rather than a physically simulated tissue model.
Clarity
Accessible and well structured; a first pass conveys the linear/sphincter distinction, a second pass repays the parameter definitions.
How to read it
Read for the durable linear-vs-sphincter muscle abstraction: focus on the muscle-vector parameters (influence zone, falloff, spring constant); a second pass pays off if you want to reimplement the deformation.

Builds on

Built upon by

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 →