← ArchivePaper2010

Example-Based Facial Rigging

SIGGRAPHAcademic253 cites3 descendantsFacial Rigging

Optimization method for constructing blendshape facial rigs directly from example expression scans using sparse deformation decomposition.

Open the paper ↗Trace in the lineage graph

Abstract

Introduces a method for generating facial blendshape rigs from a set of example poses of a CG character. The system transfers controller semantics and expression dynamics from a generic template to the target blendshape model while solving for an optimal reproduction of the training poses, enabling a scalable design process where additional poses iteratively refine the expression space. Plausible animations are obtained even from a single training pose, and formulating the optimization in gradient space yields superior results to a direct vertex optimization.

How to read this

Category: Method: optimization for building blendshape facial rigs from example scans
Contributions: An optimization that constructs a blendshape facial rig from a set of example poses, transferring controller semantics and expression dynamics from a generic template to the target
A scalable, iterative workflow where additional poses refine the expression space, with plausible animation obtainable even from a single training pose
A gradient-space formulation of the optimization that yields better results than direct vertex optimization
Context: Builds on deformation transfer for triangle meshes (Sumner et al.) to carry template rig semantics onto a new character's example poses.Builds on: Deformation Transfer for Triangle Meshes
Correctness: Assumes a generic template's controller semantics and dynamics transfer meaningfully to the target; readers should keep in mind that single-pose results are 'plausible' rather than exact and that quality scales with the number and coverage of example poses provided.
Clarity: Accessible; a first pass conveys the example-driven rig construction idea, do a second pass for the optimization objective and the gradient-space formulation.
How to read it: Read the pipeline and the role of deformation transfer first; a second pass on the gradient-space optimization is worthwhile if you plan to build rigs from scans.

Builds on

Deformation Transfer for Triangle Meshes 2004

Built upon by

Realtime Performance-Based Facial Animation 2011

Related work

Reusable Facial Rigging and Animation: Create Once, Use Many 2007 / PhD Thesis
Direct Manipulation Blendshapes 2010 / TOG
Smooth Contact-Aware Facial Blendshapes Transfer 2013 / DigiPro
Transferring the Rig and Animations from a Character to Different Face Models 2008 / CGF

Keywords

example-based blendshapes facial rigging blendshape model expression transfer controller semantics expression dynamics gradient space optimization training poses deformation transfer facial animation blendshape weights generic template performance capture facial expressions

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 →