← ArchivePaper2017
Facial Retargeting with Automatic Range of Motion Alignment
Roger Blanco i Ribera, Eduard Zell, J. P. Lewis, Junyong Noh, Mario Botsch
Automatically aligns source and target blendshape ranges of motion for retargeting, reducing artist intervention while preserving intended expression style.
Abstract
While facial capturing focuses on accurate reconstruction of an actor's performance, facial animation retargeting has the goal to transfer the animation to another character, such that the semantic meaning of the animation remains. Because of the popularity of blendshape animation, this effectively means to compute suitable blendshape weights for the given target character. Current methods either require manually created examples of matching expressions of actor and target character, or are limited to characters with similar facial proportions (i.e., realistic models). In contrast, our approach can automatically retarget facial animations from a real actor to stylized characters. We formulate the problem of transferring the blendshapes of a facial rig to an actor as a special case of manifold alignment, by exploring the similarities of the motion spaces defined by the blendshapes and by an expressive training sequence of the actor. In addition, we incorporate a simple, yet elegant facial prior based on discrete differential properties to guarantee smooth mesh deformation. Our method requires only sparse correspondences between characters and is thus suitable for retargeting marker-less and marker-based motion capture as well as animation transfer between virtual characters.
How to read this
- Category
- Method: a facial animation retargeting algorithm
- Contributions
- Automatic alignment of source actor and target character ranges of motion, framing retargeting as a special case of manifold alignment between blendshape motion spaces
- A facial prior based on discrete differential properties to keep mesh deformation smooth
- Retargeting to stylized (non-realistic) characters from sparse correspondences and an expressive actor training sequence
- Context
- Sits in the blendshape facial retargeting lineage, extending the goal of artist-friendly transfer raised by Seol et al.'s Artist Friendly Facial Animation Retargeting toward automatic, proportion-agnostic alignment.Builds on: Artist Friendly Facial Animation Retargeting
- Correctness
- Demonstrated for transferring expressions to stylized targets using only sparse correspondences and a training sequence; reader should note results hinge on how well the actor's expressive sequence spans the motion space and that semantic fidelity to artistic intent is the success criterion, not metric accuracy.
- Clarity
- Accessible at the conceptual level; a first pass conveys the manifold-alignment idea, a second pass is needed for the alignment formulation and the differential prior.
- How to read it
- Focus first on the manifold-alignment framing and what 'range of motion' means here; do a second pass on the alignment math and smoothness prior if you intend to implement or compare against it.
Builds on
Built upon by
Nothing yet.
Related work
- Transferring Facial Expressions to Different Face Models 2006 / SIACG
- Transferring the Rig and Animations from a Character to Different Face Models 2008 / CGF
- Neural Face Rigging for Animating and Retargeting Facial Meshes in the Wild 2023 / SIGGRAPH
- Realtime Facial Animation with On-the-fly Correctives 2013 / SIGGRAPH
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 →