Topology-agnostic method improves facial mesh animation accuracy
TopoRig: Topology-Agnostic Facial Rigging via Multi-Source Supervision
GraphicsComputer Vision and Pattern Recognition
Summary
Animating facial expressions on 3D models with different structures is tough because most teaching examples come from a single standard face shape. The authors created TopoRig, which can add facial movements directly to any face model without changing its structure. They taught it using a mix of good quality but limited examples, noisier diverse examples, and special image clues, making it better at mimicking expressions on new faces and shapes. This approach keeps the original model’s design intact and produces more natural-looking facial animations.
What this means in practice
- •For 3d character animators: Create realistic facial animations on characters with diverse mesh structures without needing to remap or simplify the mesh topology.
- •For game developers: Integrate flexible facial rigging tools to support customized character designs while preserving animation quality across different face geometries.
Authors
Andrew Fleet, Soroush Mehraban, Vida Adeli, Cole Clifford, Babak Taati
Abstract
Automatic facial rigging across heterogeneous mesh topologies remains challenging because high-quality expression supervision is often tied to canonical templates, while deformation transfer to arbitrary meshes can introduce geometric artifacts and correspondence errors. We present TopoRig, a topology-agnostic facial rigging framework that predicts FACS-conditioned deformations directly on input mesh vertices while preserving the original topology. Starting from the ICT FaceKit expression model, we construct complementary supervision from accurate but template-biased common-topology rigs, topology-diverse but noisier transferred rigs, and targeted image-based cues for controls poorly captured by geometric transfer. TopoRig combines local surface geometry, landmark-relative semantic features, global shape context, and FACS controls to predict per-vertex displacements. We train on 3,496 generated identities using 45 non-gaze expression controls from the 53-control ICT FaceKit vocabulary. On held-out identities and unseen mesh topologies, TopoRig more faithfully reproduces the reference expression space than prior neural facial-rigging methods, while qualitative results show consistent localized deformations across diverse character geometries. Ablations demonstrate that semantic landmark features and complementary supervision improve cross-identity and cross-topology generalization. Overall, TopoRig amortizes heterogeneous and imperfect expression supervision into a single topology-preserving deformation model.