Secondary motion in contour-driven rig inversion

Extend the contour-driven rig-parameter inversion method to production rigs with physics-driven secondary motion, including hair dynamics, cloth folds settling under gravity, and soft-body jiggle, by representing deformation dependence on the trajectory through pose space rather than only on current rig-parameter values.

Background

The presented NeuralRig is a static differentiable map from rig parameters to mesh-vertex positions. This representation is adequate for deformations determined by the current pose, but it cannot model physics-driven effects whose state depends on the sequence of poses and motion history. Consequently, the current contour-based inversion pipeline does not support production rigs that include simulators in the deformation loop.

The paper explicitly identifies extending the method to such secondary motion as unresolved. Solving this problem would enable artist-drawn contour edits to control characters with dynamic hair, cloth, or soft-tissue behavior while preserving the method’s rig-level optimization interface.

References

Extending the method to secondary motion is an open problem. The NeuralRig is currently a static map from rig parameters to vertex positions, but physics-driven secondary motion---hair dynamics, cloth folds settling under gravity, soft-body jiggle---is shaped by the trajectory the rig has taken through pose space rather than by the current parameter values.

— Inverse Rig Optimization from Line Drawings  (2609.00732 - Zhu et al., 1 Sep 2026) in Section 5, Discussion, subsection “Future Work,” paragraph “Support for Secondary Motion”