← ArchivePaper1998
Verbs and Adverbs: Multidimensional Motion Interpolation
Radial basis function interpolation framework for blending motion clips along semantic dimensions like speed and emotion for interactive characters.
Abstract
This paper presents a technique for interpolating between example motions of complex linked figures, derived from motion capture or traditional animation tools, to create real-time controllable character animation. Parameterized motions called verbs are controlled by parameters called adverbs, and a combination of radial basis functions and low order polynomials builds the interpolation space between example motions. Inverse kinematic constraints augment the interpolations to prevent artifacts such as feet slipping on the floor during a walk cycle. Verbs are assembled into a verb graph with smooth transitions, allowing an animated figure to exhibit a repertoire of expressive behaviors driven interactively or programmatically at runtime.
How to read this
- Category
- Method: multidimensional motion interpolation
- Contributions
- Parameterized motions (verbs) controlled by semantic parameters (adverbs), interpolated from example motions
- An interpolation space built from radial basis functions plus low-order polynomials, augmented with inverse-kinematic constraints to avoid artifacts like foot slipping
- A verb graph with smooth transitions for an interactive, expressive runtime repertoire
- Context
- Connects to the interactive-actor lineage (Perlin and Goldberg's Improv), turning example motions into a continuously controllable, runtime-driven blend space.Builds on: Improv: A System for Scripting Interactive Actors in Virtual Worlds
- Correctness
- Quality depends on having representative example motions spanning the adverb axes; radial-basis interpolation can extrapolate poorly outside the examples, and IK is used to patch kinematic artifacts rather than to enforce dynamics.
- Clarity
- Accessible with clear terminology; a first pass conveys the verb/adverb framing, a second pass pays off for the RBF interpolation construction.
- How to read it
- First pass for the verb/adverb/verb-graph concepts; second pass to understand the radial-basis-plus-polynomial interpolation and the IK constraints if building a blend system.
Built upon by
Related work
- Geostatistical Motion Interpolation 2005 / SIGGRAPH
- Automated Extraction and Parameterization of Motions in Large Data Sets 2004 / SIGGRAPH
- Authoring Motion Cycles 2017 / SCA
- Motion Warping 1995 / 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 →