← ArchivePaper1998
Retargeting Motion to New Characters
Spacetime constraint-based motion retargeting adapting motion capture data to new character proportions while preserving dynamic qualities.
Abstract
This paper presents a technique for retargeting motion, the problem of adapting an animated motion from one articulated character to another character with identical structure but different segment lengths. Important features of the original motion, such as feet touching the floor or hands grabbing an object, are identified as constraints that must be maintained. A spacetime constraints solver computes an adapted motion that re-establishes these constraints while preserving the frequency characteristics of the original signal, representing the changes as a motion displacement curve built from frequency-limited B-splines. The approach is demonstrated on motion capture data across characters of differing proportions, including walking, ladder climbing, and swing dancing examples.
How to read this
- Category
- Method: a motion retargeting technique
- Contributions
- Formulates retargeting as adapting motion between same-structure characters with different segment lengths while maintaining identified constraints (e.g. feet on floor, hands grabbing)
- A spacetime-constraints solver that re-establishes constraints while preserving the original signal's frequency characteristics
- Represents edits as a motion displacement curve built from frequency-limited B-splines, shown on walking, ladder climbing, and swing dancing
- Context
- Builds directly on Popovic and Witkin's Motion Warping and on Witkin and Kass's Spacetime Constraints, extending curve editing into constraint-preserving retargeting.Builds on: Motion Warping · Spacetime Constraints
- Correctness
- Assumes the source and target share identical skeletal structure differing only in proportions; it preserves kinematic constraints and signal frequency rather than enforcing physics, so dynamic realism is inherited from the source, not guaranteed.
- Clarity
- Clearly written but constraint-solver heavy; a first pass conveys the goal and approach, a second pass pays off for the optimization formulation.
- How to read it
- First pass for the constraints-plus-displacement-curve framing; second pass to follow the spacetime solve and the B-spline displacement representation if implementing retargeting.
Builds on
- Motion Warping 1995
- Spacetime Constraints 1988
Built upon by
- Deformation Transfer for Triangle Meshes 2004
- Real-Time Motion Retargeting to Highly Varied User-Created Morphologies 2008
- Capturing The Last of Us: Motion Capture Pipeline 2015
- Motion Retargeting for Crowd Simulation 2015
- Animation Bootcamp: The Animate Button: Mocap Automation Techniques at Ubisoft Montreal 2016
- Normalized Euclidean Distance Matrices for Human Motion Retargeting 2017
- Aura Mesh: Motion Retargeting to Preserve the Spatial Relationships between Skinned Characters 2018
- Surface Based Motion Retargeting by Preserving Spatial Relationship 2018
- How to Retarget Motion Capture in MotionBuilder 2019
- Learning Character-Agnostic Motion for Motion Retargeting in 2D 2019
- Interactive Multi-Character Motion Retargeting 2021
- Motion Builder Characterization and Retargeting Tutorial 2022
- The Future of Animation Retargeting in Unreal Engine | Unreal Fest Orlando 2025 2025
Related work
- Motion Warping 1995 / SIGGRAPH
- SketchiMo: Sketch-based Motion Editing for Articulated Characters 2016 / SIGGRAPH
- Normalized Euclidean Distance Matrices for Human Motion Retargeting 2017 / MIG
- Dog Code: Human to Quadruped Embodiment Using Shared Codebooks 2024 / MIG
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 →