← ArchivePaper2024
Factorized Motion Diffusion for Precise and Character-Agnostic Motion Inbetweening
Justin Studer, Dhruv Agrawal, Dominik Borer, Seyedmorteza Sadat, Robert W. Sumner, Martin Guay, Jakob Buhmann
Diffusion model for motion inbetweening that factorizes motion into character-agnostic and character-specific components for precise controllability.
Abstract
Animation is a challenging and time-consuming process where animators must manipulate hundreds of controls over space and time to create compelling motions. Recent advances in motion diffusion models have shown impressive results for general motion generation and hold the potential to reduce the number of controls manipulated by animators to achieve high quality results. However, these models are limited by their inability to match sparse constraints precisely, preventing frame-level joint control required by artists. Additionally, recent models are trained for specific characters, preventing reuse, and are incompatible for characters with only a small datasets available. To tackle these shortcomings, we propose a novel factorization of motion between a character-agnostic Bézier Motion Model (BMM), which can be trained on a large motion dataset, followed by a character-specific posing model, trainable on a much smaller pose dataset, that enables reuse across many characters. BMM provides accuracy for meeting sparse joint-level constraints by working in a reduced space of Bézier curves that better aligns the condition signal with the prediction space of our model. Additionally, the Bézier curves offer animators an intuitive interface compatible with existing authoring software.
How to read this
- Category
- Method: motion diffusion model for inbetweening
- Contributions
- A factorization of motion into a character-agnostic Bezier Motion Model (BMM), trainable on a large dataset, and a character-specific posing model trainable on a much smaller pose dataset.
- Working in a reduced space of Bezier curves to better align sparse joint-level constraints with the model's prediction space, improving precision at meeting sparse constraints.
- Reuse across many characters, including those with only small datasets available.
- Context
- Sits in the line of learned motion inbetweening and motion diffusion, building on robust motion in-betweening (Harvey et al., 'Robust Motion In-Betweening') while addressing diffusion models' difficulty matching sparse constraints precisely.Builds on: Robust Motion In-Betweening
- Correctness
- Motivated by the artist need for frame-level joint control that general diffusion models miss; the factorization assumes a useful split between character-agnostic curves and per-character posing, and a reader should note that quality on novel characters depends on the small per-character pose data available.
- Clarity
- Accessible framing of the problem; a first pass conveys the factorization idea, a second pass clarifies the Bezier reduced space and the diffusion conditioning.
- How to read it
- First pass for the agnostic-vs-specific factorization and why Bezier curves help constraint matching; second pass for the model conditioning and posing stage if applying it to your own character rigs.
Builds on
Built upon by
Nothing yet.
Related work
- SKEL-Betweener: a Neural Motion Rig for Interactive Motion Authoring 2024 / SIGGRAPH Asia
- Dog Code: Human to Quadruped Embodiment Using Shared Codebooks 2024 / MIG
- Improv: A System for Scripting Interactive Actors in Virtual Worlds 1996 / SIGGRAPH
- Character Motion Synthesis by Topology Coordinates 2009 / CGF
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 →