← ArchivePaper2003

Motion Synthesis from Annotations

Okan Arikan, David A. Forsyth, James F. O'Brien

User paints a timeline with semantic annotations and the system assembles motion-capture frames via dynamic programming to satisfy them.

Open the paper ↗Trace in the lineage graph

Abstract

This paper describes a framework that allows a user to synthesize human motion while retaining control of its qualitative properties. The user paints a timeline with annotations --- like walk, run or jump --- from a vocabulary which is freely chosen by the user. The system then assembles frames from a motion database so that the final motion performs the specified actions at specified times. The motion can also be forced to pass through particular configurations at particular times, and to go to a particular position and orientation. Annotations can be painted positively (for example, must run), negatively (for example, may not run backwards) or as a don't-care . The system uses a novel search method, based around dynamic programming at several scales, to obtain a solution efficiently so that authoring is interactive. Our results demonstrate that the method can generate smooth, natural-looking motion.The annotation vocabulary can be chosen to fit the application, and allows specification of composite motions (run and jump simultaneously, for example). The process requires a collection of motion data that has been annotated with the chosen vocabulary.

How to read this

Category: Method: annotation-driven motion synthesis from mocap
Contributions: A framework where a user paints a timeline with a freely chosen annotation vocabulary (walk, run, jump) and the system assembles matching frames from a motion database
Support for positive, negative, and don't-care annotations plus positional and configuration constraints, including composite simultaneous motions
A multi-scale dynamic-programming search that solves the assembly interactively for smooth, natural-looking motion
Context: Builds directly on Kovar et al.'s Motion Graphs, adding a semantic annotation layer and a dynamic-programming search so users specify qualitative content rather than just paths.Builds on: Motion Graphs
Correctness: Validated by smooth, natural-looking results that satisfy painted constraints; reader caveat is that it requires a motion database pre-annotated with the chosen vocabulary, so coverage and label quality bound what can be synthesized.
Clarity: Accessible; a first pass conveys the paint-the-timeline workflow, a second pass clarifies the multi-scale dynamic programming.
How to read it: Focus on the annotation interface and the multi-scale dynamic-programming search; a second pass is worth it for the search formulation and how constraints are encoded.

Builds on

Motion Graphs 2002

Built upon by

Nothing yet.

Related work

A Deep Learning Framework for Character Motion Synthesis and Editing 2016 / SIGGRAPH
Easy Generation of Facial Animation Using Motion Graphs 2017 / CGF
Dog Code: Human to Quadruped Embodiment Using Shared Codebooks 2024 / MIG
Physically Based Motion Transformation 1999 / SIGGRAPH

Keywords

motion synthesis annotations motion database dynamic programming support vector machine combinatorial optimization k-means clustering frame constraints position constraints continuity motion graphs principal component analysis radial basis function timeline painting human motion data-driven motion hierarchical search

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 →