High-level multi-parameter synthesis of human animation from motion capture data

AMR AHMED*, Adrian Hilton, Farzin Mokhtarian

*Corresponding author for this work

Research output: Contribution to journalArticle (journal)peer-review


This paper presents a high-level parametric approach for synthesis of novel human animation sequences from short clips data. The parametric approach generates motions according to user-defined high-level motion parameters, such as speed, slope and direction, while maintaining the quality and realism of captured motions, and preserving the motion constraints. The approach allows simultaneous control over multiple motion parameters. This is also done with efficient computations. Animation sequences are synthesized by parametric blending, according to high-level parameters. The mapping between high-level parameters and low-level blending coefficients is automatically pre-computed, enabling efficient motion synthesis with high-level parametric control. This allows realistic synthesis of novel motions intuitively and without expert animation skills. Hence, the presented high-level parametric approach allows both novice and expert users/animators to define their desired motion using intuitive parameters. This improves the learning curve for novice users and productivity for expert users.
Original languageEnglish
Pages (from-to)156-177
JournalJournal of Intelligent Computing
Issue number4
Publication statusPublished - 1 Dec 2012

Research Centres

  • Centre for Intelligent Visual Computing Research


Dive into the research topics of 'High-level multi-parameter synthesis of human animation from motion capture data'. Together they form a unique fingerprint.

Cite this