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

Ahmed, Amr and Hiton, Adrian and Mokhtarian, Farzin (2012) High-level multi-parameter synthesis of human animation from motion capture data. Journal of Intelligent Computing, 3 (4). pp. 156-177. ISSN 0976-9005

Full content URL: http://www.dline.info/jic/v3n4.php

Documents
High-level Multi-Parameter Synthesis of Human Animation from Motion Capture Data - J Intelligent Computing-Dec2012.pdf
[img] PDF
High-level Multi-Parameter Synthesis of Human Animation from Motion Capture Data - J Intelligent Computing-Dec2012.pdf
Restricted to Repository staff only

1MB
Item Type:Article
Item Status:Live Archive

Abstract

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.

Another novel contribution is the introduction of quantitative, in addition to the qualitative, evaluation of the results which shows that synthesized animations are perceptually and quantitatively indistinguishable from captured animations. Example applications demonstrating the proposed approach are also presented.

Keywords:Human animation, Character Animation, motion capture, motion control, motion synthesis, motion editing, mocap, motion interpolation
Subjects:G Mathematical and Computer Sciences > G490 Computing Science not elsewhere classified
G Mathematical and Computer Sciences > G400 Computer Science
G Mathematical and Computer Sciences > G450 Multi-media Computing Science
Divisions:College of Science > School of Computer Science
Related URLs:
ID Code:9406
Deposited On:11 May 2013 19:23

Repository Staff Only: item control page