Data-driven online motion analysis

Author

Huang, Tianyu ; Yang, Jia ; Li, Lijie

Author_Institution

Sch. of Software, Beijing Inst. of Technol., Beijing, China

fYear

2009

fDate

26-29 Nov. 2009

Firstpage

1407

Lastpage

1411

Abstract

The paper proposed an online analysis for motion captured motion data. Traditional manifold-based motion analysis can only perform in offline mode, which limits the applications and developments of the data-driven computer animation. By reserving the eigenspace and reconstructing the space using only the new instance, online motion analysis can be achieved. Numerical new training instances can be extended while maintaining the reasonable dimensions. The experiments results demonstrated the feasibility and merits.

Keywords

computer animation; image motion analysis; principal component analysis; data-driven computer animation; data-driven online motion analysis; incremental online principal component analysis; manifold-based motion analysis; Algorithm design and analysis; Animation; Application software; Humans; Joints; Motion analysis; Performance analysis; Principal component analysis; Skeleton; Space technology; Incremental online PCA; Principal Component Analysis; captured motion data; eigenspace reconstruction; online motion analysis;

fLanguage

English

Publisher

ieee

Conference_Titel

Computer-Aided Industrial Design & Conceptual Design, 2009. CAID & CD 2009. IEEE 10th International Conference on

Conference_Location

Wenzhou

Print_ISBN

978-1-4244-5266-8

Electronic_ISBN

978-1-4244-5268-2

Type

conf

DOI

10.1109/CAIDCD.2009.5374981

Filename

5374981