Abstract
We present a Matlab toolbox for synthesis and visualization of human motion style. The aim is to support development of expressive virtual characters by providing implementations of several style related motion synthesis methods thus allowing side-by-side comparisons. The implemented methods are based on recorded (captured or synthetic) motions, and include linear motion interpolation and extrapolation, style transfer, rotation swapping per body part and per quaternion channel, frequency band scaling and swapping, and Principal/Independent Component Analysis (PCA/ICA) based synthesis and component swapping.
Author supplied keywords
Cite
CITATION STYLE
Förger, K., & Takala, T. (2015). A motion style toolbox. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 9238, pp. 343–347). Springer Verlag. https://doi.org/10.1007/978-3-319-21996-7_36
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
