Learning from Examples and Comparing Models of Human Motion
Marek Kulbacki*, Bartosz Jablonski**, Ryszard Klempous**, and Jakub Segen***
*Systems Research Institute, Polish Academy of Sciences, 6 Newelska St, 01-447 Warsaw, Poland
**Institute of Engineering Cybernetics, Wroclaw University of Technology, 11/17 Janiszewskiego St, 50-372 Wroclaw, Poland
***Consultant, 337 Third St, Fair Haven NJ 07704, USA
This paper addresses a problem of creating character animations from motion capture clips. The main problem we want to solve is partition set of primitive motions into appropriate groups according to similarity between motions. We construct motion models to easier extract features of given motions and make animation process more flexible. Using these models we propose measure of discrepancy between motions. Moreover it normalizes length of motions and decreases high dimension of considered motion data, so clustering may take place in dimensionally reduced space. In addition we examine different motion representations for the sake of the best clustering results.
This article is published under a Creative Commons Attribution-NoDerivatives 4.0 Internationa License.