Style translation for human motion

DMCA

Style translation for human motion (2005)

Cached

Download Links

Other Repositories/Bibliography

DBLP

by Eugene Hsu , Kari Pulli , Jovan Popović

Venue:	ACM Transactions on Graphics
Citations:	88 - 1 self

BibTeX

@ARTICLE{Hsu05styletranslation,
    author = {Eugene Hsu and Kari Pulli and Jovan Popović},
    title = {Style translation for human motion},
    journal = {ACM Transactions on Graphics},
    year = {2005},
    volume = {24},
    pages = {1082--1089}
}

OpenURL

Abstract

Figure 1: Our style translation system transforms a normal walk (TOP) into a sneaky crouch (MIDDLE) and a sideways shuffle (BOTTOM). Style translation is the process of transforming an input motion into a new style while preserving its original content. This problem is motivated by the needs of interactive applications which require rapid processing of captured performances. Our solution learns to translate by analyzing differences between performances of the same content in input and output styles. It relies on a novel correspondence algorithm to align motions and a linear time-invariant model to represent stylistic differences. Once the model is estimated with system identification, the system is capable of translating streaming input with simple linear operations at each frame.

Keyphrases

style translation human motion input motion captured performance linear time-invariant model sideways shuffle interactive application new style style translation system system identification output style simple linear operation normal walk sneaky crouch stylistic difference novel correspondence algorithm original content rapid processing

DMCA