Individualising Human Animation Models

Loading...
Thumbnail Image
Date
2001
Journal Title
Journal ISSN
Volume Title
Publisher
Eurographics Association
Abstract
Realistic human body animation requires accurate body geometry, realistic textures and believable behaviours. Existing human animation models provide excellent tools to control articulated motion and body surface deformations according to body postures, but modelling a specific individual requires a large degree of skill to sculpt a model resembling that individual. The advent of 3D imaging collection techniques means that the highly accurate 3D surface of a specific person can now be collected in a matter of milliseconds, but most captured 3D surfaces have little semantic information. Consequently, a significant degree of laborious manual intervention is still required to produce a virtual character from their 3D image using traditional animation approaches. The authors present a method for combining 3D images of real people with existing nonspecific, i.e. generic, human animation models to achieve highly realistic human animation models of the 3D imaged individuals. This has been achieved by developing a two-step method for conforming a generic human animation model to fit or “clone” the 3D geometry captured from a specific individual. A segmentation procedure is first applied to the 3D imaged human body surfaces to impose a semantic labelling of the data into body surface parts. Having thereby established correspondence between the segments of the scanned human body and those of the generic model, a rigid body transformation followed by a global 3D mapping is applied to bring the generic mesh into approximate alignment with the 3D imaged model. The second step comprises a conformation procedure driven by (minimising) the distance between the closestpoints between the respective meshes to bring the generic model surfaces into close alignment with those of the real-world 3D surfaces.
Description

        
@inproceedings{
10.2312:egs.20011044
, booktitle = {
Eurographics 2001 - Short Presentations
}, editor = {}, title = {{
Individualising Human Animation Models
}}, author = {
Ju, Xiangyang
and
Siebert, J. Paul
}, year = {
2001
}, publisher = {
Eurographics Association
}, ISSN = {
1017-4656
}, ISBN = {}, DOI = {
10.2312/egs.20011044
} }
Citation