Generating Statistically Plausible Body Shape Variations for Perception Studies

Veröffentlicht 2016

We present a novel method for morphing 3D body scans according to semantic constraints 
for generating realistic variations of existing body shapes. By non-linearly optimizing the 
body shape in a statistical shape space, we ensure that user defined constraints are fulfilled 
while the main characteristics of a body are still maintained. We demonstrate the quality of 
the proposed algorithm on various real world use cases.


Quelle

Mensch und Computer 2016 - Tagungsband
2016;
Mensch und Computer 2016 (04.09.2016-07.09.206)
Aachen

Autoren


Herausgeber

  • Gesellschaft für Informatik e.V.