| Abstrakt: | The introduction of this work deals with the most widely used
methods of motion capture data synthesis. Next, we present method called
'Timewarp curve' derived from 'Registration curves' proposed by Lucas Ko-
var and Michael Gleicher and make our modification of this method, which
should speed up motion data synthesis. We implement mentioned method
and compare it with another in our software product (Motion Blender). As
the result, we get the comparison of eciency, speed and reliability of those
algorithms. We change original 'Timewarp curve' and get a novel algorithm
that preserves realism of human motion. In the end, we describe the struc-
ture, features and usability of the program and show the results of experi-
ments.
|
|---|