Abstract
dc:description.abstractComputer Animation is constantly extending the range of activities that virtual actors can perform. One of the areas almost untouched previously is performing arts, within which the requirements on motions are not easily fulfilled by most popular animation methods, although keyframing and motion capture can be used, but at a cost of considerable human intervention. 1 his thesis considers one such activity - dancing (ballroom dancing, in particular). Ballroom dancing as a simulated activity poses a number of challenges. The two that this thesis focuses on are style and interaction. Both are important not only in dancing, but also in a number of other areas, and neither has prompted much research interest before. As Character Animation does not contain any methods that would be immediately applicable to these problems, other domains are explored for approaches. The statistical analysis approach is taken from the Artificial Intelligence domain. Weighted-directed graphs are used to capture the style properties of the motion-captured sequences, and to create motion fragments that are consistent with that style. A similar technique is used to solve the recognition problem during interaction. Weightcd-directed graphs are used to store and recognise the motion patterns. A concept of control points is introduced to reduce the dimensionality of the recognition task by disregarding irrelevant data. It was not possible to model the ballroom dancing scenario it its entirety, and the methods have been tested on simplified scenarios, which, nevertheless, retain the important properties of ballroom dancing.
Degree
thesis:*- Name dc:type.qualificationname
- PhD
- Level dc:type.qualificationlevel
- Doctoral
- Grantor dc:publisher.institution
- De Montfort University
- Year dc:date.issued
- 2003
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Pugach, Stanislav