VISUALIZATION OF MASSIVE MOTION DATA USING GEOMETRIC FEATURES(International Workshop on Advanced Image Technology 2006)

スポンサーリンク

概要

• 論文の詳細を見る

• This research proposes a visualization mechanism of massive motion data for efficiently retrieving motion clips with key poses. Our previous method generates self-organizing maps by using a pose distance metric whose value deeply depends on skeletal structures of performers. Therefore, it is inadequate to integrate various motion data sets captured with different skeletal definitions. To overcome this problem, we introduce geometric features that represent qualitative metric of each pose with geometric relations between specified body locations. This approach enables integrated display of multiple motion maps while preserving consistency of qualitative relations between poses. Our map organization therefore can decompose massive motion data into categorized subsets in a layered or hierarchical manner. Such multiple view of the maps increase the efficiency in selecting key-poses.

• 一般社団法人電子情報通信学会の論文
• 2006-01-02

著者

• Kuriyama Shigeru

  Department Of Information And Computer Sciences Toyohashi University Of Technology:digital Human Res

• Yamaguchi Naoaki

  Department of Information and Computer Sciences Toyohashi University of Technology

• KURIYAMA Shigeru

  Department of Gastroenterology, Yaizu Municipal Hospital

関連論文

• Extensible Task Simulation with Motion Archive(Cyberworlds)
• VISUALIZATION OF MASSIVE MOTION DATA USING GEOMETRIC FEATURES(International Workshop on Advanced Image Technology 2006)

スポンサーリンク