Perceptual Influences by the Revision of the Motion Captured Data of Dancing

2019 Fiscal Year Final Research Report

• Project/Area Number: 16K13185
• Research Category: Grant-in-Aid for Challenging Exploratory Research
• Allocation Type: Multi-year Fund
• Research Field: Art at large
• Research Institution: Kyoto City University of Arts
• Principal Investigator: Tsuzaki Minoru 京都市立芸術大学, 音楽学部, 教授 (60155356)
• Co-Investigator(Kenkyū-buntansha): 川上 央 日本大学, 芸術学部, 教授 (20307888)
• Project Period (FY): 2016-04-01 – 2020-03-31
• Keywords: コンピュータ・グラフィックス / バイオロジカルモーション / fMRI / 知覚評価 / ダンス / 脳活動

Outline of Final Research Achievements

When adding a human-like motion to a human body model in three-dimensional computer graphics by using a motion capture system, it is often necessary to make post-processing corrections for measurement errors. In this research, as a means of error correction of captured dance motions, the ground contact state of the toes based on an auxiliary visual image was used. The effect of this correction was investigated by a forced-choice method in a paired comparison task using biological motion stimuli. We also measured brain activity using fMRI. As a result, while there are certain changes due to the correction, and some evaluators judged the corrected version to be better than the original, we obtained results that suggest that the corrected version is less dynamic than the original version.

Free Research Field

知覚心理学

Academic Significance and Societal Importance of the Research Achievements

近年，コンピュータ・ゲームや映画の映像制作では人間の動作を取り込んでコンピュータ・グラフィックス上の人物モデルの動作データに活用することが増えてきている。しかし現状では誰もが簡単に使える状態ではなく，手作業の言わば職人芸が自然で現実感のある動作の作成には必要である。この研究成果は，そのような手作業の自動化につながる要素技術へのヒントとなる。その一方で視聴者の知覚的評価には多様性が存在し，動作の滑らかさと躍動感の間には相殺関係があることも見いだした。