Motion Laboratory

This research develops a creation-support system in broadcasting and art performances using the CG contents of realistic human body motions acquired by a motion capture system. We developed a system that controls human body animation with LEMUR which is a multi-touch device. This system allows users to intuitively control human body motions, reproduction, viewpoints by switches and sliders that are designed on LEMUR. Since the system supports input of plural points, the user puts many motions at the same time by arbitrary timing and generates new motions. System evaluations showed that the created motions have originality and that system can be used for learning dance and generating new motions.

We have developed a learning and simulating system for classical ballet solo dances using Web3D technologies and motion capture data. The purpose of this research is to produce Web-based contents and to develop systems which anyone can access and use it through the Internet. This system has two modes, the learning mode and the simulation mode, and user can change it depending on a purpose. In the learning mode, user can interactively control animation and change viewpoint it in 3-dimensional computer graphics. In the production mode, user can simulate and play ballet pieces with costumes, background scenes, music, and camera works. This system prepares plural camera animations for each ballet piece, and user can see it from various viewpoints. As a result of a questionnaire evaluation, particularly this system was confirmed to be useful as a simulation system.

The purpose of this research is to develop a system for supporting dance. We have developed a database system for fundamental steps of classical ballet to manage and to utilize 3-dimensional motion data. First we have archived 583 fundamental steps obtained by motion capture systems, then constructed 3-dimentional motion database for them. User can search a fundamental step and preview its motion by 3-dimensional CG. In order to retrieve motions, two approaches are applied; the first one based on the classification of ballet fundamental motions the second one is based on the characteristics of 3-dimensional motion data, which is evaluated by calculating the sum of each position data along the direction of the hardest step. As a result of an evaluation, we confirmed that the approach using the characteristics of 3-dimensional data is valid for objective retrieval of motions.