Representation of fundamental movements and pauses for archiving traditional skills

Abstract

Considering the reduction of the number of persons engaging in traditional skills, archiving various types of traditional skills is required to preserve and transmit them to future generations. We focus on representing fundamental movements and pauses in traditional skills because they are key components in describing traditional skills and archiving them. The fundamental movements and pauses can be described based on the movements of a number of body parts obtained using motion capture system. In this paper, we propose an efficient method to represent fundamental movements and pauses using the motion data. The proposed method generates concise and informative feature values from the motion data on the basis of dimensionality reduction and feature selection. The effectiveness of the proposed method is evaluated through an experiment to describe several types of fundamental movements in Japanese traditional tea ceremony.