Presentation Information
[2Yin-B-30]Development of a System for Analyzing Individual Differences in Motion Capture Data
〇Koki Hinaiji1, Yihsin Ho1, Eri Sato Shimokawara2, Hiroki Shibata2, Takenori Obo2, Ichiro Kobayashi1 (1. Takushoku University, 2. Tokyo Metropolitan University)
Keywords:
Signal processing,SVM,Gesture recognition
In recent years, motion capture data has been expected to be utilized in areas such as physical AI, as it can digitize human movement in detail. However, motion capture data present challenges, notably significant individual differences depending on the person capturing it. While motion capture technology, which is increasingly being implemented in society, mitigates the decline in sensor accuracy due to individual differences by averaging data using expensive, large-scale equipment, it does not directly address these individual variations. Therefore, this research describes the development of an analysis system to investigate the characteristics of individual differences by analyzing motion capture data acquired using electromyography (EMG). As a preliminary step to the system, the authors aim to extract data segments that contain individual difference features. A machine learning model was constructed and evaluated using experimentally acquired EMG data to objectively distinguish between periods of primary motion and static states.