Skill Level Classification Using Motion Data via Spatial Temporal Graph Convolutional Network
Tatsuki Seino, Naoki Saito, Takahiro Ogawa, Satoshi Asamizu, Miki Haseyama · 2023
A skill level classification method using motion data based on Spatial Temporal Graph Convolutional Network (ST-GCN) is presented in this paper. Main contribution of this paper is realization of the accurate skill level classification by considering time-series characteristics. Motion data are closely associated with behaviors and tacit knowledge necessary for the classification of the skill levels. In addition, ST-GCN enables classification considering time-series characteristics to achieve high classification performance. Then the proposed method performs the skill level classification by ST-GCN using motion data. Consequently, the proposed method realizes the performance improvement of the skill level classification.