Cluster Dominance Analysis of Strength Training Motion Characteristics

This paper presents an approach to analyze clusters as a means to determine the characteristics of strength training motion patterns. The proposed method emphasizes the observation of dominance sequences within clusters and is reinforced by the formation of specific characteristics within each cluster. Data collection is performed using video-guided strength training exercises equipped with 1 kg dumbbells and recorded by a sensor embedded in smartwatches. The analysis method involves applying the concept of density affinity, which calculates the density ratio of clusters to the recognized motions. Subsequently, the dominance sequence is observed to identify which clusters exhibit distinct characteristics, ultimately determining the intended motions. The research findings demonstrate the potential for further investigation into a more comprehensive understanding of motion patterns, leading to the development of models that can be integrated into mobile devices or smartwatches.