In modern sports education, computer vision techniques have become increasingly important for analyzing team strategies and tactics in fixed-field team sports. This research presents a combination approach using object detection, multi-object tracking, and social network analysis (SNA) to investigate the dynamics of ice hockey strategies. Specifically, we utilize YOLOv8 object detection algorithm to detect players and ByteTrack to track their movements. The passing information between players is then used to construct a network representation of the team's strategy. By using weighted-edges and modularity network community detection, this research demonstrates the team roles of each player in communities analysis and captures the impact of team strategies. The goal of this research is to promote teamwork, strategic analysis, and the development of innovative knowledge of sports rules and strategies in sports education.

Authors:
Boyang Zhang, Tampere Universities, Finland

About the Presenter(s)
Dr. Boyang Zhang is expertise in AI, ML, computer vision, NLP, data mining, social media analytics, crisis communication, risk management. Excels in applying ML for corporate communication and leveraging computer vision in reality.

See this presentation on the full schedule – On Demand Schedule

Bookmark