The University of Tokyo
Football Lab

Addressing the challenge of limited access to specialized goalkeeper coaching, our project leverages computer vision and deep learning to analyze goalkeepers' movements during practice. This system provides detailed performance insights, filling a critical gap in training. Recognized by the 2023 MITOU IT Program, this innovative project has secured funding of 2,730,000 yen, showcasing its potential to transform goalkeeper training through advanced skeleton estimation technology.

#Computer Vision #Deep Learning #Pose Estimation

In collaboration with Shingo Akimoto, the sprint coach of the University of Tokyo Football Team, we are engaged in research and development of a system that supports the analysis of running movements.
#Deep Learning #Pose Estimation

We are harnessing the power of data analytics through the Bepro system. This involves gathering extensive game data, including player positions, event timestamps, and details of actions like passes and shots. Our analysis of these data points is instrumental in providing actionable insights to our coaching staff. These insights are key to improving strategies and overall team performance, showcasing the practical application of data analysis in enhancing competitive football.

#Statistical Analysis #Time Series Data Analysis

As the diversity of players' career paths expands, our goal is to uncover insights that support the formation of appropriate career trajectories for these athletes. We aim to understand the various backgrounds and experiences that contribute to a soccer player's professional journey, ultimately aiding in the development of more effective career strategies.

#Statistical Analysis #Career Development

We are working on the development of a system that utilizes object detection and tracking technologies to identify and track the positions of each player from game footage. By using this system, we can automatically acquire each player's coordinates and movement distances, enabling scene searching and more detailed tactical analysis.
#Deep Learning #Object Detection #Object Tracking

In collaboration with Takeyuki Tadokoro, an alumnus of the University of Tokyo Football Team and currently a kick specialist coach at Kicking Lab, we are engaged in the research and development of a system that supports the analysis of kicking actions.
#Deep Learning #Pose Estimation

Existing GPS devices are very expensive, costing around 100,000 yen, and obtaining real-time data is challenging. We have conducted research and development on a device that can achieve highly accurate position measurement to within a few centimeters at a much more affordable price of around 7,000 yen.
#GPS Device #Micro Controller