Investigation of Evaluation Indices for Ball Game Players' Movements Using Deep Learning: Toward Computers vs. Coaches

2022 Fiscal Year Final Research Report

• Project/Area Number: 19K04911
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 25010:Social systems engineering-related
• Research Institution: Juntendo University
• Principal Investigator: Hirotsu Nobuyoshi 順天堂大学, スポーツ健康科学部, 教授 (90360726)
• Co-Investigator(Kenkyū-buntansha): 吉村 雅文 順天堂大学, 大学院スポーツ健康科学研究科, 教授 (10210767)
• Project Period (FY): 2019-04-01 – 2023-03-31
• Keywords: 深層学習 / 球技 / 評価指標

Outline of Final Research Achievements

In this study, we used deep learning and statistical methods to classify the movements of players during a game. We then extracted feature values and investigated indices to evaluate players from a new perspective. We obtained time-series data on the acceleration in each of the three axes during a game of college female soccer players wearing the latest inertial sensor, and classified the player's movements in the frequency domain by performing a fast Fourier transform. As a result, it was confirmed that the movement classification was mainly affected by the vertical acceleration, and that the distribution of the movements during the game was different for each player. By considering the distribution of the movements as a feature value during the game of the players, it is possible to present an evaluation index using the average value of the distribution, etc.

Free Research Field

オペレーションズ・リサーチ

Academic Significance and Societal Importance of the Research Achievements

本研究は、試合中の選手の動きに関する膨大な時系列データから、指導者の「眼力」でも認識することが難しい細かい動きも含めて分類することにより、特徴量を抽出し、新たな視点での評価指標を提示できたことに学術的意義がある。また、サッカーの指導現場において、試合中での選手の動きを評価できる定量的な指標として活用できる可能性を見出し、選手評価の視点を広げることができたことに社会的意義があるといえる。