A Similarity Indicator for Differentiating Kinematic Performance Between Qualified Tennis Players

详细信息    查看全文

• 关键词：Multi ; channel data ; Kernel methods ; Kinematics ; QKLMS ; Similarity indicator
• 刊名：Lecture Notes in Computer Science
• 出版年：2017
• 出版时间：2017
• 年：2017
• 卷：10125
• 期：1
• 页码：309-317
• 丛书名：Progress in Pattern Recognition, Image Analysis, Computer Vision, and Applications
• ISBN：978-3-319-52277-7
• 卷排序：10125

文摘

This paper presents a data-driven approach to estimate the kinematic performance of tennis players, using kernels to extract a dynamic model of each player from motion capture (MoCap) data. Thus, a metric is introduced in the Reproducing Kernel Hilbert Space in order to compare the similarity between models so that the built kernel enhances groups separability: the baseline reference group and the group including players developing their skills. Validation is carried out on a specially constructed database that contains two main testing actions: serve and forehand strokes (carried out on a tennis court). Besides, the classical kinematic analysis is used to compare our kernel-based approach. Results show that our approach allows better representing the performance for each player regarding the ideal group.