Athlete body power and strength estimation using skeleton point cloud

Abstract

Vertical jump assessment is one of the most frequently used techniques to evaluate lower limb strength and power in the context of strength and conditioning. Countermovement jump (CMJ) and squat jump (SJ) have been recognized as the most valid and reliable vertical jump tests to assess body power and strength. The traditional method requires professional involvement and equipment for parameter assessment. In the context of college athletes, it is not practical to acquire direct supervision from coaches and use laboratory-based equipment such as force plates. Hence, the main objective of this study is to reduce the physical involvement in the traditional method by developing a system to identify lower body power and strength, which can be handled at an athlete’s own pace. Thus, we propose a single camera system that captures athlete skeleton joint variation from the lateral view to identify the accuracy of CMJ and SJ along with measurements for power, maximum jump height, reactive strength index, and ground reaction force. These performance parameter values are measured using a rule-based model developed with motion equations and skeleton joint variations. The biomechanics of the exercises are identified based on the clinically accepted jump protocol. The experimental results show that the system identifies athlete biomechanics with an accuracy of 91.7% for CMJ and an accuracy of 95.8% for SJ. Also, the system measures maximum jump height with a standard error of 0.88 cm.