Uniplanar aquatic exercise quantified with inertial sensors and pose estimation

Abstract

Abstract This study presents a concurrent validity analysis of two measurement techniques for quantifying uniplanar human movement during aquatic exercise. Established marker-based biomechanical motion capture systems are labour intensive, and command significant prices, limiting their accessibility in non-specialist environments. Therefore, this study assesses the applicability of alternate measurement methods; inertial measurement unit (IMU) sensors and markerless computer vision (CV) tracking in the form of pose estimation , applied to the quantification and analysis of aquatic exercise. This analysis establishes the validity of each method by deriving several performance-related metrics; range of motion, duration, and angular velocity, of uniplanar aquatic exercise repetitions. Using the proposed methods it was observed that each performance metric demonstrated excellent agreement (ICC $$\ge$$ 0.94) across methods, based on intraclass correlation coefficients in 20 healthy young adults (n = 9 women, aged 20-26). This analysis encompassed two exercise types, knee flexion-extension and hip flexion-extension. The authors highlight that this contribution represents a step toward establishing robust, accessible methods for objective exercise quantification in aquatic settings, where previously there has been an absence of convenient technical solutions.