Extracting Player's Stance Information from 3D Motion Data: A Case Study in Tennis Groundstrokes

This study presents a novel approach towards computing elements of balance, movement fluidity and reaction time, the foundations of which are commonly introduced in tennis as swing stance. The achieved results utilising presented algorithms, show 100 % recognition of tennis swings (forehands and backhands) and swing stance angle extractions from the 3D test data set. The data set was captured at 50 Hz without ball impact information using a stationary multi-camera setup. The next generation of exergames and augmented coaching technologies, utilising the presented approach for processing players' footwork and stance will enable research advancements in human performance and further developments towards improving proprioception and kinaesthetic awareness. Determining body orientation within a temporal and spatial activity pattern to predict the follow-up action(s) may also enable advancements such as improving safety in surveillance, robot and automotive vision.