Explaining rifle shooting factors through multi-sensor body tracking1

There is a lack of data-driven training instructions for sports shooters, as instruction has commonly been based on subjective assessments. Many studies have correlated body posture and balance to shooting performance in rifle tasks, but mostly focused single aspects postural control. This study finding relevant factors by examining the entire over sequences time. A data collection was performed with 13 human participants carrying out live scenarios while being recorded multiple tracking sensors. pre-processing pipeline produced novel skeleton sequence representation, which used train transformer model. The predictions from this model could be explained per sample basis using attention mechanism, visualised an interactive format humans interpret. It possible separate different phases scenario high classification accuracy (80%). Shooting detected extent separating their strong weak hand. dataset pipeline, well techniques generating explainable presented laid groundwork future research domain.