Non-invasive gait analysis from broadcast video (2017)

The New Balance 5th Avenue Mile race is a mile-long race held in New York City. In 2017, the race was attended by men and women (approximately 7,700 in total), of which runners were elite and non-elite. New Balance Athletics Inc. (NBA) is a global sports shoe and apparel manufacturer and is sponsor to some of the participating athletes. NBA wish to better understand how their shoes and sponsored athletes perform during competition and training, to help characterise running performance and fatigue.
As part of shoe development, NBA engages with traditional biomechanical analysis techniques of walking and running (e.g. pressure insole analysis, shoe deformation etc.). These techniques provide detailed information of an athlete’s gait during single or multiple running strides. However, these techniques are also restricted to lab-based environments and are rarely used in competition.
Middle-distance athletes typically modify step frequency to mitigate the effects of fatigue (Kadono and Ae, 2015). Recently, non-invasive methods, suitable for use in competition, have been developed by the Centre for Sports Engineering Research (CSER) to automatically analyse gait parameters such as step frequency. The purpose of this project was to analyse broadcast video footage of athlete step strategy throughout the entire, 5th Avenue Mile race.