WAVC 2018 paper: Foot Contact Timings and Step Length for Sprint Training



We are delighted that our research has been presented at WAVC 2018.

Technology has become a key part of monitoring and improving the performance and welfare of elite level athletes. Through intelligent measurement systems, coaching staff can optimise technique, avoid and care for injuries, and observe the effect of training programmes. Computer vision has the potential to provide powerful and flexible tools in this area, allowing measurements that do not interfere with the natural behaviour of the athlete, and which can be conducted in normal training sessions, or even during competition. The key challenge is for systems that can meet the rigorous precision and accuracy requirements.

As part of its ongoing work to apply computer vision techniques to sports, a team of computer vision and sports biomechanics researchers have developed a system for measuring two fundamental aspects of a sprint athlete’s technique: their step frequency, and their step length. In a paper to be published in this years IEEE Winter conference on Applications of Computer vision (WACV2018), the researchers compare the performance of the system to gold-standard measurement approaches currently used in sports biomechanics, force-plates (for timing) and marker-based optical motion capture (for step length). The passive, non contact, camera based system was able to measure foot-ground contact times to within 0.008 s (1.5-frame duration in 180 fps video) and step-lengths to within 1cm. Such a system has the potential to be in regular use during athlete training sessions, allowing coaches to accumulate useful information about athlete progress throughout the training season.

Evans, M., Colyer, S., Cosker, D. & Salo, A. 12 Mar 2018 2018 IEEE Winter Conference on Applications of Computer Vision: WACV 2018. IEEE, p. 1652-1660


Written by CAMERA Centre Coordinator