Validation of a Single Camera, Spatio-temporal Gait Analysis System

The measurement of foot-surface contact position would aid the understanding of player movement and step strategy in sports such as tennis. However during competition, it is impracticable to instrument players. A markerless, view-independent, footsurface contact identification (FSCi) system was developed and validated. The FSCi system analysed standard colour video sequences of walking and running (barefoot and shod) from four unique camera perspectives; output data were compared to three-dimensional motion analysis. Results demonstrated that data for 99.6% of foot contacts (all camera perspectives) were identified. The calculation of gait variables, i.e. step length etc., was performed automatically for 91.3% of foot contact data; 8.7% of data required manual intervention for analysis. Resultant direction root-mean square error (RMSE) for foot contact position was 52.1 and 52.2 mm for barefoot and shod walking respectively. Resultant direction RMSE for foot contact position during running was 91.4 and 103.4 mm for barefoot and shod conditions respectively. The FSCi system measured basic gait parameters of walking and running without interfering with the activity being observed. The system represents a flexible approach which could be used for in situ gait analysis. The FSCi system could be used for gait analysis in competitive tennis however performance of the system when applied to larger filming areas, e.g. tennis courts, must be evaluated.