Coping with full occlusion in fronto-normal gait by using missing data theory

Gait is a relatively new biometric which shows promise in its use. In this paper, we examine the monocular frontal view of gait. When tracking body parts in this view, complete occlusion of body parts may occur. To compensate for this, we offer a fresh standpoint where occluded data may be considered as data missing from a time series. Thus we can consider this as a novel application of the “missing data” problem studied in other fields dealing with time series data. Using this approach, we consider three ways of coping with occlusion by using a gait dataset and analysing the motion of coloured markers attached to body parts. The occluded motions are compensated for and the actual and predicted positions are compared which show our approach has promise for coping with complete occlusion.