Assessment of gait parameters during unconstrained straight-line walking using a single camera

Abstract

Gait parameters can reflect health status and functional abilities of people, so assessment of gait parameters especially in the daily life has recently become a hot issue in the field of health monitoring. Vision-based gait monitoring is contactless and unobtrusive for long-term home monitoring, especially for elders. Although gait monitoring based on a single camera is usually easy to use, most studies require the subjects to walk perpendicularly to the camera's optical axis or along some specified routes. This limits the application of single camera-based gait monitoring systems in elder home monitoring. In our study, on the basis of projective geometry theory, we present a method to assess the subjects' relative gait parameters during unconstrained straight-line walking using a single camera. We derive a calculation formula for step length ratio independent of the subject's walking direction, and hence can directly obtain step length ratio from a monitoring video, without need of measuring the absolute lengths of these steps in advance. Experimental results indicate that our method can achieve very close estimates of step length ratios to the real ones, independently of the subjects' walking directions when they are walking along straight lines.