Markerless gait characterization using single video camera setup

Abstract

Gait analysis is now a standard aspect for diagnosing and treating conditions such as cerebral palsy, Parkinson’s disease, rheumatoid arthritis and other neurological disorders. It is used as an early diagnosis approach to assessing a patient’s walking pattern. The state-of-the-art gait analysis techniques require a dedicated laboratory setup along with highly skilled professionals to operate the equipment involved to get the gait parameters for analysis. In addition to this, it also involves placing external markers on the body of the subjects which may not be the actual representation of how the involved subjects walk in natural settings. In this paper, a workflow for a pipeline application has been presented that provides spatiotemporal gait parameters and joint kinematics characterization from a video shot on a single smartphone camera. The proposed method for gait analysis can be used in home conditions thus eliminating constraints caused by a lack of specialised equipment, technical expertise, time and money.