Video based human gait bilateral symmetry analysis and its validity for functional assessment after an orthopaedic surgery

A color marker based computer vision system was proposed to provide spatio-temporal and kinematic information of human gait. By detecting the markers from video sequences, this system provides quantitative gait pattern information for clinicians to evaluate the rehabilitation progress of the patients who had undertaken total knee replacement (TKR) and/or total hip replacement (THR) surgeries. The leg bilateral symmetry as an efficient feature for this evaluation purpose was proposed. To calculate this parameter, leg angle curve as a gait signature was introduced to describe the gait pattern. The symmetry is denoted by dynamic time warping (DTW) distance of this walking signature. Normal and abnormal gait can be distinguished based on the leg bilateral symmetry.