Automatic system for the analysis of flexion angle of the knee using a probabilistic model

In human locomotion, the way of walking depends of multiple factors; some of them, alien to the person: treading surface, climatic variables, available space and others. Factors related to morphology also generate difficulties or anomalies in the people movement. This paper analyze the human march and generates a model of human walk based on captured data from people walking using two Kinect cameras. For modeling, we take as reference the points from the hip, knee and ankle; additionally pressure data were captured in two platforms of weight to obtain the pressure exerted in each foot; we calculate the probabilistic model based on the results of 30 persons. To evaluate its efficiency, the model was tested in 100 persons. The results show that 58% of subjects present an apparent normal gait, 21% apparent normal gait with possible anomaly in the left leg, 18% apparent normal gait with possible anomaly in the right leg and 3% apparent abnormal gait.