Pattern recognition applied to the prediction of post-operative gait in man

The application of pattern recognition methods to the analysis of human gait is presented. Pattern features to describe both normal and abnormal walking are derived from measurements of the electrical activity of muscles at the ankle. Least-squares techniques are used to obtain a transformation matrix which predicts the change in the gait pattern vector following surgery on the ankle. The results obtained from applying this method to actual data from 12 patients are discussed.