Development of an automated image processing system for kinematic analysis of human gait

We propose a simple, inexpensive, portable and real-time image processing system for kinematic analysis of human gait. We view this as a feature based multi-target tracking problem. Here we track the artificially induced features appearing in the image sequence due to the non-impeding contrast markers attached at different anatomical landmarks of the subject under analysis. This paper describes a real-time algorithm for detecting and tracking the feature points simultaneously. By applying a Kalman filter, we recursively predict the tentative features location and retain the predicted point in case of occlusion. A path coherence score is used for disambiguation along with tracking for establishing feature correspondences. Experimentations on normal and pathological subjects in different gait was performed and results illustrate the efficacy of the algorithm