Extracting human gait signatures by body segment properties

We describe a new method for extracting human gait signatures by topological analysis, using properties of body segments. The gait signature is extracted in three stages: extraction of the body contour by a thresholding and morphological filter; extraction of the leg angles based on regression analysis of contour data; finding the body points guided by known anatomical knowledge. A 2D stick figure is used to represent the human body model and trajectory-based kinematic features are extracted from the image sequences for describing and analyzing the gait motion. Also, the inherent periodicity in gait motion is analyzed by delay coordinates and a phase-space portrait. The utility of the proposed method is demonstrated in experiments, with comparison to medical data.