Skip to Main Content
An approach to human action modelling and recognition in continuous action sequences is proposed. A star figure enclosed by a bounding convex polygon is used to effectively represent the extremities of the silhouette of a human body. Thus, human actions are recorded as a sequence of the star figure's parameters, which is then used for action modelling. To model human actions in a compact manner while characterising their spatiotemporal patterns, star figure parameters are represented by a 2D feature map, which is used and regarded as a spatiotemporal template. Experiments to evaluate the performance of the proposed framework show that it can recognise human actions in an efficient and effective manner.
Date of Publication: January 2011