In this paper, a framework for automatic atomic human action segmentation in continuous action sequences is proposed A star figure enclosed by a bounding convex polygon is used to effectively and uniquely represent the extremities of the silhouette of a human body. Thus, human actions are recorded as a sequence of the starJigure 's parameters, which is then used for action modeling. To model human actions in a compact manner while characterizing their spatiotemporal distributions, star $@re parameters are represented by Gaussian mixture models (GMM). Experiments to evaluate the performance of the proposed framework show that it can segment continuous human actions in an eficient and effective manner.