Gesture Data Modeling and Classification Based on Critical Points Approximation

Human-Computer Interaction (HCI) using natural gestures is one of the promising developments in User Interface technology. One of key issues in its design is reliable modeling and classification of gesture data. In this article, we present a method for abstraction of gesture movement information, by reducing it to a sequence of approximated critical points (locations and types). The sequence of such critical points has good feature extraction properties. We present the method, results of classification, and discussion of the properties of the representation based on example gesture dataset recorded with motion capture equpiment.