Distributed System of Real Time Head Gesture Recognition in Development of Contactless Interfaces

Abstract

23 3 Abstract: The article discusses the gesture recognition system based on head distributed multi-agent network. A model of head movement in 3D space. Head position described by the matrix configurations. Developed a method of recognition of fuzzy gestures, based on recognition of the graph constructed on the basis of fuzzy nodes. Graph is generated for each dynamic gesture, perfect head. It is shown that the recognition process can be performed by a fuzzy neural network. Software agents implement recognition system. Presented work distributed recognition system in real time. A general scheme for the formation of classifying gestures. The estimation of the volume of data in the form of a matrix. Presented data rate in a distributed system. The data capacities of the most common data channels. Choose the method of transmission of a video stream.