Fuzzy hand posture models in man-machine communication

Abstract

Ever since the assemblage of the first computer, efforts have been made to improve the way people could use machines. This ambition is still present nowadays: indeed, intuitively operated systems are currently under intensive research. Intelligent Space (or iSpace) based systems are good examples: they strive to be comfortable and easy to use, even without demanding technical knowledge from their users. However, their aim is not limited to this: in fact, their ultimate goal is to achieve an intelligent environment for higher quality, natural, and easy to follow lifestyle. The system described in this chapter can be used to create a new, intuitive man-machine interface for iSpace applications. The solution exploits one of the basic human skills, namely the ability to assume various hand postures. The proposed system first processes the frames of a stereo camera pair and builds a model of the hand posture visible on the images and then classifies this model into one of the previously stored hand posture models, by using neural networks and fuzzy reasoning. © 2010 Springer-Verlag Berlin Heidelberg.