CONTINUOUS 6 DOF GESTURE RECOGNITION : A FUZZY LOGIC APPROACH

The method described in this paper recognizes previ ously learned gestures which the system was taught by performing them. Any kind of 6DOF input device (e.g. data glove, vis ual systems, space mouse etc.) can be used to gathe r the motion data that carries out the gesture. The relia bility of the recognition process can be improved b y repeating the same gestures several times and correcting wron g recognition results. This extends the system’s kn owledge. Once learned, the system translates the recognized gestures into (for a computer) identifiable objects (numbers, strings, events, etc.) that can be further processe d. The system’s advantages are its usability (e.g. wit h 2D or 3D input devices or in combination with fin ger status information, etc.), the minimum of information need d to recognize a gesture, and, consequently, the h ig speed of its scanning and comparison process.