Gesture recognition using syntactical hand-sign parsing

Abstract

Gesture recognition is needed for a variety of applications and is now gaining in importance as one method of enabling natural and intuitive human machine communication. The crucial point in recognizing gestures is that it requires great computational power and complicated structural analysis of the hand image, which current systems cannot yet handle at practical speeds. In this paper, a new method using syntactical hand-sign parsing is introduced to provide a simple and compact recognition algorithm. The method translates a sequence of hand-sign images into a canonical symbolic representation. As a result of this translation, the gesture-sequence becomes applicable for processing by conventional parsing algorithms. In order to verify the proposed method, we developed and evaluated an experimental gesture-recognition system for communicating command-messages to an intelligent mobile robot.