Real Time Gesture Recognition in A Real Complex Environment Using Interactive Model Generation

Abstract

This paper describes a real time gesture recognition method for interactive systems. Our method can recognize gesture robustly in a variety of real complex environments. For the robust gesture recognition, our method performs an interactive model generation before the recognition processing, and an individual gesture model is constructed by the interaction between a user and the computer. The individual gesture model is based on the template matching technique and it is constituted by the set of color template images which represent a set of specific poses of human part in a gesture. Using our method, we realize a real time interactive system, the Gesture Game System, which can activate the characters of video game by gestures in real-time and indicates the usefulness of our method.