Bimodal speech recognition for robot applications

Abstract

By the earliest motivation of building robots to take care of human being in the daily life, the researches on humanoid robots have been explored and developed over the recent decades. One challenge of humanoid robots is its capability of communication with people via audio-visual speech. Certainly, the integration of audio module and visual module improve the robot performance. In this paper, we present a novel system combines these two modules in one environment to be utilized alone or in humanoid robots. The proposed system adapts the user facial movements that cannot be avoided in reality. The experimental results show how these two modules enhance each other and yield an effective speech recognizer.