Design of Integrated Vision and Speech Technology for a Robot Receptionist

Service robot, which is proposed to relive repeated labor work for human, are widely researched nowadays. In this work, we propose a robot receptionist system. The robot receptionist system follows a modularize design style which is consists of several independent functions modules. This system design enables us easily to be able to integrate other functions. Driven by the rapid development of deep learning, visual recognition tasks including such as image classification, object detection and semantic segmentation have made remarkable progress. Combining some existing state-of-art approaches, the robot receptionist system integrates face recognition, human-computer interaction and facial expression display. In the following of this paper, we elaborate the design and implementation of the system from two aspects, software and hardware, respectively. The contribution of this paper mainly includes two aspects. Firstly, we propose a method for combining several software modules of speech recognition and vision. Secondly, we construct a simple and effective experimental system to realize our design.