Emotion detecting humanoid – the use of GPSO Algorithm

Abstract

Recently, we implemented a real-time expert system that receives a live video stream of the face of a human subject and recognizes any of the six basic facial emotions expressed by the subject. The expert system was implemented using the Guided Particle Swarm Optimization (GPSO) algorithm which we earlier developed and was shown to have a very good recognition success rate. We suggested in our earlier publications about the expert system that a possible application is to embed the system into a robot that could be deployed to assist bed-ridded and highly disabled patients who can only communicate through facial expressions. In this paper, we present our first attempt towards achieving this goal by using the expert system to control a humanoid robot. The system receives input from a wireless video camera mounted on the robot, evaluates and detects the emotions expressed in the video in real-time, and instructs the robot to perform some preprogrammed actions according to the detected emotion.