Roboticon: A realistic platform to imitate facial expressions

Abstract

This paper presents a concept that quickly detects the facial expressions of humans with a sufficient accuracy and imitates the expressions remotely in terms of robot-based mimics. In traditional online communication of expressions, it is not certain that the sender remains in the same state of expression, when the state is actually received in the remote side. In addition, the expression received in the remote side is mimicked by special visual characters commonly known as emoticons. The approach presented in this paper addresses the issues of real-time detection, communication, and imitation of expressions in order to solve the existing shortcomings. In particular, a computationally efficient intensity threshold-based algorithm is adopted to detect facial expressions from video signals. Because of such a simplicity, the proposed system ensures fast detection of an expression, which is quite convenient for real-time communication. At the same time, the system copies the expression on the face of a humanoid robot at the receiving end. It is expected that this novel approach, which is referred to as Roboticon will play a key role to imitate facial expressions in real-time.