Real-time facial expression recognition system

Abstract

Facial expressions are visible manifestation of the affective state, cognitive activity and personality of a person. Over the last two decades, the advances in imaging and computing technology have led to significant research effort on facial expression recognition. In the final year project, the recognition system is designed to recognise drivers’ expressions from a real-time video. As a crucial part of the future driving enhancement system, facial expression recognition could be used to analyse drivers’ emotion states so that it will predict risky driving behaviours. By providing proper corrections, some accidents could be avoided. To implement the face expression recognition, histogram-of-oriented-gradient (HOG) based object detector, active shape model (ASM) and extreme learning machines (ELM) algorithms are used in this project. The real-time recognition system could locate face regions, extract facial features and recognise basic human expressions that could be considered as an important study or exploration for the future driving safety enhancement.