Facial Expression Recognition: A Review of Methods, Performances and Limitations

Abstract

Facial expression is one of the profound nonverbal channels through which human emotion state is communicated, its automation involves analysis and recognition of facial features. Facial Expression Recognition (FER) is categorized as behavioral biometrics, and also applicable in the field of computer vision and human computer interaction. Variations in the nature of the images to be processed; head pose, image background, light intensity and occlusion are some of the sources of the challenges with facial expression recognition system. Achieving a robust automatic facial expression recognition system invariant to the aforementioned challenges, is the goal of this research area. This paper presents an analysis of major feature extraction and classification methods, their performances in terms of accuracy and their respective limitations.