An Empirical Study of Facial Expression Recognition Methods

Abstract

Facial Expression Recognition (FER) is a field that is being well researched due to its great impact on decision-making application in the domain of medical, security, corporate and other systems. In this paper, a high-level overview of the FER techniques, databases, and analysis of significant research is performed in four-folds. First, the importance of FER applications in the industries is discussed. Second, the performance of the latest frameworks is analyzed. Third, the challenges faced in adaptability are described. Fourth, the gaps in the literature are addressed. The study reveals that FER techniques are useful in the development of AI applications for pattern analysis that work by data pre-processing, feature extraction, feature selection, and expression recognition. The main challenge faced is concerned with description-level data. This study provides a better understanding and updated information for future researchers who wish to explore this domain.