Publication

Abstract : Automatic facial emotion perception has gained massive popularity over the last two decades. It is because of an growing need for behavioral biometric systems and the interaction between person and computer where facial emotion detection and emotional strength play a vital role. In general, the present works do not encode the intensity of the facial emotion observed and even less do they together model multi-class facial compliance data. The emotions and the corresponding intensities of these emotions are identified by our activities. Recently deep learning methods were proposed as an alternative to traditional SER techniques.This paper provides an overview of deep learning approaches and analyzes many recent studies used to classify emotions based on expression. The analysis includes used repositories, extracted emotions, contributions to the identification of speech emotions, and related limitations.