A Review of Advancements in Facial Emotion Recognition and Detection Using Deep Learning

Abstract

Nowadays, Emotion recognition and detection technology are trending among researchers. Automatically recognizing facial emotions is a challenging task in computer vision a face picture has a wide range of potential applications, including security while driving, interaction between humans and computers, healthcare, psychology, video conferencing cognitive research, and others. In this study, a deep-learning-based method is recommended to assess a person’s expressions with their faces. While these emotions are extremely complicated and hard for robots to comprehend, they are simple to comprehend for humans. This FER will used in lie detection by analyzing the micro-facial emotions of humans it will predict the accurate result. By using deep learning and CNN methods it can detect automatically in a real-time environment from human facial expressions. We can implement deep learning to develop robust and reliable systems. The FER is capable of recognizing and detecting facial expressions automatically. It involves training in convolutional neural networks to analyze them and examine the classification of facial emotions. The result of the dataset will depend on the specific application and the desired model accuracy. Some of the most popular and well-regarded datasets for facial expression recognition include AffectNet, FER-2013, JAFFE, CK+, and DISFA. However because they haven’t considered the difficulties presented by variations in head position, their usefulness is limited, and the accuracy still falls short of expectations. The work focused on methods and datasets that will be used to predict all kinds of such as joy, sorrow, rage, contempt, indifference, surprise, etc. And also this review provides brief information about the working methods of FER and evaluates the future challenges. This review influence will increase the efficiency of user experiences in applications like educational software, virtual assistants, and entertainment. It provides highly accurate results.