Review of Machine and Deep Learning Techniques for Expression based Facial Emotion Recognition

Abstract

Facial Emotion Recognition (FER) is a technique that learns features from raw images and achieve Emotions, is used to detect and analyze any person’s emotional state from expressions based on facial cues. The ability to accurately recognize facial emotions must be improved to enhance virtual interpersonal and Human-robot interaction that can be beneficial in many applications, this can be achieved by leveraging Artificial Intelligence and computer vision techniques. This paper studied the advancements in the field of FER, that can influence performance and compared different Machine Learning(ML) and Deep Learning(DL) based techniques by focusing on several aspects: like Nature of dataset, the input representation for providing more standardized input data, and choice Network architectures for feature extraction and classification purposes. In essence, this paper serves as a comprehensive guide to FER, providing insights into its current landscape, challenges, and future directions.