Comparative Analysis of Traditional Machine Learning and Deep Learning Techniques for Facial Expression Recognition

Facial expression recognition (FER) serves as a rapidly developing and important research space in the domain of pattern identification. Nonverbal communication plays a big part in daily life, and it accounts for roughly 55% to 93% of all communication. Robotics, video footage, expression recognition, home automation, electronic games, patient monitoring, mood disorders and anxiety diagnosis, psychoanalysis, and paralinguistic communication are all effective applications of facial expression identification. A thorough analysis of FER is provided in this paper. The literature is compiled from a variety of credible studies that were released during the most recent ten years. In this review, a variety of deep learning techniques are employed in addition to traditional machine learning methods. In addition, a number of freely available FER datasets are analysed, and standard outcomes are contrasted with them. In order to point out the subsequent research needed in this field for future researchers, this study provides an in-depth analysis of FER using conventional machine learning and deep learning strategies. Finally, by providing an overall comprehension of the latest cutting-edge techniques, this research study serves as a manual that is extremely beneficial for both experienced researchers looking for fruitful directions to pursue in the future and academicians in the FER field.