Exploring the application domains of ML-based facial emotion recognition systems: Framework, techniques and challenges

• Published in: AIP conference proceedings Vol. 2919; no. 1
• Main Authors: Rani, Sita, Bhambri, Pankaj, Kaur, Jaskiran, Sangwan, Yashwant Singh
• Format: Journal Article Conference Proceeding
• Language: English
• Published: Melville American Institute of Physics 25.03.2024
• Subjects: Communication; Emotion recognition; Emotions; Face recognition; Feature extraction; Human motion; Verbal communication
• DOI: 10.1063/5.0184852
• ISSN: 0094-243X

Human facial expressions are one of the important techniques of non-verbal communication. Facial expressions are the most tender signs for larger communication and are complemented by other gestures like eye contact, hand movement, etc. This is the direct method of communication of human emotions and intent. In this paper, the authors present the Facial Emotion Recognition (FER) framework and a brief survey of various FER techniques. It also presents the various phases of the FER system i.e., face detection, pre-processing, feature extraction, and classification. Various FER databases like JAFFE, YALE, MUG, etc. are also summarized in terms of the number of emotions, number of images, and resolution. The importance of the domain in other related subject areas like medicine, neuroscience, psychology, decision science, gaming, mental research, etc., is also introduced. The authors explored the application areas of FER techniques. The authors also present the various challenges faced in the real-time implementation of FER models. Finally, the paper is concluded by discussing future research directions.