Document Type : Review Paper

Authors

• seyed sadegh hosseini ¹
• Mohammadreza Yamaghani ²
• Soodabeh Poorzaker Arabani ³

¹ Ph.D. candidate

² Department of Computer Engineering, Lahijan Branch, Islamic Azad University, Lahijan, Iran

³ Assistant professor at Islamic Azad University of Lahijan

Abstract

Emotional computing synergizes the understanding and quantification of emotions, drawing on diverse data sources such as text, audio, and visual indicators. A challenge arises when attempting to discern authentic emotions from those concealed deliberately via facial cues, vocal nuances, and other communicative behaviours. By integrating multiple physiological and behavioural signals, more profound insights into an individual's emotional state can be achieved. Historically, research has predominantly concentrated on a singular facet of emotional computing. In contrast, our study offers an in-depth exploration of its pivotal domains, encompassing emotional models, Databases (DBs), and contemporary developments. We commence by elucidating two prevalent emotional models, followed by an examination of a renowned sentiment analysis DB. Subsequently, we delve into cutting-edge methodologies for emotion detection and analysis across varied sensory channels, elaborating on their design and operational principles. In conclusion, the fundamental principles of emotional computing and its real-world implications are discussed. This review endeavours to provide researchers from academia and industry with a holistic understanding of the latest progress in this domain.

Keywords

• Emotion Recognition
• Machine Learning
• Deep Learning
• Multimodal Emotions
• Dataset
• Information Fusion
• Feature Extraction

Main Subjects

• Other