BIOSENSOR FUSION STRATEGIES FOR EMOTION RECOGNITION

Abstract

This article focuses on the analysis and classification of emotion recognition strategies based on physiological signals, as well as on the study of available databases for multimodal analysis of psychological states using information from physiological signals. The article presents a classification of recognition methods, examines the influence of factors such as modalities and fusion strategies onto recognition performance, and determines the importance of choosing physiological signals for accurate emotion recognition. Advantages of deep learning models and practical applications of emotion recognition in real scenarios are also discussed.