Research Repository

A nonlinear heartbeat dynamics model approach for personalized emotion recognition

Valenza, G and Citi, L and Lanata, A and Scilingo, EP and Barbieri, R (2013) 'A nonlinear heartbeat dynamics model approach for personalized emotion recognition.' Proceedings of the Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBS. 2579 - 2582. ISSN 1557-170X

Full text not available from this repository.

Abstract

Emotion recognition based on autonomic nervous system signs is one of the ambitious goals of affective computing. It is well-accepted that standard signal processing techniques require relative long-time series of multivariate records to ensure reliability and robustness of recognition and classification algorithms. In this work, we present a novel methodology able to assess cardiovascular dynamics during short-time (i.e. < 10 seconds) affective stimuli, thus overcoming some of the limitations of current emotion recognition approaches. We developed a personalized, fully parametric probabilistic framework based on point-process theory where heartbeat events are modelled using a 2nd-order nonlinear autoregressive integrative structure in order to achieve effective performances in short-time affective assessment. Experimental results show a comprehensive emotional characterization of 4 subjects undergoing a passive affective elicitation using a sequence of standardized images gathered from the international affective picture system. Each picture was identified by the IAPS arousal and valence scores as well as by a self-reported emotional label associating a subjective positive or negative emotion. Results show a clear classification of two defined levels of arousal, valence and self-emotional state using features coming from the instantaneous spectrum and bispectrum of the considered RR intervals, reaching up to 90% recognition accuracy. © 2013 IEEE.

Item Type: Article
Subjects: T Technology > TK Electrical engineering. Electronics Nuclear engineering
Divisions: Faculty of Science and Health > Computer Science and Electronic Engineering, School of
Depositing User: Luca Citi
Date Deposited: 23 Jan 2015 12:51
Last Modified: 05 Feb 2019 19:15
URI: http://repository.essex.ac.uk/id/eprint/12345

Actions (login required)

View Item View Item