Can An Affect-Sensitive System Afford To Be Context Independent?

Keywords

Affect recognition; Affective computing; Context-centric; Contextual knowledge; Speech Paralinguistic

Abstract

There has been a wave of interest in affect recognition among researchers in the field of affective computing. Most of these research use a context independent approach. Since humans may misunderstand other’s observed facial, vocal, or body behavior without any contextual knowledge, we question whether any of these human-centric affect-sensitive systems can be robust enough without any contextual knowledge. To answer this question, we conducted a study using previously studied audio files in three different settings; these include: no contextual indication, one level of contextual knowledge (either action or relationship/environment), and two levels of contextual knowledge (both action and relationship/environment). Our work confirms that indeed the contextual knowledge can improve recognition of human emotion.

Publication Date

1-1-2017

Publication Title

Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

Volume

10257 LNAI

Number of Pages

454-467

Document Type

Article; Proceedings Paper

Personal Identifier

scopus

DOI Link

https://doi.org/10.1007/978-3-319-57837-8_38

Socpus ID

85020897256 (Scopus)

Source API URL

https://api.elsevier.com/content/abstract/scopus_id/85020897256

This document is currently not available here.

Share

COinS