Unobtrusive Sensing of Emotions (USE).

Egon van den Broek, Marleen H. Schut, Joyce H.D.M. Westerink, Kees Tuinenbreijer

    Research output: Contribution to journalArticleAcademicpeer-review

    23 Citations (Scopus)
    112 Downloads (Pure)

    Abstract

    Emotions are acknowledged as a crucial element for artificial intelligence; this is, as is illustrated, no different for Ambient Intelligence (AmI). Unobtrusive Sensing of Emotions (USE) is introduced to enrich AmI with empathic abilities. USE coins the combination of speech and the electrocardiogram (ECG) as a powerful and unique combination to unravel people's emotions. In a controlled study, 40 people watched film scenes, in either an office or a home-like setting. It is shown that, when people's gender is taken into account, both heart rate variability (derived from the ECG) and the standard deviation of the fundamental frequency of speech indicate people's experienced valence and arousal, in parallel. As such, both measures validate each other. Thus, through USE reliable cues can be derived that indicate the emotional state of people, in particular when also people's environment is taken into account. Since all this is crucial for both AI and true AmI, this study provides a first significant leap forward in making AmI a success.
    Original languageUndefined
    Pages (from-to)287-299
    Number of pages13
    JournalJournal of ambient intelligence and smart environments
    Volume1
    Issue number3
    DOIs
    Publication statusPublished - 7 Jul 2009

    Keywords

    • Ambient Intelligence (AmI)
    • METIS-260051
    • EWI-18292
    • speech features
    • heart rate variability
    • unobtrusive sensing
    • IR-73225
    • Emotion

    Cite this