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.
CITATION STYLE
Marpaung, A., & Gonzalez, A. (2017). Can an affect-sensitive system afford to be context independent? In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 10257 LNAI, pp. 454–467). Springer Verlag. https://doi.org/10.1007/978-3-319-57837-8_38
Mendeley helps you to discover research relevant for your work.