Authors:
Eun-Hye Jang
1
;
Mi-Sook Park
2
;
Byoung-Jun Park
1
;
Sang-Hyeob Kim
1
;
Myung-Ae Chung
1
and
Jin-Hun Sohn
2
Affiliations:
1
Electronics and Telecommunications Research Institute, Korea, Republic of
;
2
Chungnam National University, Korea, Republic of
Keyword(s):
Emotion, Affective Dimension, Physiological Responses.
Abstract:
In HCI, emotion recognition using physiological signals have been noticed lately because physiological signals can be simply acquired with some sensors and are less sensitive to social and cultural difference, in particular, there is strong correlation between human emotional states and physiological reactions. We have investigated the relation between affective dimensions, i.e., arousal, valence, intensity and approach, and physiological responses such as electrocardiograph (ECG), electrodermal activity (EDA), skin temperature (SKT), and photoplethysmograph (PPG). Three hundred college students participated in the experiment. To successfully provoke basic emotions (anger, fear, sadness, boredom, interest, surprise, joy, pain, and neutral), emotion-provoking film clips were excerpted for each target emotion. Physiological signals (ECG, EDA, PPG and SKT) as emotional responses were measured during participants’ exposure to emotional stimuli and participants were asked to rate the spec
ific emotions they had experienced on four affective dimensions, valence, arousal, intensity and approach. The result showed that there are correlations between affective dimensions and physiological responses. Contrary to valence and approach, arousal and intensity were positively related to heart rate (HR), skin conductance level (SCL) and skin conductance response (SCR), and showed negative relation to BVP and PTT. Our result suggests an availability of physiological signals for emotion recognition in HCI and can be helpful to provide the basis for the emotion recognition technique in HCI.
(More)