An Estimation Method of Emotion from Facial Expression and Physiological Indices by CCA Using Fuzzy Membership Degree

DOI

Bibliographic Information

Other Title
  • ファジィメンバシップ値を利用した正準相関分析による顔表情と生理指標からの感情の推定

Description

This report proposes a method for estimating human emotion from facial expression and three physiological indices including facial temperature, heart rate and pupil opening. It is commonly known that these indices are influenced by activity of autonomic nerve system which is sensitive to stress. Therefore, it is possible to obtain more reliable results by analyzing combined data vectors of facial expression and physiological indices than analyzing separately. These indices were acquired in the experiments where subjects watched movies of comedy or horror films which can cause emotions of pleasure or displeasure for the subjects. After the experiment, each subject assessed valence values and arousal values each in 9 levels, where each assessment is further classified into three categories by using the combined data vectors. The data vectors and membership degrees were analyzed by Canonical Correlation Analysis (CCA) and class of subjective assessment values was estimated. The results show that the recognition rates of the proposed method are about 55%.

Journal

Keywords

Details 詳細情報について

  • CRID
    1390282680650884992
  • NII Article ID
    130004591913
  • DOI
    10.14864/fss.27.0.303.0
  • Text Lang
    ja
  • Data Source
    • JaLC
    • CiNii Articles
  • Abstract License Flag
    Disallowed

Report a problem

Back to top