Home // International Journal On Advances in Software, volume 9, numbers 1 and 2, 2016 // View article

Facial Part Effects Analysis using Emotion-evoking Videos focused on Smile Expression Process

Authors:
Kazuhito Sato
Momoyo Ito
Hirokazu Madokoro
Sakura Kadowaki

Keywords: Psychological stress measures; Intentional facial expression; Machine learning approaches; Behavior modeling

Abstract:
This study specifically examines the expressive process of "happiness" related facial expressions after giving a stress stimulus. In addition, it presents a quantitative analysis of expressive tempos and rhythms using mutual information. By acquiring image datasets of facial expressions under states of pleasant-unpleasant stimulus for 20 participants, we calculated the information in three region of interests (ROIs): ROI 1, the whole face and the upper face; ROI 2, the whole face and the lower face; and ROI 3 between the upper face and the lower face. Additionally, we tried to express complexity and ambiguity objectively during facial expressions because of human psychological states. The results clarified the possibility of estimating the impression of facial expressions from the magnitude relation and order relation of mutual information of each ROI. More than male participants, female participants were able to create facial expressions of "happiness" easily and intentionally, and were less susceptible to discrepancy expressions. Finally, we discussed the differences of expressive paths between intentional and spontaneous facial expressions based on the order of the mutual information of the ROIs. As a result, we figured out the validity of our hypotheses concerning to the individual expressive path of each facial expression.

Pages: 128 to 140

Copyright: Copyright (c) to authors, 2016. Used with permission.

Publication date: June 30, 2016

Published in: journal

ISSN: 1942-2628