Color Analysis of Facial Skin: Detection of Emotional State

Geovany A. Ramirez, Olac Fuentes, Stephen L. Crites Jr., Maria Jimenez, Juanita Ordonez; Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops, 2014, pp. 468-473

Abstract


Humans show emotion through different channels such as facial expression, head poses, gaze patterns, bodily gestures, and speech prosody, but also through physiological signals such as skin color changes. The concentration levels of hemoglobin and blood oxygenation under the skin vary due to changes in a person's emotional and physical state; this produces subtle changes in the hue and saturation components of their skin color. In this paper, we present an evaluation of facial skin color changes as the only feature to infer the emotional state of a person. We created a dataset of spontaneous human emotions with a wide range of human subjects of different ages and ethnicities. We used three different types of video clips as stimuli: negative, neutral, and positive to elicit emotions on subjects. We performed experiments using various machine learning algorithms including decision trees, multinomial logistic regression and latent-dynamic conditional random field. Our preliminary results show that facial skin color changes can be used to infer the emotional state of a person in the valence dimension with an accuracy of 77.08%.

Related Material


[pdf]
[bibtex]
@InProceedings{Ramirez_2014_CVPR_Workshops,
author = {Ramirez, Geovany A. and Fuentes, Olac and Crites, Stephen L. Jr. and Jimenez, Maria and Ordonez, Juanita},
title = {Color Analysis of Facial Skin: Detection of Emotional State},
booktitle = {Proceedings of the IEEE Conference on Computer Vision and Pattern Recognition (CVPR) Workshops},
month = {June},
year = {2014}
}