Cluster Analytic Detection of Disgust-Arousal

Abstract

Automated detection of disgust-arousal could have applications in diagnosing and treating obsessive-compulsive disorder and Huntington’s disease. For achieving this ability, experimental data was used first to examine the thermal response of “facial muscles of disgust” to other common negative and positive expressions of emotive states. An attempt was then made to detect disgust-arousal through classification of affect-educed thermal variations measured along the facial muscles. Initial results suggest (i) muscles of disgust experience different levels of thermal variations under the influence of various emotive state and (ii) emotion-educed facial thermal patterns can be modeled as stochastically independent clusters to be separated as linear spaces and making automated detection of disgust-arousal possible.