Role of pupil dilation and facial temperature features in stress detection

In order to differentiate the affective state of a computer user as it changes from relaxation to stress, features derived from pupil dilation and periorbital temperature are processed with machine learning techniques. When absolute signal values are used together with entropy based features, the accuracy of affective classification is observed to increase. When decision tree (C4.5) is tested for classification, best accuracy of detection of neutral versus aroused states is above 90%.