Non-intrusive Physiological Monitoring for Automated Stress Detection in Human-Computer Interaction

Abstract

Affective Computing, one of the frontiers of Human-Computer Interaction studies, seeks to provide computers with the capability to react appropriately to a user's affective states. In order to achieve the required on-line assessment of those affective states, we propose to extract features from physiological signals from the user (Blood Volume Pulse, Galvanic Skin Response, Skin Temperature and Pupil Diameter), which can be processed by learning pattern recognition systems to classify the user's affective state. An initial implementation of our proposed system was set up to address the detection of "stress" states in a computer user. A computer-based "Paced Stroop Test" was designed to act as a stimulus to elicit emotional stress in the subject. Signal processing techniques were applied to the physiological signals monitored to extract features used by three learning algorithms: Naïve Bayes, Decision Tree and Support Vector Machine to classify relaxed vs. stressed states.

A prototype for a conversational companion for reminiscing about images

“…The model will be constructed to test the underlying assumptions of the more complex proposal, demonstrating an initial first step that acknowledges the technical chal-lenges (Greasley et al, 2000;Cowie et al, 2001;Hopkins et al, 2005;Barreto et al, 2007) remaining within this field. We believe that this implementation demonstrates how various, future, technical improvements can be exploited by a complete dialogue model.…”

Bio-sensing for Emotional Characterization without Word Labels

“…To create the data set, we computed features from the recorded signals. The features we computed were the same as the ones that are assessed by Barreto et al [12] with the only difference that we did not consider each movie as a complete segment from which each feature is computed over the whole segment, but we assessed the signals in intervals of 40 seconds, in order to have a sequence of feature values for each elicited emotion.…”

“…Figure 3 shows an example of such a GSR response. An electrodermal response is often described with a few specific characteristics from these responses: amplitude, rise time and the half-recovery time [12]. The specific features that we computed from the GSR signal are: the number of GSR responses, Mean value of the GSR, average Amplitude of the GSR responses, average Rising time of the GSR responses and the average Energy of the responses (the total area under the half-recovery time).…”

“…The specific features that we computed from the GSR signal are: the number of GSR responses, Mean value of the GSR, average Amplitude of the GSR responses, average Rising time of the GSR responses and the average Energy of the responses (the total area under the half-recovery time). All these features were computed as described in [12].…”

Detection of Sympathetic Activation through Measurement and Adaptive Processing of the Pupil Diameter for Affective Assessment of Computer Users

“…The postprocessing operated on the MPD signal to obtain the PMPDmean feature is detailed in the next section. The processes to obtain the remaining 9 features are reflected in their names (further details for the calculation of these 9 features can be found in [16]. They are not included in this paper as the emphasis is on the PD signal).…”