University of Houston
Department of Computer Science

In partial fulfillment of the Requirements for the Degree of
Master of Science

Colin Puri
will defend his thesis

StressCam

Abstract
This thesis introduces a novel methodology for quantifying mounting stress of a computer user and within a Human-Computer Interaction (HCI) framework. The method is based on thermal imaging of the face. The sensor can be employed as computer peripheral, much like a video-conferencing camera. The sensing modality does not require contact with the subject and it is passive; therefore, monitoring can be continuous and transparent to the computer user without posing any health risk. We have found that mounting stress is correlated with increased blood flow in the frowning vessel of the forehead. This increased blood flow dissipates convective heat, which can be captured through thermal imaging. Feelings of frustration on the subjects was invoked through a computerized Stroop test. We computed a measure quantifying the relative difference in physiological activity on the frowning vessel between a baseline and a Stroop testing session for a series of subjects. We compared this against a gold standard, which we chose to be the Energy Expenditure (EE) indicator measured though a cardiopulmonary stress device.  We have found that the proposed novel and contact-free thermal measurement methodology is highly correlated with the established invasive EE methodology. Therefore, our method has the potential to be incorporated in HCI and gauge transparently frustration levels of the computer user. The importance of such an HCI augmentation towards building more responsive computers cannot be overestimated.

With this success in mind, we then aimed our thermal camera at another phenomenon of stress that becomes apparent as a subject is put under duress, perspiration.  The measurement of both of these indicators of stress has uses in HCI and other avenues such as the next generation of polygraphy testing as the nature of human habits are still being discovered. This work is sponsored by an NSF research grant on information assurance.

Date: 1, December, 2005
Time: 02:00 PM
Place: 550-PGH

Faculty, students, and the general public are invited.
Thesis Advisor: Dr. Ioannis Pavlidis