Abstract

We have developed a non-invasive thermal image analyzer for deceptive detection (TAD²) where the far-infrared data around the periorbital and nostril areas are simultaneously analyzed. Measured change in maximum skin temperature around two periorbital regions is converted to a relative blood flow velocity. A respiration pattern is also simultaneously determined via the ratio of the measured maximum and minimum temperatures in the nostril area. In addition, our TAD² employs a simple normalized cross correlation scheme to independently track locations of the two periorbital and nostril areas. Our field case study from 7 subjects in two real crime scenes and with the use of our baseline classification criteria shows two-fold improvement in classification rate compared to our analysis using either the periorbital or nostril area alone.

© 2007 SPIE