Translator Disclaimer
23 April 2010 Biometric study using hyperspectral imaging during stress
Author Affiliations +
To the casual observer, transient stress results in a variety of physiological changes that can be seen in the face. Although the conditions can be seen visibly, the conditions affect the emissivity and absorption properties of the skin, which imaging spectrometers, commonly referred to as Hyperspectral (HS) cameras, can quantify at every image pixel. The study reported on in this paper, using Hyperspectral cameras, provides a basis for continued study of HS imaging to eventually quantify biometric stress. This study was limited to the visible to near infrared (VNIR) spectral range. Signal processing tools and algorithms have been developed and are described for using HS face data from human subjects. The subjects were placed in psychologically stressful situations and the camera data were analyzed to detect stress through changes in dermal reflectance and emissivity. Results indicate that hyperspectral imaging may potentially serve as a non-invasive tool to measure changes in skin emissivity indicative of a stressful incident. Particular narrow spectral bands in the near-infrared region of the electromagnetic spectrum seem especially important. Further studies need to be performed to determine the optimal spectral bands and to generalize the conclusions. The enormous information available in hyperspectral imaging needs further analysis and more spectral regions need to be exploited. Non-invasive stress detection is a prominent area of research with countless applications for both military and commercial use including border patrol, stand-off interrogation, access control, surveillance, and non-invasive and un-attended patient monitoring.
© (2010) COPYRIGHT Society of Photo-Optical Instrumentation Engineers (SPIE). Downloading of the abstract is permitted for personal use only.
Sheela Nagaraj, Shafik Quoraishee, Gabriel Chan, and Kenneth R. Short "Biometric study using hyperspectral imaging during stress", Proc. SPIE 7674, Smart Biomedical and Physiological Sensor Technologies VII, 76740K (23 April 2010);

Back to Top