A novel approach using mechanical physiological activity as a biometric marker is described. Laser Doppler Vibrometry
is used to sense activity in the region of the carotid artery, related to arterial wall movements associated
with the central blood pressure pulse. The non-contact basis of the LDV method has several potential benefits in
terms of the associated non-intrusiveness. Several methods are proposed that use the temporal and/or spectral
information in the signal to assess biometric performance both on an intra-session basis, and on an intersession
basis involving testing repeated after delays of 1 week to 6 months. A waveform decomposition method that
utilizes principal component analysis is used to model the signal in the time domain. Authentication testing
for this approach produces an equal-error rate of 0.5% for intra-session testing. However, performance degrades
substantially for inter-session testing, requiring a more robust approach to modeling. Improved performance
is obtained using techniques based on time-frequency decomposition, incorporating a method for extracting informative
components. Biometric fusion methods including data fusion and information fusion are applied in
multi-session data training model. As currently implemented, this approach yields an inter-session equal-error
rate of 9%.
Small movements of the skin overlying the carotid artery, arising from pressure pulse changes in the carotid during the cardiac cycle, can be detected using the method of Laser Doppler Vibrometry (LDV). Based on the premise that there is a high degree of individuality in cardiovascular function, the pulse-related movements were modeled for biometric use. Short time variations in the signal due to physiological factors are described and these variations are shown to be informative for identity verification and recognition. Hidden Markov models (HMMs) are used to exploit the dependence between the pulse signals over successive cardiac cycles. The resulting biometric classification performance confirms that the LDV signal contains information that is unique to the individual.
A hybrid sustained attention task was developed in order to examine the relationships between manual response times and the timing and morphology of horizontal saccades involved in shifting gaze to a source of task relevant visual information. Twelve subjects performed this task for 60 min with no breaks. Performance and gaze control measures were aggregated across 20 min intervals comprising early, middle and late segments of the task. Response time variability was significantly increased during later task segments (p<0.05). These segments were also associated with increased variability in the amplitude of saccades (p<0.05). Saccade durations during the late task segments were also longer and more variable (p<0.05). Correlations between response times and measure of saccadic activity were also computed across consecutive 5 min intervals for each individual subject. The obtained correlations between saccade latency and response times exceeded 0.70 for six of the twelve subjects. Additional analyses examined the relationship between trials characterized by extreme values on either the performance or the gaze control measures. Trials characterized by extremely long response times were also associated with increased saccade amplitudes, durations and latencies (p<0.01). Conversely, response times were abnormally long on trials categorized as extreme on the basis of the saccade morphology and timing measures (p<0.01). These results confirm the utility of the sustained attention task as a laboratory platform for the development of real-time systems for alertness monitoring. The data also support the contention that measures of gaze control behavior can reflect aspects of cognitive activity and, therefore, should be seriously considered for inclusion in any physiologically-based alertness assessment battery.