Computational Intelligence and Bio-Identification Lab

Reza Derakhshani, Ph.D. and his research group in 454 FH (CIBIT Lab) conduct research in biomedical signal analysis, biometrics (physical and psychophysiological), and physiological system identification using computational intelligence paradigms such as artificial neural networks. Their recent projects include noninvasive brain computer interfacing, a new patented biometric modality based on vasculature on the white of the eye, data-driven modeling of human joints, and non-obtrusive psychophysiological pattern recognition using postural and ocular dynamics. A partial summary of Dr. Derakhshani’s research grants (current and recent projects) is given below.

Recent and Current Research

Post-mortem Ocular Biometric Analysis

Funded by: National Science Foundation Center for Identification Technology Research
Duration: 8/1/2010 to 12/31/2011 (projected)
Goal: This is a unique study of potential postmortem degeneration of ocular biometric identification metrics through analysis of multispectral time-lapse sequence of ocular captures ante to postmortem.

Computational Simulation of Canine Biomechanically Induced Unicompartmental Osteoarthritis: A Concurrent Multiscale Approach

Principal Investigator(s): Trent Guess, Ph.D.
Funded by: Missouri Life Sciences Review Board
Duration: 2/1/2009 to 2/1/2012
Goal: To develop a predictive, computationally efficient, patient level simulation tool of mechanical osteoarthritis indicators. Specifically, a computational model of the canine knee (stifle) that includes surrogate models of cartilage tissue behavior will be combined with musculoskeletal models of movement and validated through in-vivo canine models of osteoarthritis. The computationally efficient cartilage surrogates will predict key tissue indicators of osteoarthritis (shear stress, peak stress, and stress transients) in response to organ level loading and learn from a finite element model solution database.

Acquisition of an Experimental Platform to Support Research and Educational Activities in Human Motion

Principal Investigator(s): Trent Guess, Ph.D.
Funded by: National Science Foundation
Duration: 9/1/2008 to 8/31/2011
Goal: This major research instrumentation grant provided us with a state of the art comprehensive human gait capture lab. Equipments include Vicon MX-T40 (motion capture with comprehensive software suite), AMTI OR6-6 (force plate), Delsys Myomonitor IV (wireless EMG & HR), and 16TB of NAS.

Economical, Unobtrusive Measurement of Postural Correlates of Deception

Principal Investigator(s): Dr. Lovelace
Funded by: National Science Foundation Center for Identification Technology Research
Duration: 8/1/2009 to 4/1/2011
Goal: This project is a novel adaptation of force platform technology to analyze the postural shifts that may accompany deception. By applying non-linear, data-driven signal classification methods to traditional features of postural sway extracted from force platform data, this project will provide a reliable, efficient means of credibility assessment in security screening environments.

An Integrated Platform for a new Biometric Identification System based on Ocular Surface Vasculature

Funded by: University of Missouri’s Office of Research and Economic Development, IP Fast track Initiative
Duration: 8/1/2009 to 7/31/2010
Goal: This hardware-software development grant integrates computational imaging with adaptive optics for high resolution scans of ocular surface vasculature on the sclera, episclera, and conjunctiva and its fusion with iris recognition for bimodal biometric identification.

A Multimodal Electroencephalographic Brain-Computer Interface

Funded by: UMKC Research Board, Faculty Research Grant
Duration: 01/14/2007 to 10/14/2008
Goal: This research studies feature multimodality in geometric and spatiotemporal characteristic spaces of imagined tasks as reflected in EEG.

Bi-modal, Adaptive Thought Translation Device

Funded by: University of Missouri Research Board
Duration: 1/1/2007 to 6/1/2008
Goal: This study investigates the utility of various spatiotemporal EEG features for real-time detection of intended motor movements using data-driven machine learning methods.

Enhancing Iris Systems Using Conjunctival Vascular Patterns-Phase II

Funded by: National Science Foundation Center for Identification Technology Research
Duration: 1/1/2008 to 12/31/2008
Goal: This study investigates multi-algorithmic classification of conjunctival, episcleral, and scleral vasculature patterns photographed in the RGB-IR spectrum via texture-based image analysis.

An Acquisition Platform for Non-Cooperative, Long Range Ocular Biometrics

Funded by: National Science Foundation Center for Identification Technology Research
Duration: 1/1/2008 to 12/31/2008
Goal: This study investigates optical and computational requirements and methods for high-resolution imaging of the human iris in near infrared from distances up to 10 meters.

Psychophysiological Biometrics

Principal Investigator(s): Dr. Burgoon
Funded by: National Science Foundation Center for Identification Technology Research
Duration: 1/1/2008 to 12/31/2008
Goal: This study investigates computerized analysis of blink reflex and pupil dilation to detect individuals’ specific affective-cognitive states.

Dynamic Simulation of Joints Using Multi-Scale Modeling

Principal Investigator(s): Trent Guess, Ph.D.
Funded by: National Science Foundation
Duration: 09/01/2005 to 08/31/2008
Goal: This study investigates new data-driven, computationally efficient surrogate models of the human knee validated by cadaver knees.