Jeffrey R. Paone, Ph.D.
|| Home | Research | Teaching | Biography | Personal ||
Face and iris biometrics perform very when the image is acquired from a close range - several feet from a face or inches from an iris. It becomes a challenge to take a high-resolution and good quality image from ten to fifteen feet away from a subject. My research in this area developed a system to automatically track a subject across a room while capturing images of their face, iris, and fingerprints at a distance up to fifteen feet. This project made use of the Microsoft Kinect for skeletal tracking.
Camera Calibration & Naturalistic Driving
The Second Strategic Highway Research Program (SHRP2) collected two years of naturalistic driving video data. The videos were collected from four different cameras placed within the vehicle. We were responsible for developing a camera calibration model that accurately represented the intrinsic and extrinsic properties of all four cameras. Several performer teams then used this calibration data to develop tools to identify elements both inside and outside the vehicle. Our team created baselines to compare the performer teams' results.
My research examined the biometric menagerie and its stability under varying conditions. The conditions under scrutiny are when the illumination or subject's expression changes, or when the images representing an individual subject in the dataset are changed. The biometric menagerie is a classification system that labels subjects based on their matching tendencies.
Last Updated: August 13, 2015