“Big Data Computational Imaging: Designing Instruments and Algorithms for Gigapixel Cameras and Real-time Millimeter-wave Radar”
By Dr. Daniel Marks, Associate Research Professor, Electrical and Computer Engineering Dept., Duke University
Fri. April 10th, 2015 from 3:05 – 3:55 p.m. in WEB 1230
Abstract
Gigapixel cameras stream billions of pixels per second of data which have to be assembled into contiguous, seamless video frames. Real-time millimeter-wave radar systems, designed for checkpoint security, capture millions of samples per second, which must be rapidly decoded to determine if threats objects are present. The solutions to these imaging problems are demonstrated by a process of joint design of the physical layer and the use of graphics processing unit accelerated algorithms. New capabilities are obtained by designing the instrumentation and inference methods to complement each other. An example of these new capabilities are the DARPA AWARE Wide-Field project gigapixel cameras, based on microcamera arrays that require image tiling algorithms to produce a seamless, contiguous image. A second example is a novel real-time W-band millimeter-wave security scanner based on frequency diversity.
Biography
Daniel Marks received his BS in Electrical Engineering from the University of Illinois at Urbana-Champaign in 1995, his MS from UIUC in 1998, and a PhD in Electrical Engineering from UIUC in 2001. He is currently an Associate Research Professor at the Electrical and Computer Engineering Department at Duke University in Durham, NC. He is currently an editor of Applied Optics, and has 85 research articles and 18 patents or patents-pending in areas including optical design, optical coherence theory, nonlinear optics, image and signal processing, optical coherence tomography, X-ray and coded aperture imaging, compressive sensing, holography, and microwave and millimeter wave imaging, and teaches graduate courses in optical design and optical coherence theory.