TOCS Event-Silicon Valley Campus - Carnegie Mellon University

TOCS Event-Silicon Valley Campus - Carnegie Mellon University

TOCS Event


David Stork


Research Director - Computational Sensing and Imaging Group, Rambus Labs


October 29, 1:30 pm



CMUSV, Rm 118 [directions]

[login as guest]

Title: Ultra-Miniature Lensless Computational Sensors and Imagers

The traditional lensed camera obscura architecture-in which each point in the scene illuminates a very small region in the image plane thereby producing an image-has dominated the technology of imaging systems (cameras, microscope, telescopes, etc.) for nearly a millennium.  The recent revolution in computational imaging, in which sophisticated digital signal processing is co-designed with the optics, has led to a broad expansion in the types of imagers and sensors, their form factors, and especially their functionality. No longer must the sensed information "look good" like a traditional image; instead a final digital image is estimated or computed from the raw sensed signals. We describe a new class of computational imagers and sensors that rely on spiral odd-symmetry phase gratings integrated with CMOS photosensor matrices. These gratings have a number of special optical properties and preserve signals' Fourier information (up to the two-dimensional Nyquist rate).  Because our sensors are lensless, they can be far smaller in spatial volume than imagers based on lenses-as small as the cross section of a human hair. These sensors can, moreover, exploit the "Moore's Law" of reduction in pixel pitch further than can diffraction-limited lens-based imagers.  Most importantly, both the grating and signal processing can be tailored to a wide range of image sensing and pattern recognition problems. This talk will describe the special optics, signal processing, hardware prototypes and a number of application areas for these computational sensors.  

Joint work with P. Gill, P. Johnstone, A. Kumar, and T. Vogelsang

Speaker Bio:

Dr. David G. Stork is Distinguished Research Scientist and Research Director at Rambus Labs. A graduate of MIT and the University of Maryland, he has held faculty positions in physics, mathematics, electrical engineering, statistics, computer science, neuroscience, psychology and art and art history variously at Wellesley and Swarthmore Colleges and Clark, Boston and Stanford Universities. He holds 42 patents and has published eight books or proceedings volumes, including Pattern classification (2nd ed.) with R. Duda and P. Hart, and Seeing the light:  Optics in nature, photography, color, vision and holography with D. Falk and D. Brill.  He is a senior member of IEEE and ACM and a Fellow of SPIE and the International Association for Pattern Recognition.  The work presented in this talk received a Best Paper Award from SensorComm 2013