Event Location: NSH 1507
Bio: Amit Agrawal is a visiting research scientist at Mitsubishi Electric
Research Labs (MERL). He received the B.Tech degree in EE from Indian
Institute of Technology (IIT), Kanpur in 2000 and M.S. & PhD in ECE from
University of Maryland, College Park in 2003 and 2006 respectively. His
research interests are in computer vision, image processing and
computational photography. He organized a tutorial at ICCV 2007 on
‘Gradient Domain Manipulation Techniques in Vision and Graphics’. His
papers have appeared in several vision and graphics conferences including
SIGGRAPH, CVPR, ICCV and ECCV.

Abstract: Computational photography is an emerging field combining smart optics,
electronics, computer vision and graphics. Recent trends in enhancing the
capabilities of a camera involve combining multiple photos taken with
different camera parameters. In this talk, first I will describe gradient
domain algorithms to combine photos taken with and without a flash for
recovering high dynamic range images, removing glass reflections, shadows
and flash hotspots. Then, I will discuss techniques of coding and
modulation in cameras that go beyond such photo manipulations. I will
describe several new camera designs including (a) gradient camera for high
dynamic range imaging, (b) coded exposure camera for motion deblurring,
(c) coded aperture for extending depth of field and (d) mask-enhanced
cameras for capturing lightfields using optical modulation.

Joint work with Ramesh Raskar, Jack Tumblin, Ashok Veeraraghavan and
Ankit Mohan