"Color Imaging Fundamentals"

by Dr. Farhan Baqai, Friday, April 15, 2005

Abstract
This talk is intended as a precursor to my talk to Monday (April 18). I
will introduce the concepts of color imaging, describe the inter-disciplinary
nature of the field, and provide motivation for various problems in color
image processing. I will start by discussing the pros and cons of color
spaces. The remainder of the talk will be biased towards the role of
spatial characteristics in color perception. Spatial effects play an important,
but frequently ignored, role in color vision. I will briefly cover the optics
of the eye, contrast sensitivity functions of the human visual system, the
effect of surround on color appearance, and describe how spatial factors
can be incorporated into vision and color appearance models. Finally, I will
point out application areas where these concepts can be applied.

Digital Imaging: High Dynamic Range Imaging, Color Constancy, and Printing

by Dr. Farhan Baqai, Monday, April 18, 2005

Abstract
In the last decade, digital imaging has become ubiquitous. It is hard to
find imaging technologies that really are not digital or electronic. It is
now possible to obtain very high quality prints of digital images from a
cheap ink-jet printer. Recently, the explosion of consumer digital cameras
has placed new pressures on designers. There are a lot of very interesting
technical challenges. Consumers don't want to give up features when cost
drops. There is a relentless push for more and more quality. As digital
images flow around the world, thanks to improvements in wireless and
wireline communication technology, researchers are developing better and
faster algorithms for presenting, storing, searching, and verifying those
images.

In this talk, I will review some of my digital imaging research in
important areas such as high dynamic range imaging, colors constancy, and printing.
For high dynamic range imaging and color constancy, I will present our
modifications to the Frankle-McCann Retinex algorithm and show that our
extensions suppress artifacts and reduce the number of iterations required
to get to the final image. I will also discuss an algorithm that combines
Retinex with fractal compression and decompression. This algorithm does
the recoloring of the image, without an assumption of piecewise smoothness,
which is inherent in most color constancy algorithms. Furthermore, we can
simultaneously perform color constancy and compression. For rendering,
first I will show that by using a stochastic printer model within an iterative
halftoning scheme, we can obtain printed images of excellent quality, and
secondly I will discuss a systematic method of designing artifact-free
clustered-dot color screens based on lattice theory, a color device model,
and a spatio-chromatic human visual system model.

Biography

Farhan Baqai received the BSEE degree from UET, Lahore, Pakistan, the
M.Eng.Sc. degree in Electrical and Electronic Engineering from the
University of Melbourne, Australia, and the M.S. and Ph.D. degrees in
Electrical and Computer Engineering from Purdue University, West
Lafayette, IN. He has worked for Carrier Telephone Industries, Pakistan, GIK
Institute of Engineering Sciences and Technology, Pakistan, and Xerox Corporation in
Webster, NY. Since 2001, he is with US Advanced Technologies Center of
Sony Corporation in San Jose, CA, where he is engaged in the design of signal
and image processing algorithms for digital photography.

Dr. Baqai is a Senior Member of the IEEE, a member of the IEEE Signal
Processing Society, and the Society for Imaging Science and Technology.
His research interests include color imaging, digital
photography, statistical signal processing, multi-dimensional signal
processing, and radar imaging.