|Continuous Convex Relaxation Methods
for Image Processing: Optimal Solutions and Fast Algorithms
Tony F Chan
|Visual Signal Analysis
and Compression: Rethinking Texture
Thrasyvoulos N. Pappas
|Multimedia Social Networking: A
New Paradigm for Signal and Image Processing
K. J. Ray Liu
Convex Relaxation Methods for Image Processing: Optimal Solutions
and Fast Algorithms
This talk will introduce recent methods to compute solutions to
fundamental problems in image processing. Several meaningful problems
in image processing are usually defined as /non-convex/ energy minimization
problems, which are sensitive to initial condition and slow to minimize.
The ultimate objective of our work is to overcome the bottleneck
problem of non-convexity. In other words, our goal is to "convexify"
the original problems to produce more robust and faster algorithms
for real-world applications. Our approach consists in finding a
convex relaxation of the original non-convex optimization problems
and thresholding the relaxed solution to reach the solution of the
original problem. We will show that this approach is able to convexify
important and difficult image processing problems such as multiphase
image segmentation based on the level set method and image registration.
Our algorithms are not only guaranteed to find global solution to
the original problem, they are also at least as fast as graph-cuts
combinatorial techniques while being more accurate. Joint work with
Prof Tony F Chan assumed the presidency of HKUST on 1 September
2009. Prior to joining HKUST, Prof Chan was Assistant Director of
the U.S. National Science Foundation (NSF) in charge of the Mathematical
and Physical Sciences Directorate, which is the largest directorate
at NSF. In that position, he guided and managed research funding
of about HK$10 billion a year in astronomy, physics, chemistry,
mathematical science, material science, and multidisciplinary activities.
Prof Chan's scientific background is in Mathematics,
Computer Science and Engineering. He received his BS and MS degrees
in Engineering from the California Institute of Technology (Caltech)
and his PhD in Computer Science from Stanford University. He pursued
postdoctoral research at Caltech as Research Fellow, and taught
Computer Science at Yale University before joining the University
of California at Los Angeles (UCLA) as Professor of Mathematics
in 1986. He was appointed Chair of the Department of Mathematics
in 1997 and served as Dean of Physical Science from 2001 to 2006.
He also holds honorary joint appointments with the University's
BioEngineering Department and the Computer Science Department.
Prof Chan was one of the principal investigators
who made the successful proposal to the NSF to form the Institute
for Pure and Applied Mathematics (IPAM) at UCLA. He served as IPAM's
Director from 2000 to 2001.
Prof Chan is an active member of many scientific
societies, including the Society of Industrial and Applied Mathematics
(SIAM), the American Mathematical Society, the Institute of Electrical
and Electronic Engineers (IEEE) and was elected as a SIAM Fellow
in March 2010. Prof Chan is also a member of Committee of 100 and
has served on the editorial boards of many journals in mathematics
and computing, including SIAM Review, SIAM Journal of
Scientific Computing, and the Asian Journal of Mathematics,
and is one of the three Editors-in-Chief of Numerische Mathematik.
He co-wrote the proposal to start a new SIAM Journal of Imaging
Sciences and serves on its inaugural editorial board. He formerly
served on the NSF Mathematical and Physical Sciences Advisory Committee
and the US National Committee on Mathematics, and represented the
US to the 2006 General Assembly of the International Mathematics
Union in Spain.
His current research interests include mathematical
image processing and computer vision, Very Large-Scale Integration
(VLSI) physical design and computational brain mapping. He has published
over 200 refereed papers and is one of the most cited mathematicians.
He has mentored over 25 PhD students and 15 postdoctoral fellows.
Signal Analysis and Compression: Rethinking Texture
||Thrasyvoulos N. Pappas
Electrical Engineering and Computer Science Department
Date: September 28, 2010
Time: 9:00am - 10:00am
The fields of visual signal analysis and compression have made significant
advances during the last two decades, incorporating sophisticated
signal processing techniques and models of human perception. One
of the keys to further advances is a better understanding of texture.
We examine a number of applications that critically depend on texture
analysis, including image and video compression, computer vision,
content-based retrieval, visual to tactile image conversion, and
We first look at image and video compression. Traditional
compression techniques have relied on point-by-point comparisons
-- whether in the original image domain or in a transform domain
-- that cannot adequately exploit the stochastic nature of texture.
We discuss the idea of "structurally lossless" compression
that allows significant differences between the original and decoded
images, which may be perceptible when they are viewed side-by-side,
but do not affect the overall quality of the image.
In computer vision and content-based retrieval,
texture is a key (low-level) perceptual attribute that plays a critical
role in material perception, and along with color and shape, in
the extraction of semantic information. Texture analysis is also
an essential element of a new segmentation-based approach for converting
images into tactile patterns. Such a conversion can dramatically
increase the amount of information that can be made available to
the visually impaired segment of the population. Finally, a better
understanding of the joint perception of visual, acoustic, and tactile
textures is critical for the development of multimodal interfaces
for the next generation of interactive environments for entertainment,
commerce, education, and medicine.
A key problem in all of the above applications is
the inability of existing metrics to quantify specific texture attributes
and to predict perceptual texture similarity. We discuss the development
of objective texture similarity metrics that allow substantial point-by-point
deviations between textures that according to human judgment are
virtually identical. Such metrics are essential for all of the above
applications, but we show that different applications impose different
requirements on metric performance. We also discuss the development
of metrics for texture attributes (perceptual dimensions), such
as directionality and roughness.
Thrasyvoulos (Thrasos) Pappas (www.eecs.northwestern.edu/~pappas)
received the S.B., S.M., and Ph.D. degrees in electrical engineering
and computer science from MIT in 1979, 1982, and 1987, respectively.
From 1987 until 1999, he was a Member of the Technical Staff at
Bell Laboratories, Murray Hill, NJ. In 1999, he joined the Department
of Electrical and Computer Engineering (now EECS) at Northwestern
University as an associate professor. His research interests are
in image and video quality and compression, image and video analysis,
content-based retrieval, perceptual models for multimedia processing,
model-based halftoning, and tactile and multimodal interfaces.
Dr. Pappas is a Fellow of the IEEE and SPIE. He
has served as an elected member of the Board of Governors of the
Signal Processing Society of IEEE (2004-2007), chair of the IEEE
Image and Multidimensional Signal Processing Technical Committee
(2002-2003), and technical program co-chair of ICIP-01 and ICIP-09.
He has also served as co-chair of the 2005 SPIE/IS&T Electronic
Imaging Symposium. Since 1997 he has been co-chair of the SPIE/IS&T
Conference on Human Vision and Electronic Imaging. Dr. Pappas has
served on the editorial boards of the IEEE Transactions on Image
Processing, the IEEE Signal Processing Magazine, the IS&T/SPIE
Journal of Electronic Imaging, and the Foundations and Trends in
Signal Processing. He is currently editor-in-chief of the IEEE Transactions
on Image Processing.
Social Networking: A New Paradigm for Signal and Image Processing
||K. J. Ray Liu
Department of Electrical and Computer Engineering
University of Maryland, College Park
Date: September 29, 2010
Time: 9:00am - 10:00am
Within the past decade, the proliferation of multimedia social network
communities, such as Napster, and YouTube where millions of users
form a dynamically changing infrastructure to share content, have
introduced the new concept of social networking that creates a technological
revolution as well as brings new experiences to users. The massive
content production poses new challenges to the scalable and reliable
sharing of multimedia content over large and heterogeneous networks.
It also raises critical issues of intellectual property protection
and privacy issues.
In a multimedia social network, users actively interact
with each other, and such user dynamics not only influence each
individual user but also affect the system performance. To provide
a predictable and satisfactory level of service, it is of ample
importance to understand the impact of human factors on multimedia
social networks. Such an understanding provides fundamental guidelines
to the better design of multimedia systems and networking, and offers
more secure and personalized services. For example, in a peer-to-peer
file-sharing system, users pool together the resources and cooperate
with each other to provide an inexpensive, highly scalable, and
robust platform for distributed data sharing. However, since the
nature of participation in many multimedia social networks is often
voluntary and unregulated, there is a need to provide incentives
and mechanism to stimulate cooperation among users to improve system
The influence of human behavior and factors has
seldom been recognized in signal and image processing research.
Therefore, first in this talk the goal is to illustrate why understanding
of human factors and behavior plays an important role in designing
and improving multimedia communications and security. Such a journey
leads us to reconsider many classical signal and image processing
problems from the concept/notion of social networking. The second
goal of the talk is to demonstrate that the social networking approach
can indeed offer a new and unified view to many classical problems
and has the potential of becoming a new signal and image processing
Dr. K. J. Ray Liu was named a Distinguished Scholar-Teacher of University
of Maryland in 2007. He leads the Maryland Signals and Information
Group conducting research encompassing broad aspects of wireless
communications and networking, information forensics and security,
multimedia signal processing, and biomedical engineering.
Dr. Liu is the recipient of numerous honors and awards including
the 1994 National Science Foundation Young Investigator Award; best
paper awards from IEEE and EURASIP; IEEE Signal Processing Society
2004 Distinguished Lecturer; EURASIP 2004 Meritorious Service Award;
and 2009 IEEE Signal Processing Society Technical Achievement Award.
A Fellow of the IEEE and AAAS, he is recognized by Thomson Reuters
as an ISI Highly Cited Researcher. Dr. Liu is President-Elect of
IEEE Signal Processing Society. He was the Editor-in-Chief of IEEE
Signal Processing Magazine and the founding Editor-in-Chief of EURASIP
Journal on Advances in Signal Processing.
Dr. Liu also received various research and teaching recognitions
from the University of Maryland, including Poole and Kent Senior
Faculty Teaching Award and Outstanding Faculty Research Award, both
from A. James Clark School of Engineering; and Invention of the
Year Award from Office of Technology Commercialization.
His recent books include Behavior Dynamics in Media-Sharing Social
Networks, Cambridge University Press (to appear); Cognitive
Radio Networking and Security: A Game Theoretical View, Cambridge
University Press, 2010; Handbook on Array Processing and Sensor
Networks, IEEE-Wiley, 2009; Cooperative Communications and
Networking, Cambridge University Press, 2008; Resource Allocation
for Wireless Networks: Basics, Techniques, and Applications,
Cambridge University Press, 2008; Ultra-Wideband Communication
Systems: The Multiband OFDM Approach, IEEE-Wiley, 2007; Network-Aware
Security for Group Communications, Springer, 2007; Multimedia
Fingerprinting Forensics for Traitor Tracing, Hindawi, 2005.