|
|
Motorola
Distinguished Lecture Series
-
Some Issues in Video Indexing, Retrieval, Filtering and
Summarization, Thomas S. Huang, University of Illinois at
Urbana-Champain
-
Game Theory for Engineering Systems, Jose B.
Cruz, Jr., Ohio State University
-
Class and Channel Condition Based Weighted
Proportionaly Fair Scheduler, Vijay Subramanian, Motorola, Inc.
Thomas S. Huang
Beckman Institute of Advanced Science and Technology
University of Urbana-Champaign
After an overview of some of the
important issues in video management, we concentrate on two topics:
content-based image-video retrieval using relevance feedback and
automatic annotation of video. We believe that an effective way of
retrieving images and video is to combine keywords with low-level
feature similarity. The two topics we describe in this talk contribute
to this approach.
Relevance feedback can be viewed as a
two-class classification problem (the relevant class and the irrelevant
class). The difficulty is that the number of training samples is very
small. A method will be described which attempts to make use of the
unlabeled samples to improve the performance of the classifier.
A probabilistic framework is proposed
for recognizing high-level concepts (such as indoor-outdoor, people,
explosion) from low-level features (color, texture, shape, structure,
layout, motion, audio). Correlations between the high-level concepts
(e.g., helicopters and outdoor tend to co-occur) are used to improve
recognition accuracy.
Finally, some comments on how to combine
keywords (high-level concepts) and low-level features in image-video
retrieval will be given.
Thomas S. Huang received his B.S.
degree in Electrical Engineering from National Taiwan University,
Taipei, Taiwan, China; and his M.S. and Sc.D. in Electrical Engineering
from the Massachusetts institute of Technology. He is the William L.
Everitt Distinguished Professor of Electrical and Computer Engineering
at the University of Illinois, Urbana-Champaign. He is also Research
Professor at the Coordinated Science Laboratory and Head of the Image
Formation and Processing Group at the Beckman Institute for Advanced
Science and Technology, where he is Co-Chair of the institute's major
research theme, Human Computer Intelligent Interaction.
Dr. Huang's professional interests
lie in the broad area of information technology, especially the
transmission and processing of multidimensional signals. He is a member
of National Academy of Engineering and a Fellow of the International
Association of Pattern Recognition, IEEE, and the Optical Society of
America. He has received a Guggenheim Fellowship, an A.V. Humboldt
Foundation Senior U.S. Scientist Award, and a Fellowship from the Japan
Association for the Promotion of Science. He received the IEEE Signal
Processing Society's Technical Achievement Award in 1987, and the
Society Award in 1991. He was awarded the IEEE Third Millennium Medal
in 2000. Also in 2000, he received the Honda Lifetime Achievement
Award for "contributions to motion analysis." In 2001, he received the
IEEE Jack S. Kilby Medal. He is a Founding Editor of the International
Journal of Computer Vision, Graphics, and Image Processing and Editor
of the Springer Series in Information Sciences, published by Springer
Verlag.
Tuesday, October 30, 2001
Technological Institute - Room L324
4:00PM (Reception Following)
Jose B. Cruz, Jr.
The Ohio State University
Complex engineering systems generally
contain several controllers, each with a different objective to
optimize. The controls needed to optimize one criterion generally differ
from the controls needed to optimize a different criterion. How can
multiple controllers with different optimality criteria steer the same
systems to operate optimally? Examples of systems that contain
multiple controllers with different optimizing goals include: (a)
Internet-type communication networks where non-cooperating individual
users choose their routing to optimize individual costs and a network
manager optimizes network efficiency, (b) electric energy networks where
deregulated energy suppliers, blocks of consumers, and utilities
develop biding strategies for selling and buying electricity in multiple
energy markets, (c) command and control hierarchies in military
operations. In this talk we will discuss basic concepts from game
theory, which underlies the development of control strategies for such
systems, present a few examples, and sketch and overview of the status
of research in this field.
Jose B. Cruz, jr. is the Howard D.
Winbigler Chair in Engineering, and Professor of Electrical Engineering
at the Ohio State University (OSU). He received his Ph.D. from the
University of Illinois in electrical engineering. He served as Dean of
the College of Engineering at OSU from 1992 to 1997, Professor of
Electrical and Computer Engineering at the University of California in
Irvine from 1986 to 1992, and faculty member at the University of
Illinois from 1956 to 1986. Dr. Cruz was elected as a member of the
National Academy of Engineering in 1980. He is also a Fellow, Institute
of Electrical and Electronics Engineers; Recipient, Curtis W. McGraw
Research Award of the American Society for Engineering Education;
Recipient, IEEE Richard M. Emberson Award; Fellow, American Association
for the Advancement of Science; and Recipient, Richard E. Bellman
Control Heritage Award, American Control Council in 1994.
Wednesday, April 25, 2001
Technological Institute - Room L324
4:00PM (Reception Following)
Vijay Subramanian
Motorola, Inc.
Efficient management of the radio
resource of a 3-G system is important from an operator's perspective.
This, however, cannot be the only concern with QoS negotiations have
been made for various users and the operator has to uphold these. This
leads to a fairness objective that the operator has to keep in mind. In
this talk we outline a scheme to perform packet-level scheduling and
resource allocation at the wireless node that takes into account the
notions of both efficiency and fairness and presents a means to explore
the trade-off between these two notions. As a part of this scheme we
see the scheduling problem as deciding not just the packet transmission
schedule but also the power allocation, the modulation and coding scheme
allocation and the spreading code determination since the latter three
directly influence the radio resources consumed. The proposed
scheduler takes advantage of the average and current channel conditions
to improve the system throughput while maintaining proportional fairness
among the users. Using a utility maximization formulation based on the
data-rates that the mobiles can transmit at, we decide on the weights
for a weighted proportionally fair allocation based scheduling
algorithm. We also show how one can adapt the weights and the
algorithm for a time-varying channel. We conclude with a simulation
based performance analysis for infinitely-backlogged sources on a UMTS
system as well as an EDGE system.
Vijay Subramanian got his B.Tech.
degree at IIT, Madras in 1993 and his M.Sc.(Engg.) from IISc., Bangalore
in 1995. He obtained his doctorate in Electrical Engineering from the
ECE Department at UIUC in 1999. From the end of Fall'99 he has been
with the Mathematics of Communication Networks group at Motorola,
Arlington Heights. His interests are in communication networks,
communications, information theory, queueing theory and applied
mathematics.
Thursday, April 5, 2001
Technological Institute - Room L324
4:00PM (Reception Following)
|
|
|
