ECE 637: Pattern Recognition
Electrical and Computer Engineering
Pattern recognition techniques are used to design automated systems
that improve their own
performance through experience. This course covers the methodologies, technologies, and
algorithms of statistical pattern recognition from a variety of perspectives. Topics including
Bayesian Decision Theory, Estimation Theory, Line
ar Discrimination Functions, Nonparametric
Techniques, Support Vector Machines, Neural Networks, Decision Trees, and Clustering
Algorithms etc. will be presented.
Students taking this course should have graduate standing in electrical and c
engineering. Specifically, students should be familiar with linear algebra, probability, random
process, and statistics (e.g., ECE 650 or its equivalent). In addition, programming experience
(MATLAB/C/C++) will be helpful.
, Hart and Stork,
, Second Edition, Wiley, 2001.
Useful supplementary books:
, Mc Graw
Hill, New York, 1997.
S. Theodoridis, K. Koutroumbas,
, Academic Press, 1999.
After completing this course, the students should be able to:
1. Understand basic concepts in pattern recognition
2. Gain knowledge about state
art algorithms used in pattern recognition research
Understand pattern recognition theories, such as Bayes classifier, linear discriminant analysis.
4. Apply pattern recognition techniques in practical problems.
Topics Covered/Course Outline
1. Bayesian Decision Theory
2. Estimation Theory
3. EM algorithms
4. Nonparametric Techniques
5. Linear Discriminant Functions
6. Support vector Machine
7. Neural Networks
8. Stochastic Learning
9. Algorithm Independent Learning
10. Unsupervised Learning
Relationship to Program Outcomes
This course supports the
achievement of the following outcomes:
a) Ability to apply knowledge of advanced principals to the analysis of electrical and computer
b) Ability to apply knowledge of advanced techniques to the design of electrical and computer
c) Ability to apply the appropriate industry practices, emerging technologies, state
design techniques, software tools, and research methods of solving electrical and computer
d) ability to use the appropri
the art engineering references and resources, including
IEEE research journals and industry publications, needed to find the best solutions to electrical
and computer engineering problems.
e) Ability to communicate clearly and use the appropri
ate medium, including written, oral, and
electronic communication methods.
f) Ability to maintain life
long learning and continue to be motivated to learn new subject.
g) Ability to learn new subjects that are required to solve problems in industry withou
dependent on a classroom environment.
h) Ability to be competitive in the engineering job market or be admitted to an excellent Ph.D.
January 12, 2003