BOWIE STATE UNIVERSITY

chardfriendlyAI and Robotics

Oct 16, 2013 (3 years and 8 months ago)

60 views

BOWIE STATE UNIVERSITY

Department of Computer Science

Course Syllabus


COSC 890

(3 Cr)


Selected Topics: Advanced Topics in Pattern Recognition


Instructor:

Manohar Mareboyana

Office:


Computer Science Building Room 225

Phone:


301
-
860
-
3971

Email:


mmareboyana@

bowiestate.edu

Office Hours:

Semester dependent


Course Prerequisites:

COSC504


Catalog Description
:
This course introduce
s
statistical pattern recognition

with
applications in
image processing

and remote sensing
.
It includes supervised learning using Bayesian decision
theory, parametric estimation, and nonparametric techniques in probability density estimations.
Techniques for handling multidimensional data of various types and scales along wit
h algorithms
for clustering and classifying data will be explained.

Unsupervised learning and clustering are
discussed with examples from computer vision. Classification of multidimensional data using
Perceptron, multi
-
layered neural networks, and suppor
t vector machines (SVM) will be
discussed.


POLICIES REGARDING ATTENDANCE, GRADING, HOMEWORK AND ACADEMIC
INTEGRITY etc.


Evaluation
:

Following is the Evaluation system for the Final Grade. Each homework will be
graded. Students are responsible for co
mpleting them as scheduled.


Homework/Assignments



10%

Projects:





20%

Quizzes:





20%

Mid
-
Term Exam




20%

Final Exam





30%


Projects, Mid
-
term and Final exams are mandatory. Mid
-
term will not include the grades of
projects.



Grading
:

Academic
d
ishonesty will result in grade F. The grade levels used in evaluating
students’ work are:

90
-

100

A

80
-

89

B

70
-

79

C


Final grades will be computed based upon credits earned for all the five components mentioned
above.



Attendance
: Regular attendanc
e in the class is mandatory. Students will be responsible for any
loss of information, assignments, and projects due to absence from class. There will be no
make
-
up for any missed classes, projects, assignments, and exams.


Help
: During the office hours o
r by appointment. Email can be used for help but I check emails
prior to schedule class meeting. Please use the homework email to email me. The reply may
take more than 24 hours. While sending email, always mention your name and a proper
subject that ref
lects the problem in brief. Always use teacher’s email address given in this
syllabus. The instructor will check emails from students only at this address. Also, due to
virus problems, student should use only that address that is given to the teacher in St
udent
Information Sheet by the student. If you send from some other address, it might go to junk
area and you may not get a response.


Academic Dishonesty
: Academic dishonesty includes plagiarism, cheating, and other
illegal or unethical behaviors in doing

the work of the course. Plagiarism is the act of
representing another's ideas, words or information as one's own. If you receive
assistance on an assignment from someone else, you must avoid plagiarism by giving
proper credit for this assistance. Includ
e in your assignment a comment naming the
person who assisted you and stating what the assistance was. Students who are guilty of
academic dishonesty are subject to severe penalties ranging from a reduction in points
(and possible failure) for the assignm
ent/project, to failing the course, or in extreme
cases, dismissal from the University. Do not copy other student's projects, codes, and
design. A group of students working together on a project must change their forms and
codes to differentiate from other
s.


Inclement weather conditions
: In case of inclement weather conditions, call the
following number regarding cancellations: (301) 860
-

4000.



Important Numbers
:

Dept

of

Comp Sc (Secretary):

(301)

860
-
3961

Dept

of

Comp Sc (Fax):

(301)

860
-
3979



Bla
ckboard Information
:

Website:
http://classroom.bowiestate.edu

Help Desk: (301) 860
-
4357

It will be the student's responsibility to logon and be comfortable with the use of Blackboard established for
them. Call

the help desk if there is any problem in accessing the Blackboard. Also, be comfortable with zipping
the files as students are required to submit all the zipped Project files on Blackboard using digital drop box.


ADA statement
: Students with disabilitie
s who wish to receive ADA accommodations should
report to the Office of Special Populations, CLT Building Room 311 (telephone 301
-
860
-
3292)



The Important Dates

Late Registration



TBA

Last day to add a Class



TBA

Last Day to drop without a W


TBA

Fall C
onvocation



TBA

Last Day to change from Credit to Audit

TBA

Last day of Classes



TBA

Final Exams




TBA


Required text
: Richard O. Duda, Peter E. Hart, and David G. Stork, “Pattern Classification”,
Second Edition, John Wiley and Sons, Inc, 2001, ISBN 0
-
4
71
-
05669
-
3


Teaching Modes
:


The primary teaching mode will be seminar
-
discussion, problem solving. This will be
supplemented by out of class conferences, discussions.


Course

Objectives
: Upon successful completion of the course, the students will be abl
e to:

1. demonstrate the concepts in Bayes decision theory

2. demonstrate parameter estimation and non
-
parametric techniques

3. demonstrate and implement dimesionality reduction techniques

4. demonstrate knowledge unsupervised learning techniques

5. d
emonstrate advanced pattern classification techniques such as neral networks, SVM
using matlab


REQUIREMENTS:

Students are expected to:


1.

Attend all classes, participate in class discussion, and otherwise be prepared.

2.

Read all assigned selections before co
ming to class on the day those selections are to be
discussed.

3.

Submit all assignments when due (1/2 letter grade off for each day late without
documented excuse; papers more than one week late will not be accepted.)

4.

Implement projects using the software p
rovided.


COURSE OUTLINE & SCHEDULE:

(schedule subject to change with due notice) There will
be daily homework assignments administered during class.


Week

Topic

1

Theoretical

foundations of
pattern recognition


2

Bayesian decision theory

3

Maximum like
lihood and Bayesian estimation

4

Non
-
parametric techniques

for probability density estimation

5

Linear discriminant functions

6

Multilayer neural networks


7

Mid
-
term exam

8

Multilayer neural networks

Contd.

9

Dimensionality Reduction

10

Unsupervi
sed Learning Data Clustering


11

Unsupervised Learning Data Clustering

Contd.

12

Support Vector Machine

13

Support Vector Machine Contd.

14

Review for the final Exam

15

Final Exam








Bibliography:


1.

Christopher M. Bishop,

Pattern Recognition and

Machine Learning

Springer 2006.


2.

A. K. Jain,
Fundamentals of Digital Image Processin
g
, Prentice Hall1
,1989.

3.

B. Scholkopf and A. J. Smola,

Learning with Kernels
, MIT Press, 2002.

4.

N. Cristianini and J. Shawe
-
Taylor,
An Introduction to Support Vector Machi
ne,
Cambridge University Press, 2000

5.

IEEE Transactions on Image Processing (Periodicals from IEEE press)

6.

IEEE Transactions on Pattern Analysis and Machine Intelligence (Periodicals from IEEE
Press)