M633 Industrial and Applied Mathematics
Spring 2007
Special Project on the Mathematics of
Data Compression Algorithms
Individual teams will
investigate (analyze and implement) algorithms for data compression with an
emphasis on their mathematical underpinnings.
We will be particularly interested in the distinction in the compression
of spatial data (imagine digital snapshots) versus spatiotemporal data (such as
video). This is a project based course
and students will work together to develop a Matlab GUI
for testing data compression algorithms on different data sets.
Instructor: Michael Kirby
Meeting Time: T, R
Prerequisites: M530 or M560
or M561
Students will be evaluated based on weekly progress
reports summarizing their contribution to the solution effort (25%), class
participation (15%), final team project report (50%), final
individual presentation summarizing semester long contributions to the project
(10%).
The students in this class will propose and implement
a novel solution to a problem
in industrial or applied mathematics. In spring 2007 this problem will be video
compression.
Week 1. Problem presentation.
Week 2. Class formulates problem statement and proposes
several alternative methods for proceeding.
Relevant papers in literature identified.
Week 3. Scope of work
defined. Milestones identified with
timetable. Identify minimal solution and
levels of enhancement.
Weeks 4-6.
Develop solution methodology.
Weeks 7-8. Team code design. Distribution of programming
assignments.
Weeks 9-11. Computer coding and
alpha testing of computer code.
Discuss and interpret intermediate results. Modify solution approach as necessary.
Week 12. Beta testing of computer solution.
Week 13. First draft of final report will be discussed.
Week 14. Assess minimal solution and prospects for
enhancement.
Week 15. Potential
future work plan outlined.
Week 16. Class presentations. (During finals week).