EE381K Multidimensional Digital Signal Processing - FAQ
What are the pre-requisites?
Essentially, you would need to have a one-semester graduate-level
class in digital signal processing at the level of Discrete-Time
Signal Processing by Alan Oppenheim and Ronald Schafer.
That class covers the essential theory and algorithms for
one-dimensional digital signal processing.
- Pre-requisite #1: EE381K-8 Digital Signal Processing
(sampling, z-transforms, discrete-time Fourier transforms,
discrete Fourier transforms, and fast Fourier transforms)
- Pre-requisite #2:
A class on image processing, Fourier optics, or system theory.
What does this course cover?
The first half of the course covers theory and algorithms of
multidimensional signals, systems, and transforms.
The second half covers a series of applications of multidimensional
systems in the fields of image, video, and seismic processing.
The course covers the following topics:
multidimensional signals and systems, multidimensional discrete
Fourier analysis, discrete cosine transform, two-dimensional FIR
filters, multidimensional recursive systems, two-dimensional IIR
filters, beamforming, seismic processing, tomography, multidimensional
multirate systems and filter banks, image halftoning, video processing
What's the trick for doing well in the course?
Do well on the project.
In the course, 50% of the grade is based on the project. The project
will consist of two equal parts: a literature survey due just before
the middle of the semester and a final report due on the last class
day of the semester. I have several ideas for projects on the course Web
pages, and I recommend (but not require) that you work in teams of
two or three. Both reports will be no more than 8 pages long of
double-spaced, 12pt font text.
The key to doing well on the project is to do a project related
to any research, development, or design work you are doing
at UT or at your company.
Other keys include getting started as soon as possible, finding
a fun topic, and recruiting a good team.
What's the benefit of working so hard on the project?
A side benefit to doing well on the project is that you might produce
work that will be published.
I set the length of the project reports to match the length required
for papers for IEEE signal and image processing conferences and
for correspondence items (short journal papers) for IEEE journals.
To date, 13 of the 51 projects (25%) completed in this course
have been published in IEEE conferences and journals.