CSCI
341
(W
'13), Digital Image Processing
Winter 2010
|
Dr. Joshua Stough , Parmly 408
|
MWF 2:30 - 3:25PM
Office Hours: M 3:30-, F 10-12, or ANYTIME |
stoughj@wlu.edu, x8811 or (919) 357-0604 (txt welcm) |
Location: Parmly 405
|
http://www.eg.bucknell.edu/~jvs008/teaching/CS341W13/CS341W13.html
|
Course Schedule
Overview
A survey of topics in the acquisition, processing and analysis of
digital images, with much of the necessary mathematical background
developed in the course. Topics in image processing include image
enhancement and restoration, compression, and
registration/alignment. Topics in image analysis include
classification, segmentation, and more generally statistical pattern
recognition. Throughout the course, human vision and perception
motivate the techniques discussed.
Specific topics to be covered in more or less depth:
- human visual sytem and perception
- filtering and contrast enhancement (windowing, blurring,
sharpening, histogram methods)
- segmentation (Canny edges, face detection, Hough transform,
Bayesian segmentation formulation)
- registration (transforms, sum of squared-differences,
correlation, corner detection, homographies/photo stitching)
- image compression (LZW, Fourier, wavelets)
At the end of the course, you will know some of these topics in detail,
while you will have a passing knowledge of the purpose and context of
others.
Is CSCI 341 For You?
- Strongly encourage Linear Algebra coreq, although this course
requires CSCI 111 or equivalent.
- As an upper level CS elective, it would be helpful in this course
to have had some calculus and linear algebra (how to take a derivative
of a
simple function, matrix algebra).
- Mathematical and statistical entities like probability
distribution, histogram, Gaussian, and others will be developed in
class insofar as we need them.
- We will be using MATLAB for all our work.
Time Commitment
During the course of CSCI 341, you will work on several
thought-intensive projects.
This can be a time-consuming process. Expect to spend 10-12 hours on
each
programming assignment (depending on the specific
assignment).
Textbooks
Required: Gonzalez, Woods: Digital Image Processing, 3rd Edition (2008)
(available through the bookstore or amazon).
Potentially useful: Richard
Szeliski, Computer Vision: Algorithms and
Applications.
Attendance, Grading, Late
Assignments
- Attendance is required.
- Grading breakdown: Exams 50% (2
x
10,
1 x 30), Assignments 40%,
Attendance and Participation 10%.
- Late Assignments lose 10, 15,
25, 25, 25% for each additional day late (no credit on the
fifth day). This scale may be delayed given the severity of your
circumstances and my being
informed of them in a timely manner.
- I will defer to
the Counseling Center (see here).
- If you have an athletic event and will not be able to make a
deadline, you should tell me within a day of an assignment being
posted.
- Accomodations for qualified disabilities requires written
notfication in conjunction with Student Academic
Support (see here or call
458-8754).
Assistance from others, group work,
etc.
Computer Science is a difficult subject to learn on your own (like many
others). Talking through problems with mentors and peers can be
an exceptional learning tool. However, this sort of collaboration
can also be quite dangerous, as you may be underprepared to be tested
on your own. Thus, here are my guidelines for appropriate
collaboration, in order to better help you learn. Conduct in
gross violation of these guidelines will be considered a breach of
trust. In the below, an implementation
refers to a detailed explanation (such as code) of the programming
structures and object interactions for solving a problem.
- Your only sources for this course should be the book, links I
provide to you through the course schedule and email, other students in
this course with you, and
me.
- You are not to seek assistance
from students who have previously taken this course, with the exception
of the TAs for the course.
- At the end of the term, YOU will
be someone who has previously taken this course. In the future,
please do not provide assistance to future students, and encourage them
to seek appropriate help.
- Do not use other online materials. It is likely that coded
solutions to many of the homeworks are readily available on the
web. However, this material obviously constitutes implementation
details.
- I encourage you to discuss with each other broad solutions to
problems on assignments. However, when you sit down to code your
solution, it should be your work. Please acknowledge aid received
from other students. It will have no effect on my grading, and it
is the honorable thing to do.
- I have found that in the vast majority of
non-typo cases, the mental exercise of running sample data through your
code illuminates the issue and leads to better understanding on your
part. Ignoring this exercise for convenience will hurt you. I
will help you with this exercise.
- After completing the assignment, feel free to discuss any aspects
of the assignment, including implementation details with others who
have also completed the assignment.