CS 4330: Computer Vision
Fall 2008 (FY09-1)
Introduction to the Course
Synopsis and Course Goal
This course provides an overview of Computer Vision – enabling computers to understand visual input.
The course introduces the major concepts and methods, and gives you a sample of problem settings
and techniques for solving them. It is intended for those who want to get an understanding of how
computer vision works and how it can perform such diverse tasks as object recognition, robot
navigation, and increase situational awareness. The course has strong ties between theory,
examples of its application, and practical exercises in order to promote fast and persistent
learning.
The course is intended for students with a basic proficiency in linear algebra and a programming
language. Example courses at NPS: MA2043 or MA3042
Course Learning Outcomes
What tasks do you need to master in order to achieve the overall course goal? Upon successful
completion of this course, you will be able to:
- understand the potential and limitations of CV
- evaluate the usefulness and performance of CV methods
- understand models of the human visual system
- relate computer vision to computer graphics
- comprehend and evaluate a computer vision system’s technical description
- discuss recent advances in computer vision and put them in context
- converse with practitioners and researchers in the field
- build real-time, interactive applications with a software library called OpenCV
- utilize several CV tools including Matlab
Syllabus and Organization
Take a look at the Contents on the following pages to get an idea of the syllabus. Computer vision is a
“hot topic” and rapidly evolving, so expect several modifications that reflect recent developments and
advances. Every chapter is a learning module with specific learning objectives, reading and study
material, exercises, and graded assignments.
This is a preliminary schedule for the quarter. Reading assignments are from various textbooks,
see :
SS = Shapiro & Stockman,
BB = Ballard & Brown,
DIP = Gonzalez & Woods (Digital Image Processing),
DIPuM = Gonzalez et al. (Digital Image Processing using Matlab).
The reading report is due before class on the following Monday. For example, SS1 and BB1 are to
be read in the week of 09/24 and the reading report must be in my hands before class on
10/01.
|
|
|
|
|
|
| Week | Date | Modules | Topic | Reading | Assignments (due date) |
|
|
|
|
|
|
|
|
|
|
|
|
| 1 | 09/29 | 1, 2 | Introduction to Computer Vision, Pixel Operations | SS1, BB1 | 10/06: survey, C++ quiz, math quiz, OpenCV Hello World |
|
|
|
|
|
|
| 2 | 10/06 | 3, 4 | M. Shah visit, Morphology, Spatial Filtering | SS3, DIPuM3 | 10/09: bg model, histogram, 3x3 average |
|
|
|
|
|
|
| 3 | 10/13 | 4, 5 | Spatial Filtering, Color | DIP6 | (Monday is Columbus Day) 10/16: Gaussian, Laplacian, orientation histogram |
|
|
|
|
|
|
| 4 | 10/20 | 6 | Segmentation | SS10, DIPuM6 | 10/23: color-based spatial segmentation |
|
|
|
|
|
|
| 5 | 10/27 | 7, 8 | Fitting, determistic and probabilistic | Hough | 10/30: model fitting |
|
|
|
|
|
|
| 6 | 11/03 | 9 | Shapes and Regions, PCA | SS7 | Wednesday 11/05: midterm (excl. S+R) |
|
|
|
|
|
|
| 7 | 11/10 | 10 | Texture | SS4 | midterm discussion; 11/13: texture-based classification (Gabor filters) |
|
|
|
|
|
|
| 8 | 11/17 | 7,11 | Imaging Sensors, Camera Transformations | H&Z | |
|
|
|
|
|
|
| 9 | 11/24 | 12 | Classification and Learning TBD | (Thursday is Thanksgiving) |
|
|
|
|
|
|
| 10 | 12/01 | 12 | project | TBD | |
|
|
|
|
|
|
| 11 | 12/08 | 13 | Capstone | unconventional optics | 12/10: Final Exam |
|
|
|
|
|
|
| |
Textbook, Web and Data Space
The required textbook for this course is Shapiro & Stockman: Computer Vision (see below). We will
also work with a reader and several resources provided to you in class. Grades for homework
assignments etc. will be on Blackboard. Software and other course material can be found here:
http://www.movesinstitute.org/~kolsch/courses/CS4330
and here:
\\comfort.ern.nps.edu\cs4330$
Hours
Lectures and lab:
Monday, Wednesday 1000-1150, WA-275
These two time slots will be used for lectures and/or lab time. Please bring your laptop for every
meeting.
Final:
Wednesday 12/10/07 in class
Holidays:
Mon 13 Oct 2007 (Columbus Day),
Tue 11 Nov 2007 (Veteran’s Day),
Thu 27 Nov 2007 (Thanksgiving)
Office Hours:
I am available for questions and help whenever I am in my office (WA-279) and have some time. If
you would like a firm appointment, please don’t hesitate to set one up by email or phone
(656-3402).
Grading
Grades for homework assignments etc. will be on Blackboard:
50% lab projects incl. reports
10% homework assignments, surprise quizzes (if any)
15% midterm
25% final
Homework, Late Policy, Teamwork etc.
Homework assignments will be posted on or before Mondays and are due the following Thursday at
11:59pm unless otherwise noted. No credit will be given for assignments that are more than 3 days late
(Sunday 11:59pm). You can hand in at most one assignment up to 3 days late and still receive full
credit. That is: one permitted late assignment for the entire quarter. Additional late assignment will be
given a max of 50% credit.
Unless otherwise noted, written assignment can be handed in on a sheet of paper (preferred) or by
email. If you email it to me, please make sure your name appears on top of the actual assignment (e.g.,
in the text file attachment), not just in the email. Programming assignment must be turned in using
Blackboard’s Digital Dropbox. See the respective assignment assignment for the naming
convention.
Collaborative work:
- Collaboration is encouraged on understanding assignments and formulating strategies for
solution.
- Internet, textbook, and other research is encouraged.
- However, the actual writing of solutions must be done independently, without access to
any material from colleagues, or other sources.
- Similarly, you may not allow anyone to copy your solution.
- Exams must be done independently.
- If a part of your solution includes a quote, it must be clearly indentifyable as such through
quotation marks and a reference.
- Keep in mind: plagiarism is a serious offense and ”Detection can result in serious
academic sanctions, ranging from a failing grade on a test or paper, through failure in a
course, to the denial of a graduate degree.” NPS’ official words can be found here:
http://intranet.ern.nps.edu/Faculty/documents/Plagiarism\%20Statement.doc
Reading Assignments
There will be weekly reading assignments from the reader and from other reading material. A “reading
assignment” requires you to read the material and to write a short, one or two page summary.
Demonstrate your understanding of the contents, don’t provide me with a complete recount of the
entire text. You are encouraged to add a paragraph in which you critique the reading, note questions
that you have or state some other related comment. Print the reports (duplex, please) and bring them to
class on the respective due day. If you cannot make it to class, please email them to me prior to
class.
Equipment
All special equipment and software will be provided to you:
High-performance laptop, Matlab with Image Processing Toolbox, USB or FireWire (1394) camera,
OpenCV
Programming Projects
As part of the course, we will work on small programming projects with the help of a high-level
computer vision library (OpenCV) to demonstrate some of the methods we have learned
in class. Some sample projects are listed below, further ideas and detail can be found in
Appendix ??:
- face detection in live video
- background modeling for object tracking
- detection and tracking of pedestrians
- extrinsic camera parameter estimation, for ego tracking and robot movement
- live interaction through marker tracking
Resources
Textbooks
Textbooks, more or less in order of relevance (res: on reserve in the library)
- course textbook: Linda G. Shapiro, George C. Stockman, Computer Vision (Prentice
Hall; 1st edition, 2001), ISBN: 0130307963
- res: D. Ballard and C. Brown: Computer Vision (Prentice Hall, 1982)
- res: D. A. Forsyth, J. Ponce; Computer Vision: A Modern Approach (Prentice Hall),
ISBN: 0130851981
Since I was asked: there are no color pictures in this book.
- R. Hartley and A. Zisserman: Multiple View Geometry second edition, Cambridge
University Proess. ISBN 0521540518
- Digital Image Processing (2nd Edition) by Rafael C. Gonzalez, Richard E. Woods
Publisher: Prentice Hall; 2nd edition (January 15, 2002) ISBN: 0201180758
- Digital Image Processing Using MATLAB by Rafael C. Gonzalez, Richard E. Woods,
Steven L. Eddins Hardcover: 782 pages Publisher: Prentice Hall; 1st edition (September
5, 2003) ISBN: 0130085197
- Robot Vision by Berthold K. P. Horn Hardcover: 480 pages Publisher: The MIT Press;
Mit Press edition (March 13, 1986) ISBN: 0262081598 co-published with McGrawHill
→ out of print
- Computer and Robot Vision by R. Haralick, and L. Shapiro Addison-Wesley, 1992
→ two-volume book is tons of material, good references
- A Guided Tour of Computer Vision by Vishvjit S. Nalwa Addison Wesley, 1993
→ great for concepts
- Machine Vision by R. Jain, R. Kasturi, and B. G. Schunck McGraw-Hill, 1995
- Machine Vision: Theory, Algorithms, Practicalities by E. R. Davies Academic Press,
1997
→ chapter 2 is at
http://www.palantir.swarthmore.edu/maxwell/papers/html/Maxwell-WUEIC-1997.htm
- Practical Computer Vision using C by J. R. Parker 1994
- J.E. W. Mayhew, J.P. Frisby; 3D Model Recognition From Stereoscopic Cues MIT Press, 1991,
ISBN 0-262-13243-5 online at
http://homepages.inf.ed.ac.uk/rbf/BOOKS/MAYHEW/
- M. A. Fischler, O. Firschein Intelligence: The Eye, the Brain and the Computer Addison-Wesley,
1987, ISBN: 0201120011 online at
http://homepages.inf.ed.ac.uk/rbf/BOOKS/FISCHLER/
- a collection of online (and offline) computer vision/image analysis/imaging books
at:
http://homepages.inf.ed.ac.uk/rbf/CVonline/books.htm
Miscellaneous
Contents