CS 766 - Computer Vision
Fall 1995
Office: 6397 CSS
Telephone: 262-1965
Email: dyer@cs.wisc.edu
Office Hours: 2:30 - 3:30 Mondays and Thursdays, and by appointment
Teaching Assistant: Bryan So
Office: 5364 CSS
Telephone: 262-5105
Email: so@cs.wisc.edu
Office Hours: 1:00 - 2:00 Wednesdays and Fridays, and by appointment
General Course Information
Fundamentals of computer vision.
First, an introduction to low-level image analysis methods,
including image formation, edge detection, feature
detection, and segmentation.
Principles of defining modules for
reconstructing three-dimensional scene information using
techniques such as
shape from shading and depth from stereo.
Active methods for scene recovery such as depth from focus and
occluding contour detection by viewpoint control.
Motion detection and analysis including tracking.
Model-based three-dimensional object recognition.
- Schedule
1:00 - 2:15 p.m. Tuesdays and Thursdays in 1325 CSS
- Prerequisites
CS 540, fundamentals of calculus, probability
theory, linear algebra, and C
- Grading
- Midterm Exam (Thursday, November 16): 35%
- Homework assignments: 35%
- Project: 25%
- Class participation: 5%
- Syllabus
- Required Readings
- Selected parts of
Machine Vision
by R. Jain, R. Kasturi, and B. G. Schunck, McGraw-Hill,
New York, 1995
- A collection of readings
from journals and conference proceedings;
to be sold at DOIT Documentation in small batches
- Supplementary Reading Sources
- Online Information
Most course information will be available online at WWW URL
http://www.cs.wisc.edu/~dyer/cs766.html
Reading Assignments to Date
- 9/5: Chapters 1 and 8 of JKS, and Paper #1 (at DOIT)
- 9/14: Chapter 2
- 9/21: Chapter 3
- 9/28: Chapters 4 and 5
- 10/5: 4 papers at DOIT available as Handout #3
- 10/17: Chapter 6 except 6.6, 6.7, 6.8.3, and 6.9
- 10/19: Chapter 7.1 - 7.3, and 2 papers at DOIT available as Handout #4
- 11/2: 3 papers at DOIT available as Handout #5, and Chapters 9, 11.1 - 11.3
- 11/28: Chapter 14 (primarily 14.4)
Homework Assignments
- Homework #0: Image Enhancement by Histogram Modification (Optional)
Make a copy of your portrait image in ~cs766-1/public/images/
and then use xv to contrast enhance your face. Do this by
first rotating the image, then cropping a window around your head
(say down to your shoulders), and finally interactively adjusting the
Intensity modification function in the Color Editor window under the
Windows button. (You are also free to modify other things such as color
if you wish.) When you have found a good grayscale transformation
save the result as a color gif image and
put it in the same directory where
the original image is. Send me email telling me qualitatively what
intensity transformation you applied and why it improves the quality
of the image overall. I'll then use this image in the "photo board"
of students in the class. Feel free to use this image in your own
Web home page as well!
- Homework #1: Skeletons (Due: October 3)
- Homework #2: Image Splining for Mosaics (Due: October 24)
- Read the 3 papers:
- E. H. Adelson et al., Pyramid methods in image processing,
RCA Engineer 29-6, Nov/Dec 1984
- P. Burt and E. H. Adelson, The Laplacian pyramid as a compact
image code, IEEE Trans. Comm. 31, 1983, 532-540
- P. Burt and E. H. Adelson, A multiresolution spline
with application to image mosaics, ACM Trans. Graphics
2, 1983, 217-236
- Hints and FAQs
- Some splined images
produced by students in the class
- Homework #3: Segmentation by Snakes (Due: November 9)
- Read the 2 papers:
- M. Kass, A. Witkin and D. Terzopoulos,
Snakes: Active contour models,
Int. J. Computer Vision 1,
1988, 321-331
- D. Williams and M. Shah, A fast algorithm for active contours
and curvature estimation, Computer Vision, Graphics, and Image
Processing: Image Understanding 55, 1992, 14-26
- Hints and FAQs
- Homework #4: Your Own Project (Due: December 18)
Computer Accounts
- Accounts
Course accounts are on the Sun Sparcstations called sun1 - sun36
in rooms 1358, 1363 and 1368. Each account has a large disk space quota
of 50MB so you can store images for homeworks and
your project. Be sure to delete old images and compress others
(see gzip(1)), however,
in order to save space.
- Email
Email sent to cs766-1list
goes to everyone in the class including the instructor and TA
- Printers
To print images you should use one of the laserprinters,
laser1 - laser4, which are located
in room 1359. Alternatively, the generic printer name laser will
send output to one of the four printers with the shortest queue. Caution:
Before sending images to the printer, be sure to check the queue; if
there are a lot of jobs being printed it is bad manners to send images
to be printed because they take so long to print. Be considerate!
- Vision Software
- Vista
The Vista
programming environment will be used in the homework assignments.
The code is located in the directory /p/vision/ip-tools/vista/
Man pages are in /p/vision/ip-tools/vista/man/
and executables are in /p/vision/ip-tools/vista/bin/
- Xv
xv(1) is an interactive image display program for the X
window system that is very useful for displaying images in a
variety of formats.
- ImgStar
70 basic image processing operations invoked using Unix-like command lines.
Code, executables and manual are in /p/vision/ip-tools/imgstar/
- Khoros
The Khoros image processing software development environment
provides a set of basic image processing modules and a graphical
programming language interface for rapid prototyping of simple
image processing algorithms. The code is located in the directory
/p/vision/ip-tools/khoros /p/vision/ip-tools/khoros/bin/cantata
is the executable that starts up the interactive environment.
- Netpbm
A toolkit for conversion of images between a large variety of
different formats. Based on the Pbmplus package. Man pages are in
/p/vision/ip-tools/man/ and executables are in
/p/vision/ip-tools/bin/
- Matlab
Matlab(1) is a numeric computation and visualization
environment. Signal processing
and image processing toolboxes are especially relevant.
- Test Images
Most test images will be put in the directory
/p/vision/images/ although they may require format
conversion to be used. Some other images may be put in
~cs766-1/public/images/ Numerous image databases are also
accessible via the WWW; for example, see the
collection of test images at CMU
Examination
- Examination
- Solution
- The Exam will be held on Thursday, November 16 from 12:45 pm to 2:15 pm in the regular classroom, 1325 CS. Note the early
starting time! The exam will cover topics up through shape-from-shading,
including readings in the textbook, papers sold through DOIT, and
homework assignments. You may bring into the exam one (1) 8.5" x 11"
sheet of paper with any notes you want on both sides. The exam will
focus on main ideas and algorithms, not proofs. See old exams below for
the types of questions that will be asked.
- Old Exams
Links of Interest
Vision Demos and Projects
Other Computer Vision and Related Courses