CS 787 - Computational Vision

Winter 2007
Department of Computer Science
University of Waterloo

Instructor: Richard Mann, DC2510, x3006, mannr@uwaterloo.ca

Lectures: Mondays *** NEW TIME 3:30-6:00, Room DC3313 ***
First class: Monday 8 January

Important Note: Please try to send all course material from a University of Waterloo account (ie., SSH into a teaching or research machine and send from there). If you don't have a Waterloo account yet, please contact the MFCF help center (MC3011). In the meantime you can use the account you gave in the first class. Also please don't send any messages with "zip" or "gif" attachments (send "tif" or "jpg" instead, or refer me to a URL). Those are common filetypes in Microsoft viruses and are removed by my SPAM filter.

News:

Office hour time: Tuesdays, 4:30-5:30pm in DC2510. Talk about lecture material, assignments, or project topics.

Lecture #2 *postponed* due to "snow day" (yeah, right!). Next lecture will be: Monday 22 Jan, 3:30-6:00pm, DC3313. Note *new* time.

Assignment #4 (stereo) posted. See link below.

Summary

Computational vision is concerned with the automatic processing of image and video data for scene reconstruction, object recognition, navigation, and activity detection.

The objective of this course is to provide a concise treatment of some fundamental problems in computational vision. This course will focus on a core set of problems where efficient and robust algorithms can be applied. Students will be required to implement several algorithms using real datasets. In addition to providing practical approaches and algorithms, this course will provide the foundation required to pursue research computational vision.

Prerequisites

There are no formal prerequisites for this course, however, it is advisable to have some exposure to numerical computation, especially linear algebra (eg., CS370), and some basic programming experience. Programming will be done in Matlab.
Note: this course is open to undergraduates, with permission of the instructor.

Marking (tentative)

Assignments: 50%

Four assignments. Handwritten + small programming component.

Final project: 50%

Implementation or literature review.

References

All required material will be provided in lectures. Possible references include:

Introductory Techniques for 3D Computer Vision, E. Trucco and A. Verri, Prentice-Hall, 1998.

A concise treatment, focussing on geometric approaches to vision.
The course closely follows this book.
Computer Vision, A modern approach, D. A. Forsyth and J. Ponce, Prentice Hall, 2003.

Comprehensive treatment of most areas of vision. A good resource
for background reading and choosing project ideas.

Robot vision, B. K. P. Horn, MIT Press, 1986.

Vision, D. Marr. Freeman, 198x.

Three-Dimensional Computer Vision, O. Faugeras, MIT Press, 1993.

Digital Image Processing, K. R. Castleman, Prentice Hall, 1996.

I will put all of these books on reserve in the library. You may also purchase Trucco and Verri's book in the University bookstore.

Lecture Schedule (preliminary)

Lecture	Topic	References	Assignments
1. (Jan 8)	Vision Overview; Image formation	Horn Ch2, Trucco and Verri Ch2 Related link: - Combining images of varying exposure (S. Mann, Toronto)
(Jan 15)	no class "snow day".
2. (Jan 22)	Lighting and reflectance models Application: Photometric stereo	Horn Ch10 Related links: - Lightness and shading illusions (E. Adelson, MIT) - Shape from shading (M. Langer, McGill)	A1 out
3. (Jan 29)	Filtering: Linear systems, Fourier theory Note: did not cover discrete case (finite size images, sampling). This will be covered in Lecture #4.	Horn Ch6, Castleman Ch 10
4. (Feb 5)	Edge and Feature detection. Application: good features to track (corners)	Trucco and Verri Ch4	A1 due; A2 out
5. (Feb 12)	Parameterized features and model fitting. Application: fitting lines to data using robust and mixture models.
(Feb 19)	no class, reading week
6. (Feb 26)	Optical flow Application: mixture models for optical flow	Trucco and Verri Ch8, Horn Ch12 References: - Mixture models for optical flow (A. Jepson, Toronto) - Optical Snow	A2 due; A3 out
7. (Mar 5)	Structure from motion	Trucco and Verri Ch8
8. (Mar 12)	Stereo (Baseline case)	Trucco and Verri Ch7; Faugeras Sect 6.6	A3 due; A4 out
9. (Mar 19)	Stereo (Epipolar geometry) Application: 3d scene reconstruction	Trucco and Verri Ch7; Faugeras Ch7
10. (Mar 26)	Object Recognition: View-based approaches. Application: PCA for object recognition	Trucco and Verri Ch9,	A4 due
11. (Apr 2)	Object Recognition: Model-based approaches. Application: Object search	Trucco and Verri Ch9
xx. (extra?)	Computational perceivers. Applications: Scene labelling, perception of scene dynamics

Assignments

Assignment #1: Image formation and Lighting.

Assignment text (PDF), complete package (tar, zipped) Due: noon, Fri 9 Feb, 2007.

Notes:

To do the assignments you will need access to Matlab. This should be available in the Undergrad teaching environment. Information is available at:

http://www.cs.uwaterloo.ca/cscf/general/

http://www.math.uwaterloo.ca/mfcf/undergrad/Undergrad_Servers.html

All students registered for the course, grad and undergrad, should have accounts here. If you don't have course accounts please visit MC3011 (bring your "Watcard").

You will need to install the following course software (tar, zipped) in your home directory.

Oops! You need the following additional files (put in cs787-w07-software/matlab/ directory): pgmRead.m, pgmReadHeader.m, pgmWrite.m

Assignment #2: Linear systems and Feature detection

Assignment text (PDF), complete package (tar, zipped) Due: 3:30pm, Mon 26 Feb, 2007.

Assignment #3: Robust and mixture models for optical flow

Assignment text (PDF), complete package (tar, zipped) Due: 3:30pm, Mon 12 Mar, 2007.

Assignment #4: Stereo vision: Block matching and dynamic programming

Assignment text (PDF), complete package (tar, zipped). Due: noon, Fri 30 Mar, 2007.