Paul Fieguth Dept. of Systems Design Engineering Faculty of Engineering University of Waterloo Waterloo, Ontario Canada N2L 3G1 |
pfieguth@uwaterloo.ca Tel: (519) 888-4567 x84970 FAX: (519) 746-4791 |

Your handout should be at a high level, be understandable to other students in the class. Ideally, you should include a few references for students who might want to learn more on the topic.

I have here a couple of example handouts, just to give you a rough sense of
what others have done. However your handout does *not* need to look like
one of these:

Handout #1: side 1, side 2.
Handout #2: side 1, side 2.
Handout #3: only one side

Your project needs to represent your own work, cannot be copied from previous work (such as a thesis proposal or comprehensive exam), must be properly cited, and cannot have any material copied (plagiarism). Copying even single sentences or phrases is unacceptable.

You can submit a hardcopy in my mail slot or under my door (DC2643).
You can also email me an electronic copy, but which MUST be in PDF format
(no Word files!). So that I can sort and find your email easily, your email
must have "675" in the subject line,

e.g., "Subject: SD675 Project"

P. Fieguth: Office Hours, Mondays 12-1pm, DC-2643

This course is an advanced version of SY DE 372 - Introduction to Pattern Recognition. For those students lacking a background course in pattern recognition, it is recommended that you attend some or all of the SD372 lectures. The material discussed in SD675 will overlap only relatively little with that from SD372.

The course starts with a brief summary of SY DE 372: probabalistic classifiers, discriminant functions, unlabeled clustering, and feature extraction.

More advanced topics will include some information theory (as it pertains to feature extraction), statistical estimation and error analysis (relating to parameter estimation), neural networks, self-organizing maps, and syntactical/grammatical pattern recognition.

The course grade is based on a few computer labs, one or two problem sets, and a term project.

- Assignment 1 (Assigned January 14th, Due January 21st)
- Assignment 2 (Assigned January 21st, Due January 28th)

The data set assign2.mat for this assignment is located here Right-click to download, save to disk, and open in Matlab.

- Assignment 3 (Assigned January 28, Due February 4 or February 9)

To keep the results consistent from student to student, I am giving you the data points for this assignment instead of you generating them yourselves. The data set is assign3.mat; if, for some reason, you are using a very old version of Matlab, then you might try assign3_v4.mat. Right-click to download, save to disk, and open in Matlab.

Students may wish to numerically evaluate expressions involving Q(). The Matlab routine for Q() is here. - Assignment 4 (Assigned Feb 4, but only posted on-line on Feb 6, due by Feb 13th)

This lab assignment is on Boosting. The data set is the same one as Case 4 from the previous two labs, and is available here. - Assignment 5 (Assigned Feb 13, due in late Feb)

This assignment has 5 short questions, no big programming, a couple of small analytical questions. No data set needed for this lab.. - Assignment 6 (Assigned Mar 9, due in late Mar) This is a CRB assignment. The Matlab is really easy, just two simple plots. There are two CRB derivations; neither is hard, but the second is a bit tricky to sort out what to do.

- Overall maximum of one page
- Give a title
- Give a one to three paragraph summary of the *problem*. Focus on the problem, what makes it interesting, or hard. Don't focus on how you might want to solve the problem.
- Make it clear what you propose to do with the problem (write a survey, implement an algorithm, testing on what sort of data etc.).
- Give a few references (at least two, no more than five)

- Very short - aim for 3-5 minutes
- No powerpoint (no time for that), just one or two overheads, or you can do your presentation on the blackboard
- Do NOT discuss your *solution*, focus on the *problem*. How does the problem connect to the course, makes it hard, why is it interesting?
- Prepare a one-page (one or two sided) handout for the class, sort of a brief conference-paper describing the project. It does not need to be detailed, just a short description, maybe some figures, plus a few references.

- Ideas for selecting a project topic.
- Information concerning project grading.
- Advice on how to write technical reports.

Home | Teaching | Research | Students | Software | Papers |

(Page last updated September 13, 2016, [an error occurred while processing this directive])