CENG 421/ELEC 536: Computer Vision, spring 2011
Instructor: Dr.
Alexandra Branzan Albu email: aalbu at uvic dot ca
Lectures: Tuesday,
Wednesday, Friday, 9:30-10:20 am CLE 111
Office hours
(Instructor): Wednesdays, Fridays 10:30 am-12:00 noon in ECS 631.
READING BREAK
OFFICE HOURS
- Friday
Feb 25 9:30 am – 12:00 noon and 1:30 pm – 3:30 pm
- or send email for appointment
Textbook:
M. Sonka, V. Hlavac, and R. Boyle,
“Image processing, Analysis, and Machine Vision”, 3rd edition,
Thomson 2008.
Assignments
Assignment
1 posted. Image data
for assignment 1.
Project
The educational
goals of this project are:
- to expose you to the state-of-the
art in computer vision research;
- to implement one or more
published algorithms. This requires the understanding of the theoretical
details of the algorithms, as well as a certain degree of creativity in
re-engineering the method from the scientific article.
- to give you experience expressing
your self using computer vision concepts.
- by doing projects on different
topics and presenting them, we hope that you get a broad exposure on a variety
of computer vision applications
- examples of applications: medical
imaging, ocean engineering, surveillance, activity recognition,
environmental monitoring.
Projects in CENG 421 and in ELEC 536 are different in
complexity. For more details, see the project-related slides in Lecture 1.
Project
deliverables:
-
Project
description (3%) : one-two
slides outlining your project goals, image databases, algorithms that you
intend to use, and how you intend to validate your results.
-
Project
progress report (2%) : one page describing
the work that you have accomplished
to date, as well as the organization of future work with milestones.
-
Project
report (15%):
o
Content:
your report must contain the following:
1. introduction should contain: general
problem statement and related work (you must briefly discuss and reference the
papers that you have used for your project)
2. Approach:
this is the core section of your report; make sure that the approach is
described with a proper mathematical formalism and/or algorithmic flowcharts
3. Experimental
results and validation: it is critical to evaluate the performance of your
approach with quantitative measures; choose your evaluation protocol according
to the objectives of the project; this section must also contain a detailed
description of your experimental database; in case you have designed your own
database, briefly explain how you designed it.
4. Conclusions:
briefly describe what you have learned from this project and ways to further
extend and improve your work.
5. References:
make sure you use a proper format for your refs
o
Page limit:
4 (extra pages for appendixes are allowed). Exceptions can be made for
well-justified requests of longer page limits. Such requests will be examined
on a case-by-case basis.
o
Format:
according to IEEE
guidelines (word only). Complete IEEE guidelines can be found at
http://www.ieee.org/web/publications/authors/transjnl/index.html
o
Each team needs to submit one report
only. The contribution of each team member to the project needs to be clearly
specified at the end of the report.
-
CD
with database and code (10%): same deadline as for
presentation. Make sure that code is functional, easy-to-use, and well-commented.
-
Oral
presentation (10%): presentations will take place in the
last week of classes. The oral presentations will be marked according to this set
of criteria.
Syllabus (tentative)
|
Week |
Lecture notes |
|
1. Jan 5-9 |
L2.
Image Basics-1. Perspective Geometry. Sampling and Quantization. Required Reading: Textbook 2.1, 2.2 |
|
2. Jan 10-16 |
Tuesday Jan 11- no lecture L3.
Image basics-2. Sensing Illuminated objects. L4.
Example of a simple vision-based system (powerpoint
to be used as a guide; does not replace paper) |
|
3. Jan 17-23 |
Image preprocessing |
|
4. Jan 24-30 |
L9-10.
Edge and corner detection. Reading (for ELEC 536 only) |
|
5. Jan 31- Feb 6 |
L11. Guest lecture (project-related) Maia Hoeberechts and Marjolaine Matabos
(NEPTUNE
Canada) Challenges
in image and video analysis for undersea cameras Neptune presentation and examples are zipped here L12. Overview of past ELEC 536/CENG 421
projects L13.
Image Segmentation. Edge-based
|
|
6. Feb 7-13 |
L14.
Image segmentation. Thresholding L15, L16. Short presentations of
project proposals |
|
7. Feb 14-20 |
L17.
Segmentation. Region-based. L18.
Segmentation. Matching and evaluation. - Reading: Adelson and Bergen, Image pyramids Practice
midterm posted. Solution
posted |
|
8. Feb 21-27 |
Reading Break |
|
9. Feb 28 – March 6 |
Midterm (March 1, in-class) L20-21. Binary Shape analysis. Mathematical morphology. Reading: Watershed
Transform (excerpt from Gonzales and Woods) |
|
10. March 7-13 |
Motion Analysis Reading on motion
representation (mandatory for ELEC 536) Reading on texture synthesis
(mandatory for ELEC 536) L24.
Pattern recognition. Introduction Reading on minimum-distance
classifier |
|
11. March 14-20 |
Project meetings (Tuesday and
Wednesday) L25.
Pattern recognition. Bayesian classifiers Reading on Bayesian
classifiers |
|
12. March 21-27 |
Performance evaluation of Computer
Vision Algorithms Midterm 2 (Friday March 25) |
|
13. March 28-April 1 |
Project presentations |
