CPS843/CP8307
Introduction to Computer Vision
Winter 2017

with Prof. Kosta Derpanis


Course description. This course introduces the fundamental concepts of vision with emphasis on computer science and engineering.  In particular, the course covers the image formation process, image representation, feature extraction, stereopsis, motion analysis, 3D parameter estimation and applications.

Contact information. kosta[at]scs.ryerson[dot]ca

Office. Engineering and Computing Building (ENG), room 283

Homepage. www.scs.ryerson.ca/~kosta

Office hours. Thursdays at 2pm

Prerequisites. This course requires programming experience (CPS305 OR equivalent) as well as basic linear algebra (MTH108 OR equivalent) and calculus (MTH310 OR equivalent).

Course textbook. Richard Szeliski, Computer Vision: Algorithms and Applications (available for free or purchase)

Lectures. Below is the tentative schedule of topics.  Links to slides will be made available after each lecture.

# DATE TOPIC SLIDES
 quicktime
BACKGROUND MATERIAL
ASSIGNMENTS
0
1/13
Introduction to computer vision PDF MOV
Szeliski - Chapter 1

Eero Simoncelli, A Geometric Review of Linear Algebra
A0 released
IMAGE FORMATION
1
1/13 Camera optics (part 1) PDF MOV
Szeliski - Chapter 2 (Sec. 2.1)

History of Photography in 5 Minutes (video)

2
1/20
Camera optics (part 2) PDF MOV

IMAGE FILTERING AND FEATURE EXTRACTION
3
1/27
Image filtering (smoothing) PDF MOV
Szeliski - Chapter 3 (Secs. 3.1 to 3.2.2) A1 released
4
2/3
Image filtering (edge detection) PDF MOV Szeliski - Chapter 4 (Sec. 4.2)

Pedro Felzenszwalb, Edge Detection

Avidan and Shamir, Seam Carving for Content-Aware Image Resizing

5
2/10 Image features PDF MOV Szeliski - Chapter 4 (Sec. 4.1)
MODEL FITTING
6
2/17
Model fitting PDF MOV
Szeliski - Chapters 4 (Sec. 4.3.2) and 6 (Secs. 6.1.1, 6.1.2 and 6.1.4)

Ballard and Brown, Hough transform
A2 released
FREQUENCY ANALYSIS
7 3/3
Frequency analysis (part 1) PDF MOV Szeliski - Chapter 3  (Sec. 3.4)

Horn - Chapters 6 and 7
(BrightSpace: Course Readings)

8 3/10
Frequency analysis (part 2) PDF MOV
RECOGNITION
9 3/17
Machine learning crash course
PDF MOV
A3 released
MULTIPLE IMAGE ANALYSIS
10
3/24
Stereopsis PDF MOV Trucco and Verri - Chapter 7
(Brightspace: Course Readings)

11 3/31
Motion analysis PDF MOV Trucco and Verri - Chapter 8 (Secs. 8.3 and 8.4)
(Brightspace: Course Readings)

Fleet and Weiss, Optical Flow Estimation
(Secs. 1 and 2)
A4 released
12 4/7
3D structure and motion PDF MOV Trucco and Verri - Chapter 8 (Secs. 8.1 and 8.2)
(Brightspace: Course Readings)

Online viewing. Note that some videos contain audio.  To hear the audio portion of the MOV files you may have to download the video to your local machine; audio support seems to be problematic in Google Drive for Education.

Reference textbooks.
     Gilbert Strang, Linear Algebra and Its Applications (video lectures)
     Berthold Horn, Robot Vision, MIT Press.
     Emanuele Trucco and Alessandro Verri, Introductory Techniques for 3-D Computer Vision, Prentice Hall.

Evaluation. Graduate students (CP8307) will be expected to do additional work on each assignment and test.  Undergraduates (CPS843) are encouraged to attempt these question for extra credit.
     Assignment 0 (MATLAB warm-up): 5%
     Assignment 1-4: 10% each (CPS843 work in groups of at most two and CP8307 work alone)
     Midterm: 25%
     Final: 30%

Academic misconduct. Committing academic misconduct, such as plagiarism and cheating, will trigger academic penalties including failing grades, suspension and possibly expulsion from the University.  Ryerson students are responsible for familiarizing themselves with the Student Code of Academic Conduct

Useful links.
     MATLAB install for Ryerson students
    
MATLAB tutorial
    
MATLAB primer
     Basic Linear algebra review
     
Linear Algebra review and MATLAB tutorial

     Linear algebra review and reference

Acknowledgements. While a great effort has been made to assemble an original set of lecture slides, the essence of the presentation of many of the slides rely significantly on slides prepared by the following instructors: Richard Wildes, Kostas Daniilidis, James Hays, Derek Hoiem, Aaron Bobick, David Lowe, Kristen Grauman, Robert Collins, Svetlana Lazebnik, Steve Seitz, William Freeman, Robert Pless, Andrej Karpathy and Alyosha Efros.