EECS4422

This course introduces the fundamental concepts of computer vision, with a balance of theory and practical application.

Specific topics:

• Introduction
• Image formation
• Image processing
• Feature detection & matching
• Segmentation
• Dense motion estimation
• Feature-based alignment
• 3D - motion
• 3D - stereo
• 3D - single view
• Recognition

General prerequisite; LE/EECS 2030 3.00 or LE/EECS 1030 3.00; SC/MATH 1025 3.00; SC MATH 1310 3.00, LE/EECS 2031 3.00. (NOTE: The General Prerequisite is a cumulative GPA of 4.50 or better over all major EECS courses. EECS courses with the second digit “5” are not major courses.)

It also is recommended that students enter this course with a good working knowledge of the calculus of several variables and linear algebra. Familiarity with linear systems theory (e.g., EECS 3451, formerly COSC 4451 and CSE 3451), comfort with elementary manipulation of complex variables and previous experience equivalent to a university level introduction to physics course also would be of value. If in doubt, then consult with the instructor.

• James Elder
  • Office: LAS 003G
  • Office hours: W 11:30AM-12:30PM
  • Email: jelder@yorku.ca

• Hakki Karaimer
  • Office: LAS 2052
  • Office hours:
  • Email: hakkicankaraimer@gmail.com

The required textbook for this course is

Computer Vision Algorithms and Applications by Richard Szeliski, Springer, 2011.

Errata for the textbook is available here.

This text is available at the York University Bookstore in York Lanes. Also, a copy is on reserve at the Steacie Library on campus.

• Lectures: Students are responsible for all material covered in lectures. Lecture slides will be posted incrementally on this website; see the Schedule link on the side bar to this page.
• Assigned Readings: Students are responsible for all material assigned as reading in the textbook; see the Schedule link on the side bar to this page.
• Labs: There will be six two-hour labs, these will be held at 4pm-6pm on selected Mondays; see the Schedule link on the side bar to this page. The TA will be present at each lab to provide demonstrations and guidance. Students are responsible for all material presented in labs. Labs 1, 2, 4 and 5 will be associated with the four short assignments (see below). Lab 3 will be used to help students prepare for the midterm. The final lab will be used to demo projects and to discuss project reports.
• Assignments: There will be four short assignments that include theory and coding questions; details will be posted incrementally on this website; see the Schedule link on the side bar to this page.
• Midterm: In class, closed book.
• Project: Students will complete a project involving the implementation of a computer vision algorithm of their choosing. Details can found under the Project link in the side bar.

• Explain the basic terms, concepts and applications of computer vision, including reference to at least one real-world system.
• Apply basic mathematical techniques to solve problems in computer vision.
• Develop software to solve problems in computer vision.
• Analyze the effects of noise in computer vision algorithms and use appropriate techniques to reduce its effects

The weight distribution of the course components is as follows:

• 30% - 5 assignments, 6% each.
• 30% - Midterm exam
• 40% - Project
  • 10% - Proposal
  • 10% - Presentation/Demo
  • 20% - Final Report

Each piece of work will be assigned a numeric grade. A final numeric grade will be computed using the weighting given above. The final letter grade will be determined from the numeric score using the standard Computer Science and Engineering mapping.