EECS 598-08
Foundations of
Computer Vision
Fall 2014


EECS 598-08: Foundations of Computer Vision

Instructor: Jason Corso (jjcorso)
Time/Place:MW 1430-1600 in 2150 DOW
Office Hours: TR 1430-1600 in 4227 EECS

Course Description and Outline

Computer Vision seeks to extract useful information from images. This course begins the fundamentals of image formation and then organizes the remaining material according to the class of information to be extracted. We will cover early processes, such as basic features, edges and contours; motion tracking, including optical flow and filtering; shape primarily from a binocular 3D reconstruction point of view; and both object and action detection and recognition. The course has been designed to present an introduction to computer vision targeted to graduate students. The course will balance theory and application both in lectures and assignments.

A coarse sequence of topics we will cover is below followed by a detailed week-to-week schedule that will be populated throughout the semester.

  1. Data Fundamentals: camera models, image formation, range sensing and video.
  2. Early Processes: extracting basic features, edges, contours, and segmentation.
  3. Motion Tracking: extracting movement, optical flow, tracking, and filtering.
  4. Shape: extracting 3D structure, epipolar geometry, stereo, SFM, shape from X.
  5. Objects: extracting objects, detection, recognition, and matching.
  6. Actions: extraction actions, space-time localization, and detection.

Information Flow

This courses uses CTools, Piazza and the instructor's website.

  • CTools will be used for submitting work and accessing grades (via the Dropbox plugin), among some minor other things that will be noted in due course. The CTools Link is
  • This website ( will primarily hold lecture notes and problem-set downloads.
  • Piazza is used for announcements and to manage student discussions. The piazza course website is (this link is also in the CTools site). Students should ensure they are enrolled in the course. Nearly all questions you have about the course, both logistical and technical should be posted to piazza (after you have already checked piazza to ensure the same question has not already been answered). Only in the event of a concern of privacy, should you directly email the instructor.

Materials Needed

You are required to have and work with one of these two textbooks (concurrent readings will be available in both and supplemented with additional material when necessary):
Forsyth and Ponce (FP) Computer Vision, A Principled Approach. Prentice Hall, 2nd Edition, 2011. (ISBN-13: 978-0136085928 ISBN-10: 013608592X)
Szeliski (SZ) Computer Vision: Algorithms and Applications published by Springer and available for purchase on various website or as a download (free) at

Additional concurrent readings may be assigned from materials available on the web. Or from the following books (all such reading are optional and are intended to further deepen your understanding of a particular topic):

Faugeras and Luong (FL) The Geometry of Multiple Images published by MIT Press in 2001.

A working installation of MATLAB with the image processing toolbox installed.


Row color denotes topic area: Data Fundamentals  Early Processes  Motion  Shape  Objects  Actions 
1W 9/3Introduction; Data Fundamentals [pdf] FP 1; SZ 1 
2M 9/8 No Class ECCV CSC Assigned
2W 9/10 No Class ECCV  
3M 9/15 Camera Models; Geometric Image Formation [pdf ] FP 1; SZ 1, 2.1 and 2.3  
3W 9/17 Camera Calibration [pdf ]; Catadioptric Systems [Geyer ppt]FP 1.3; SZ 6.3; FL 4.6; Corso Notes PS1 Posted with PS1 Data
4M 9/22 Radiometric Image Formation; Color; Light [pdf ] FP 2, 3; SZ 2.2, 2.3 CSC Selection Due (9/21/14@23:59)
4W 9/24 Linear Filters and Image Processing [pdf ] FP 4, 6.1, 6.4; SZ 3 
5M 9/29 Local Image Features [pdf ] FP 5; SZ 4.2, 4.3 
5W 10/1 Segmentation [pdf ] FP 6.2, 9; SZ 5.2-5.4 
6M 10/6 Clustering in Vision [pdf | EM-slides pdf ] FP 6.2, 9; SZ 5.2-5.4 PS2 Posted with PS2 Data
6W 10/8 Model-Fitting and Contours [pdf | EM-slides pdf ] FP 10; SZ 4.3, 5.1 PS1 Due (10/10/14@23:59)
7M 10/13No Class Fall Break 
7W 10/15 Motion [pdf ] FP 10.6; SZ 8; TV 8 
8M 10/20 Tracking [pdf ] FP 11; SZ 8 
8W 10/22 Epipolar Geometry [pdf ] FP 7, SZ 11.1, TV 7 
9M 10/27 Stereo Vision [pdf ] FP 7, SZ 11, TV 7Midterm Exam (Take-Home 80 Minutes)
9W 10/29 Phil Torr's Matlab SAM toolkit url  
10M 11/3 Euclidean SFM (board lecture no pptx)FP 8.1 
10W 11/5 Affine SFM [pdf ] FP 8.2 PS2 Due (11/7/14@23:59)
11M 11/10 Projective SFM [pdf ] FP 8.3  
11W 11/12 Visual Recognition [pdf ]   CSC Draft Due (11/12/14@23:59)
PS2 Due (11/14/14@16:59)
12M 11/17 AdaBoost and Face Detection [pdf ]   
12W 11/19 AdaBoost and Face Detection [pdf ]   
13M 11/24 Face Recognition [pdf ]   
13W 11/26 Face Recognition [pdf ]   
14M 12/1   
14W 12/3  CSC Due (12/3/14@23:59)
15M 12/8   
15W 12/10   


  • Problem Sets (45%) There will be three problem sets (one each due at the end of September, October and November). These will include both analytical problems and programming assignments (in Matlab). All data for programming assignments will be provided by the instructor. Problem sets may be discussed in groups but must be written independently, including programming. Over-the-shoulder MATLAB debugging, for example, is not permitted. No code from other students, on-line or off-line resources other than that explicitly mentioned in the assignment is permitted.
  • Comprehension Service Component (15%) Each student will select a unique entry in Wikipedia that is related to computer vision early in the semester. The student will revise and contribute to this Wikipedia entry (off-line) throughout the term. The revision and final entry will be due in beginning of December. Upon instructor approval, the student will then make the revisions to the entry in the on-line site.
  • Exams (40%) There will be an in-class mid-term exam (mid-to-late October) and a take-home final exam (December). The mid-term exam will not have programming on it; the final exam will.

last updated: Tue Sep 6 22:19:34 2016; copyright jcorso
Please report broken links to Prof. Corso .