CS691A: Graduate Computer Vision

• Spring 2010; Tuesday and Thursday, 11:15-12:30 • in ELAB 323


Erik Learned-Miller
elm at cs.umass.edu
(413) 545-2993

Teaching Assistant: None
Syllabus: syllabus.pdf


Reading Materials

Lightness Perception and Lightness Illusions

Unsupervised learning in vision


MATLAB Tutorial

Matlab Diary for Lecture on Jan. 19

Interesting Links

The Necker Cube illusion

Checker shadow illusion

Movie on optical illusions

Early color photographs by S. M. Prokudin-Gorsky


Date Lecture topic New assignments Assignments due Reading
Jan. 19 Introduction. What is Computer Vision? What are the goals of computer vision? What should we take from humans? Also, Intro to Matlab. See Matlab Diary of in class session under "Resources" above.

Read A Review of Basic Probability by Tuesday, January 26, at class time.
Jan. 21 Review of basic probability. Samples spaces. Events. Joint Probability. Conditional Probability. Marginalization. Role of Probability and Statistics in Computer Vision.

Jan. 26 Bayes rule. Likelihoods, priors, and posteriors. Estimating likelihoods, priors, and posteriors.

Read Supervised Learning and Bayesian Classification by Tuesday, February 2, at class time.
Jan. 28 A few more Matlab tricks: dot-m files and Matlab calling conventions, the image toolbox, avoiding for loops, repmat, dot-times, dot-slash, dot-power, etc.

What features to use in classification? Modeling the entire joint distribution of images in a class. Assuming feature independence. More on estimation and smoothing.

Assignment 1: Probabilistic classification.

Digit data for assignment 1

Due by class time on Feb. 4. Please email me the solution as a zipped tar file of all necessary files.

Feb. 2 Euclidean distance functions in 1, 2, 3, and more dimensions. Non-parametric density estimators.

Feb. 4 The statistics of changing coordinates. Translation, rotation, and other image movements as changing coordinates. Nearest neighbor classification. K-Nearest neighbor classification. Consistency of K-nearest neighbors.

Feb. 9 Prokudin-Gorski photographs. Correlation alignment. Mutual information alignment. Other schemes of aligment. Problems of alignment and solutions to alignment problems. Assignment 2: Automatic alignment of Prokudin-Gorsky plates.


Due by 11:15 on Feb. 16.

Read Entropy and mutual information by Thursday, Feb. 11 at class time.
Feb. 11 Alignment continued.


Feb. 18 Types of alignment: exhaustive search, gradient descent, coordinate descent, gradient descent with restarts. Issues in alignment. Local minima. The zero-gradient problem. Computational complexity. Criteria for pairwise alignment: mutual information, correlation, norms (L2, L1, L-infinity). Introduction to joint alignment.

Feb. 23 Joint alignment and congealing. The minimum entropy criterion. Non-parametric maximum likelihood. Joint gradient descent (or joint coordinate descent). Smoothing of the optimization landscape without destroying information. Next assignment: Congealing implementation, due on March 9th, before class.

Lecture slides on congealing
Feb. 28

Mar. 2 Start light and optics. Electromagnetic spectrum. Pinhole cameras. Thin lenses.

Mar. 4 Point sources, extended sources. Radiance, irradiance, luminance, illuminance, brightness.

Handout on radiometry.
Mar. 9

Mar. 11 Learning to classify textures. Written problem set. Due Mar. 25



Mar. 23 Guest Lecture: Vidit Jain


Mar. 30 Motion in images. Part 1. Backgrounding.

Apr. 1 More on backgrounding. Modeling the background, the foreground, and the prior.

Apr. 6 Finish backgrounding. Start optical flow.

Apr. 8 Finish optical flow.

Final problem set. Due Apr. 20 train_data.mat
test_data.mat gardenImages.mat