Jump to Tutorials, Lab Assignments and Class Schedule

David Knill
phone: 5-4597
office: 275 Meliora Hall
e-mail: knill@cvs.rochester.edu
office hours: Tuesday, 1:00 – 2:00

Lectures: Tue, Thu 9:40 – 10:55 AM
Lab: Wed 3:25 – 5:05 PM

Clarc 108A

Bruce, V., Green, P. R. and Georgeson, M. A. Visual Perception.
Supplementary readings (handed out in class)

Matlab Primer: http://www.fi.uib.no/Fysisk/Teori/KURS/WRK/mat/singlemat.html

http://www.bcs.rochester.edu/courses/222.index.html

Description:

The advent of computer technology brought with it a revolution in the study of biological vision. It provided the means to simulate and test complex models of how the brain "solves" visual problems (e.g. computing depth). This course will provide hands-on learning of computational theories of biological vision systems. It will cover a broad range of material, from how cells in the retina process images, to how neurons in visual cortex perform complex computations like finding object boundaries, estimating depth and recognizing patterns. The course will include two 1-1/4 hour lectures and a 2-1/2 hour lab session each week . In the lab, students will learn central concepts by working through computer tutorials, running experiments, and developing and simulating state-of-the-art models of visual processing.

Course requirements:

1. Lab Tutorials: Students must complete a canned computer tutorial each week. Included in each tutorial will be a series of questions which will require students to explore the behavior of mathematical functions, explore the behavior of canned computational models, and even perform experiments. For each tutorial, students must hand in a record of the Matlab session in which they worked through the tutorial as well as written answers to directed questions. Tutorials are handed out in class on Tuesday and are due the following Tuesday (prior to class) (20% of grade)
2. Lab assignments: Students will complete 4 laboratory assignments in the class. These assignments will require implementing a series of computational models in Matlab. Students must hand in the Matlab code (commented) which I will be able to run to demonstrate the model. Lab are assigned on Thursdays and are due 1-1/2 weeks later on a Tuesday. (20% of grade)
3. Final Project: Students will design and implement a final project in Matlab related to the material covered in the class. Your final project can be an extension of computational models implemented in the class, an implementation of some other computational model, or an implementation of a psychophysical experiment. Students will turn in all Matlab code (commented) associated with their project and instructions for demonstrating it. Along with the code, students will turn in a ten-page, double-spaced paper describing their project and the relevant background information from the literature. Students must also meet individually with the instructor to demonstrate their project. (20% of grade)
4. Exam: There will be a cumulative final exam for the course. (20% of grade)
5. Class participation and daily questions: Since the course builds cumulatively on the material presented in class, it is essential that you attend and participate in the Tuesday, Thursday classes and that you read the assigned material prior to class. Attendance in class is mandatory, with the exception of illness and family emergencie. You should also submit two questions on the readings assigned for each class by midnight the night prior to the class. These amount to, on average, one set of questions per week, with 11 total during the term. You have one free pass on questions, meaning that you are required to turn in 10 sets of questions during the term. (20% of final grade)

Grading

1. Tutorials – Tutorials must be turned in by the due date. They are handed out in class on Tuesday and are due the following Tuesday (prior to class). LATE ASSIGNMENTS WILL NOT BE ACCEPTED. It is critical that students keep up with the class material. Getting behind on the tutorial assignments will seriously impact your learning experience.
2. Lab assignments - Labs are assigned on Thursdays and are due 1-1/2 weeks later, in class on a Tuesday. The equivalent of one letter grade will be subtracted for each class period that a lab is late (1 point out of 10).
3. Final project – The first part of the final project (Matlab code and demonstration) will count for 70% of the project grade. The paper will count for 30% of the grade.
4. Class participation – Class participation counts for 20% of your final grade. 2 points out of the 20 will be subtracted for each unexcused absence. 1 point ot of the 20 will be subtracted for each class in which you do not submit questions on the assigned reading (applies only to classes for which new reading is assigned), with the exception of the one free pass that you have for the term.

Important dates

Course schedule

Note: Lab assignments and tutorials are due one week following the day they are assigned (unless otherwise noted in class).

Week 1 (1/13 – 1/17)

Thursday: Introduction and introduction to computational vision

Week 2 (1/20 – 1/24)

Tuesday: Matlab / Linear algebra

Tutorial 1: introduction to Matlab / linear algebra

Thursday: Image formation (B&G, chapter 1)

Week 3 (1/27 – 1/30)

Tuesday: Image formation, cont. (Supplemental reading)

Tutorial 2: Linear systems analysis

Thursday: Fourier analysis (Supplemental reading)

Lab 1: Convolution, modeling blur and astigmatism (Due: Tuesday, 2/11)

Week 4 (2/3 – 2/7)

Tuesday: Fourier analysis

Tutorial 3: Fourier analysis

Thursday: Fourier analysis 2

Week 5 (2/10 – 2/14)

Tuesday: Retinal image coding (B&G, Chapter 2)

Tutorial 4: Modeling retinal ganglion cells

Thursday: Work on tutorial

Lab 2: Modeling brightness illusions (Due: Tuesday, 3/4)

Week 6 (2/17 – 2/21)

Tuesday: Off day

Thursday: Visual cortex (B&G, Chapter 3)

Week 7 (2/24 – 2/28)

Tuesday: Simple and complex cells

Tutorial 5: Modeling simple and complex cells

Thursday: Edge detection (B&G, Chapter 5)

Week 8 (3/3 – 3/7)

Tuesday: Edge detection

Tutorial 6: Edge detection

Thursday: Perceptual organization (B&G, Chapter 6)

Lab 3: Energy model for edge detection (due 3/25)

Week 9 (3/10 – 3/14)

Spring break

Week 10 (3/17 – 3/21)

Tuesday: Energy models for texture segmentation

Tutorial 7: Texture segmentation

Thursday: Stereopsis (B&G, Chapter 7, pp. 137-159)

Week 11 (3/24 – 3/28)

Tuesday: Stereopsis

Tutorial 8: Stereo energy models

Thursday: Lab day

Lab 4: Depth from stereo (Due: 4/4)

Week 12 (3/31 – 4/3)

Tuesday: Statistical limits to vision (Supplemental reading)

Tutorial 9: Limits to absolute sensitivity

Thursday: Signal detection theory – Signal detection reading

Week 13 (4/3 – 4/7)

Tuesday: Signal detection theory and ideal observers (Supplemental reading)

Tutorial 10: Signal detection theory

Thursday: Pattern recognition

Due: Project proposal

Week 14 (4/10 – 4/14)

Tuesday: Bayes' classifier

Tutorial 11: Letter recognition

Thursday: Efficient image coding – accounting for natural statistics (Wandell, Chapter 8, 247-258)

Week 15 (4/17 – 4/21)

Tuesday: Decorrelating transforms and image compression

Tutorial 12: Image compression

Thursday: Free day

Week 16 (4/24 – 4/28)

Tuesday: Free day

Thursday: Final projects due