Syllabus: CSC2539 Topics in Computer Vision, Winter 2012

Professors:
David Fleet, Office: Pratt 391
Allan Jepson, Office: Pratt 283?

Office hours: Please send email for appointments

Lectures:
Thursdays 11:00am-1:00pm,
Room BA4010

Synopsis

This is a seminar course to explore recent topics in video processing, with emphasis on tracking complex objects and tne analysis of people. The course is structured as a participitory seminar course. Each meeting will consist of some combination of lecture, paper presentation, and discussion.

The topics of the readings are flexible as there are many exciting papers to read and discuss. Our initial plans are to include topics such as image-based tracking, model-based tracking, part-based human detection/parsing, 2d and 3d articulated people tracking, models for biological model, activity and gesture recognition, and the analysis of scene dynamics. The first six to eight weeks will be somewhat structured, covering well established research problems and exemplary articles published in journals and conferences in the last few years. The topics for the last few weeks be depend on student interests.

Students are expected to have a reasonable background in computer vision, with knowledge comparable to that covered in CSC2503, Foundations of Computer Vision.

Readings

There is no course textbook. Each week we will select readings, some of which will be assigned as required readings.

Course Work

Students are required to present papers to the class, to complete short (1-2 page) reports on assigned readings each week (when not presenting), to complete one introductory assignment, and then a course project.

The assignment is to be completed individually. The goal of the initial assignment is to give you hands-on experience with image sequences, thereby providing students with richer perspective when presenting papers and choosing a course project.

The course project is to be completed individually.