This course will cover current topics in computer vision with an emphasis on motion estimation, learning methods, and probabilistic models. Readings will be from recent research papers. Computational techniques such as Expectation-Maximization, Hidden Markov Models, and Belief Propagation will be introduced. Applications to tracking, recognition, image enhancement, and human motion analysis will be considered. Prerequisites: basic probability, linear algebra, and calculus.
This year will be unlike previous years. There will be no big real-world project and no murders to solve.
Instead this will be a much more traditional seminar course. Each class will focus on an important topic in vision and we will read at least 3 papers on that topic. One will be a seminal paper that introduced or defined the field; this will provide the historical context. This will also expose you to some great older papers that are often cited and seldom read. The second will be a paper by me and my collaborators (unless we have the seminal paper). This will give us one paper where the "expert" is in the room. This will also be useful for people who are interested in working with me. The third (and possibly fourth) will be recent, state-of-the-art methods that show how the problem has evolved and where the field is heading.
The course will be divided into 3 topic areas: Image motion and optical flow (5 weeks), Markov random fields and statistical image models (3 weeks), and Human pose estimation and tracking (4 weeks).
A computer vision course, e.g. CS143 (Introduction to Computer Vision) or an equivalent course.I'll assume good familiarity with linear algebra, calculus, probability, statistics, (e.g. CS155, AM0040, AM165, AM169, or AM264).
Michael Black
Office: CIT 521
Email: black<at>cs<dot>brown<dot>edu
Office Hours: by appointment
There will be few if any formal lectures by me. Everyone is expected to read all papers and participate in the discussion. Every paper will have a student presenter who is repsonsible for preparing a talk of approximately 20 minutes and for leading the discussion. You will be graded on this presentation. Everyone will get to present several papers.
Things I want you to learn: 1) how to read a paper critically, find its "nugget", its flaws and the potential for improvement; 2) improve your skills at preparing and giving talks; 3) the history of some important areas in computer vision; 4) the current state of the art in these areas.
