CS342 Computational Photography

Fall 2007

Instructor Prof. David R. Martin http://vision.bc.edu/~dmartin dmartin@cs.bc.edu Office: Fulton 424C Phone: 552.2757	Admin. Support Jane Costello costeljl@bc.edu Office: Fulton 460 Phone: 552.3975	Tech. Support Phil Temples support@cs.bc.edu Office: Fulton 255 Phone: 552.0153
Location Fulton 310 T-Th 12:00-1:15pm	Lab Fulton 160 http://cslab.bc.edu

Course Catalog Description

How are digital images created and enhanced? How can a computer recognize people from their fingerprints, faces, or eyes? How are special effects done in the movies? This class introduces the computational and mathematical problems of modern digital photography primarily through applied linear algebra. Mathematical topics will be introduced as needed and in context, emphasizing intuition; examples are eigenvectors, the SVD, linear least squares, density estimation, affine maps and homographies, constrained optimization, and Lagrange multipliers. This is a hands-on course focused on programming projects using MATLAB. No prior experience with MATLAB is assumed, but students must have some programming experience. [Flier]

Prerequisites

Computer Science: CS101 Computer Science I or equivalent (see below)

Mathematics: MT202 Multivariable Calculus MT210 Linear Algebra

Basic programming skills are required. Programming will be done in Matlab. Students must know some modern high-level language such as C or Java, but no Matlab experience is necessary. The material assumes comfort with differential and integral calculus and linear algebra.

Grading

Review the University's policy on academic integrity.

Homework (75%) - The lab-like assignments are the most important part of the class, giving you hands-on experience implementing and experimenting with image processing algorithms.

Final project (25%) - The final project may be done individually or in pairs, and will consist of implementing an extension to one of the homework assignments or an algorithm discussed in class. Projects will be presented by students at the end of the term.

Textbook

There is no textbook for this class. Readings will be taken from various books and research papers. I recommend the following MATLAB reference book (available at the BC bookstore), though it is not necessary since MATLAB has exceedingly good online help.

MATLAB Primer, Seventh Edition. Timothy A. Davis and Kermit Sigmon. Chapman & Hall/CRC, 2005.