EE/CNS/CS148 -  Spring 2004

Selected Topics of Computational Vision

Visual Recognition
Summary:
 Humans use vision to recognize objects that they have seen before. They can also generalize from a few examples and recognize classes of objects, e.g. cars, sneakers, dogs. We will explore the computational theory and the engineering issues that underpin object recognition. In particular: image processing, principal component analysis, pattern classification, elements of decision theory, probabilistic models of shape, and geometric hashing. The homework and lab assignments will guide the students in building a software system that learns object categories from examples and recognizes objects in pictures.
Prerequisites: Elements of linear algebra, probability, statistics, signal processing. Some familiarity with either Matlab or C/C++ computer programming. The necessary background in computer vision, pattern recognition, geometry and probabilistic modelling will be developed in class.
Final grade: 100% based on weekly homework.
Instructor:
 Pietro Perona
TAs:
 Pierre Moreels Alex Holub
Time and Place:
 T-Th 10:30-12, 102 Steele
 
Details:
 Background material
Class organization
Lecture Notes
References
Homework Assignments
Links