TEL 613E - Statis.Pattern .Analy.&Classif

Course Objectives

To teach fundamentals of Statistical Pattern Analysis and Classification concepts and algorithms,
To teach students programming techniques for machine learning and to encourage them to incorporate these techniques into their own research.

Course Description

The statistical theory of pattern analysis, including both parametric and nonparametric approaches to classification. Multidimensional probability distributions. Classification with likelihood functions and Bayesian estimation, linear discriminant functions, density estimation, supervised and unsupervised classification, performance estimation, and classification using sequential and contextual information, including Markov and hidden Markov models. Gaussian mixture models. Eigenvalue decomposition. Feature reduction, combination of classifiers and sensor fusion. Application areas include image and video processing