contains material from
Template Matching Techniques in Computer Vision: Theory and Practice
Roberto Brunelli © 2009 John Wiley & Sons, Ltd
The chapter formally introduces template matching as a hypothesis testing problem. The Bayesian and frequentist approaches are considered with particular emphasis on the Neyman-Pearson paradigm. Matched filters are introduced from a signal processing perspective and simple pattern variability is addressed with the normalized Pearson correlation coefficient. Hypothesis test often requires the statistical estimation of the parameters characterizing the associated decision function: some subtleties in the estimation of covariance matrices are discussed.
keywords: hypothesis testing, classification, Bayes risk criterion, Neyman-Pearson criterion, matched filters, correlation coefficient, maximum likelihood estimation, James-Stein estimator, shrinkage.