GRADUATE NEWS

Feature extraction is a fundamental problem in pattern recognition and image processing. Learning machines such as support vector machines (SVM) can address this problem, but there are many issues with their use. For any machine learning classification method, the benchmark is Bayes method that establishes an optimal classification framework. I show that proper regularization of linear SVM allows convergence to the Bayes optimal solution for data sets that have the same covariance matrix. Principal components are essential system features for the classification problem. I show how eigenstructures are an inherent part of the SVM functional basis that encodes the geometric features of a separating hyperplane. SVM architectures based on insufficient eigenstructures are shown to have insufficient learning capacity. SVM returns large numbers of features for overlapping distributions. I develop a principal component feature extraction method for both linear and nonlinear SVM. The PCA compression technique offers an effective method for selectively tuning in a refined manner, the amount of support vectors (SV) used to construct SVM decision boundaries. SV-PCA features provide core components for matched filter banks for both binary and m-ary classification problems.

Posted on: May 14, 2009