On the extraction of pattern features from continuous measurements

A sub-optimum method of extracting features from continuous data belonging to two pattern classes is presented. The set of features selected minimize bounds on the probability of error obtained from the Bhattacharyya distance and the Hajek divergence. The random processes associated with the two pattern classes are assumed to be Gaussian with different means and covariance functions. The results represent an extension of the existing results for classes with the same means and different covariance functions.