Abstract: A system and method for classifying input patterns into two classes, a class-of-interest and a class-other, utilizing a method for estimating an optimal Bayes decision boundary for discriminating between the class-of-interest and class-other, when training samples or otherwise, are provided a priori only for the class-of-interest thus eliminates the requirement for any a priori knowledge of the other classes in the data set to be classified, while exploiting the robust and powerful discriminating capability provided by fully supervised Bayes classification approaches. The system and method may be used in applications where class definitions, through training samples or otherwise, are provided a priori only for the classes-of-interest. The distribution of the other-class may be unknown or may have changed. Often one is only interested in one class or a small number of classes.

Abstract: A system and method for estimating the a priori probability of a class-of-interest in an input-data-set and a system and method for evaluating the performance of the adaptive Bayes classifier in classifying unlabeled samples from an input-put-data-set. The adaptive Bayes classifier provides a capability to classify data into two classes, a class-of-interest or a class-other, with minimum classification error in an environment where a priori knowledge, through training samples or otherwise, is only available for a single class, the class-of-interest. This invention provides a method and system for estimating the a priori probability of the class-of-interest in the data set to be classified and evaluating adaptive Bayes classifier performance in classifying data into two classes, a class-of-interest and a class-other, using only labeled training samples, or otherwise, from the class-of-interest and unlabeled samples from the data set to be classified.

Abstract: This invention relates generally to a system and method for correlating two images for the purpose of identifying a target in an image where templates are provided a priori only for the target. Information on other objects in the image being searched may be unavailable or difficult to obtain. This invention treats the design of target matching-templates and target matched-filters for image correlation as a statistical pattern recognition problem. By minimizing a suitable criterion, a target matching-template or a target matched-filter is estimated which approximates the optimal Bayes discriminant function in a least-squares sense. Both Bayesian image correlation methods identify the target with minimum probability of error while requiring no prior knowledge of other objects in the image being searched. The system and method is adaptive in that it can be re-optimizing (adapted) to recognize the target in a new search image using only information from the new image.

Abstract: This invention relates generally to a system and method for classifying input patterns into two classes, a class-of-interest or a class-other, utilizing an Adaptive Fisher's Linear Discriminant method capable of estimating an optimal Fisher's linear decision boundary for discriminating between the two classes, when training samples are provided a priori only for the class-of-interest. The system and method eliminates the requirement for any a priori knowledge of the other classes in the data set to be classified. The system and method is capable of extracting statistical information corresponding to the “other classes” from the data set to be classified, without recourse to the a priori knowledge normally provided by training samples from the other classes. The system and method can re-optimize (adapt) the decision boundary to provide optimal Fisher's linear discrimination between the two classes in a new data set, using only unlabeled samples from the new data set.

Abstract: A system and method for extracting “discriminately informative features” from input patterns which provide accurate discrimination between two classes, a class-of-interest and a class-other, while reducing the number of features under the condition where training samples or otherwise, are provided a priori only for the class-of-interest thus eliminating the requirement for any a priori knowledge of the other classes in the input-data-set while exploiting the potentially robust and powerful feature extraction capability provided by fully supervised feature extraction approaches. The system and method extracts discriminate features by exploiting the ability of the adaptive Bayes classifier to define an optimal Bayes decision boundary between the class-of-interest and class-other using only labeled samples from the class-of-interest and unlabeled samples from the data to be classified. Optimal features are derived from vectors normal to the decision boundary defined by the adaptive Bayes classifier.