Boosting multiple classifiers constructed by hybrid discriminant analysis

In this paper, a set of hybrid dimension reduction schemes is constructed by unifying principal component analysis (PCA) and linear discriminant analysis (LDA) in a single framework. PCA compensates LDA for singular scatter matrix caused by small set of training samples and increases the effective dimension of the projected subspace. Generalization of hybrid analysis is extended to other discriminant analysis such as multiple discriminant analysis (MDA), and the recent biased discriminant analysis (BDA), and other hybrid pairs. In order to reduce the search time to find the best single classifier, a boosted hybrid analysis is proposed. Our scheme boosts both the individual features as well as a set of weak classifiers. Extensive tests on benchmark and real image databases have shown the superior performance of the boosted hybrid analysis.