Boosting in Random Subspaces for Face Recognition

Boosting is an excellent machine learning algorithm. In this paper, we propose a novel boosting method - boosting in random subspaces. Instead of boosting in original feature space, whose dimensionality is usually very high, multiple feature subspaces with lower dimensionality are randomly generated, and boosting is carried out in each random subspace. Then the trained classifiers are further combined with simple fusion method. Compared with boosting in original feature space, there are two advantages. The first is that the computation complexity of training is reduced, which is obvious. The second is that fusion further improves accuracy, which is verified by our extensive experiments on FERET database