Boosted Bayesian Kernel Classifier Method for Face Detection

Abstract

In this paper, we present a novel face detection approach based on adaboosted relevance vector machine (RVM). The novelty of this paper comes from the construction of the kernel classifier with different kernel parameters. We use Fisher's criterion to choose a subset of Haar-like features. The proposed combination outperforms in generalization in comparison with support vector machine (SVM) on imbalanced classification problem. The combination of boosting algorithm and RVM classifier will yield accurate and sparse model which will perform well in real-time application. This method is compared, in terms of classification accuracy, to other commonly used methods, such as SVM and RVM without boosting, on CBCL face database. Results indicate that the performance of the proposed method is overall superior to previous approaches with very good sparsity.