Interest and research activities in automatic face recognition (FR) have increased significantly within the last several years. Such a large growth is motivated by the growing application demands in many areas, including: identity authentication, access control, face-based video indexing or browsing, human-computer interaction or communication, etc. Up to now, numerous algorithms have been proposed for FR, and two issues are central to all these algorithms: 1) what features can be used to represent a face, and 2) how to classify a new face image based on the chosen representation. The main work of this paper is to design an efficient classifier, we employ boosting algorithm to improve the correct recognition rate (CRR) of classic face recognition algorithms.The ensemble learning algorithms combine the weak classifiers into a composite classifier, which is more robust than the original one. Boosting algorithm is an excellent ensemble learning algorithm that is very popular in a wide range of applications. It divides the whole classifier into many sub-classifiers; each sub-classifier concentrates on a special hard task, then combine the sub-classifiers into a composite classifier. In this paper, two new algorithms are proposed, called B-JD-LDA (Boosting on JD-LDA), B-PRM (Boosting on Probabilistic Reasoning Model). Experiments on the FERET and ORL database demonstrate that both algorithms get higher CRR than original algorithms. |