Support vector machines (SVMs) have been recently proposed as a new learnin
g network for bipartite pattern recognition. In this paper, SVMs incorporat
ed with a binary tree recognition strategy are proposed to tackle the multi
-class face recognition problem. The binary tree extends naturally, the pai
rwise discrimination capability of the SVMs to the multi-class scenario. Tw
o face databases are used to evaluate the proposed method. The performance
of the SVMs based face recognition is compared with the standard eigenface
approach, and also the more recently proposed algorithm called the nearest
feature line (NFL). (C) 2001 Elsevier Science B.V. All rights reserved.