We develop an approach for gender recognition based on human faces. We combine rectangle features extracted from the human-face region into a rectangle-feature vector (RFV). The RFV is computationally fast and effective in encoding intensity variations of local regions of human face. By only using few rectangle features learned by AdaBoost, we present an effective gender identification approach. We then use nonlinear support vector machines for classification, and obtain more accurate recognition results. Experimental results show that our approach performs well for the Feret database.
Relation:
Far East Journal of Experimental and Theoretical Artificial Intelligence 3(2), pp.101-112
