Component-Based Gender Identification Using Local Binary Patterns

Abstract

In this paper a component-based gender identification model from facial images has been proposed. The paper enhances the gender identification by using individual facial components (forehead, eyes, nose, cheeks, mouth and chin). Group of frontal facial images are used to validate the proposed model, feature extraction technique Local Binary Patterns (LBP) is implemented, then KNN and SVM classification techniques are applied to accomplish the gender identification model. The results achieved in this research work show an improved accuracy rate when face components (eyes, nose, mouth) are used for gender identification instead of the whole facial image. These results indicate that there are some facial parts which are not necessary for facial image recognition related application like gender identification.