Gender Classification based on Fusion of Facial Components Features

The last few years have seen a great interest in image processing and as a result research in this field has been undertaken into the classification of gender based on face components because of its applications in database searching, marketing and knowledge that the human face carries a lot of information which may be extracted and used for many various purposes. This paper presents gender classification based on the fusion of facial components, eyes, nose, mouth, forehead and cheeks as many factors such as occlusion may affect visibility as they block the camera from viewing a facial image and to offset the effect of some features which have a lower classification rate. An artificial neural network with back propagation is used here for classification. The findings of this research showed us that with fusion the accuracy rates of components with lower classification can be improved. A good example being the mouth having a percentage accuracy of 75%, however after fusing it with the feature vector of the nose which has an individual classification of 90% we get an accuracy of 87%. This is of great use when large part of the component with a higher classification accuracy is occluded.