A method for face gender recognition based on blocking-LBP and SVM

As an important field of face recognition, gender recognition based on face has been paid more and more attention. This paper proposes a method for gender recognition based on blocking local binary pattern (LBP) and support vector machine (SVM). With the difference from those traditional methods for face image feature extraction, we divide a face image into several blocks, overlap or non-overlap, and the LBP histogram characteristics of these blocks are extracted and cascaded to form face feature vector. The SVM is used to carry out the gender recognition on the feature vectors. The analyses are given about the effect of the attachments of face and the different partitions of face image on recognition results. The detail experiment results show that our method gives higher accuracy.