Linear separability of gender classification
LI Wei-jie · Computer Engineering and Applications Journal · 2008
In this paper,the authors first analyze linear separability of face images for gender classification.The authors compare the popular linear feature extraction,nonlinear feature extraction,and a reformative nonlinear feature extraction method,and conduct gender classification experiments using these methods under different conditions.These experiments clearly reveal linear separability and classification performance of the feature extraction results obtained using different methods.The authors are also the first to state that gender classification issue should take complexion into account.Additionally,the authors propose a novel significant gender classification strategy and scheme.