Human biometrics is a popular method for security applications of personal identification. In past years, most research topics focused on face, iris, and fingerprint or signature recognition. Apart from these features, human noses can also be used as a biometric feature for gender classification according to our daily experience. In this work, four distinct features are extracted from nose images to distinguish between male and female faces. By comparing with other nose features obtained from 3D stereoscopic photographs, the proposed features are easily extracted from ordinary face images. Image preprocessing methods are applied to extract those four features automatically, and the linear discriminant analysis (LDA) method is applied to classify those extracted features for gender classification. Experimental results demonstrate that average classification accuracy can reach to 77%. In particular, the curvature feature calculated from the nose wing achieves the best classification performance.