Gender recognition is one of the most popular research areas in security, biometrics and human computer interaction applications . In previous studies, structural and textural features of facial expressions were mostly used to identify gender. One of the biggest challenges of gender recognition is differentiating textual features of faces that decrease the accuracy of the proposed method, and there are lots of factors such as media, ambient lighting and environmental conditions. In order to overcome these disadvantages, firstly, a new database which has different expressions of face images is created. Then, some feature extraction and classification methods are used to improve recognition accuracy. Principal Component Analysis is used both for feature extraction and dimension reduction. Also, Local Binary Pattern Analysis which is frequently used in gender recognition is used at that stage. In classification stage, Euclidean and Manhattan classifiers are used. Finally, all methods' recognition performances are compared using the classification accuracy of applied methods.