Ear Recognition Using Texture Features - A Novel Approach

Ear is a new class of relatively stable biometric that is invariant from childhood to old age. It is not affected with facial expressions, cosmetics and eye glasses. Human ear is one of the representative human biometrics with uniqueness and stability. Ear Recognition for Personal Identification using 2-D ear from a side face image is a challenging problem. This paper analyzes the efficiency of using texture features such as Gray Level Co-occurrence Matrix (GLCM), Local Binary Pattern (LBP) and Gabor Filter for the recognition of ears. The combination of three feature vectors was experimented with. It is found that the combination gives better results compared to when the features were used in isolation. Further, it is found that the recognition accuracy improves by extracting local texture features extracted from sub-images. The proposed technique is tested using an ear database which contains 442 ear images of 221 subjects and obtained 94.12% recognition accuracy.