Hybrid feature-based teeth recognition system

Biometric technology is used in several security applications at present. The biometric information and classification have large impact in efficiency of the application. Teeth are chosen as a characteristic for biometric system in this research since teeth are difficult to be reshaped by intent surgery like well-known features such as face, palmprint and fingerprint. In this paper, appropriate features including machine learning model for teeth recognition are proposed. The features of our system can be classified into two categories: global and local features. Global features are composed of singular values and color histogram of teeth image whilst local features are obtained from normalized teeth width. For the experiments, our proposed features are applied into a system of the multilayer perceptrons network with Levenberg-Marquart backpropagation training algorithm. The performance of our system is better than other existing techniques in terms of accuracy and false acceptance.