Se-Woong Oh, Hyungwoo Park,Jong-Bae Kim

Biometric information used in biometrics includes face, fingerprint, iris, and vein. Among them, fingerprint is widely used as an authentication method in mobile devices. However, there are not many mobile devices that can use fingerprint authentication technology. Currently, fingerprint authentication technology in mobile devices requires a separate sensor, so it can be used only in high-end devices. Fingerprint authentication technology has a limitation of recognition rate due to environmental factors in the processing of existing image, and there are limitations in the case where authentication fails continuously. Therefore, a method of mounting a separate oath to improve performance is widely used. This paper addresses these limitations by using a camera that is included as a basis for mobile devices. We have studied a model to measure the similarity of fingerprint image obtained by camera using SIFT and SURF algorithm. Experiments were carried out by obtaining 10 different fingerprint images for each 10 experiments. The FER and FRR were used for the performance of the fingerprint recognition system. The ERR value of the SIFT algorithm was 1.4065% and the SURF algorithm was 4.4017%.

Volume 11 | 05-Special Issue

Pages: 264-272