Research Of Fingerprint Identification System Based On Spectral Minutiae

With the rapid development of information technology, the automatic fingerprintidentification system have wide application, the precision and speed of automatic fingerprintidentification system demands were also increased. But, the displacement, rotation and scale offingerprint images would be changed, and they were usually caused by the acquisition of fingerposition, the collecting pressure and spot light. And the existing fingerprint identification systemfor these linear change fingerprint couldn’t always meet the actual need.To solve the above problem, this paper proposed a novel fingerprint identification methodbased on the spectral of fingerprint minutiae. Firstly, this paper extracted the feature offingerprint minutiae in frequency domain. Secondly, this paper used the characteristics ofFourier-Mellin transform to eliminate the linear change to obtain the alignment-free feature.Finally, this paper utilized matlab to build a platform, and the new feature were applied to theautomatic fingerprint identification system. Because the new spectral minutiae feature wererepresented as a fixed-length feature vector,so the calibration did not require matching. Thus, itcould greatly reduce the matching time and speed up the recognition rate of the system. And italso could improve the recognition accuracy of the system, so as to improve the performance ofthe whole system.The experiment showed that, the proposed method could not only save time but alsoimprove the precision. And using this minutiae feature in automatic fingerprint identificationsystem could effectively improve the system performance.