Fingerprint recognition using local features and Hu moments

Person identification systems based on fingerprint patterns called Automatic Fingerprint Identification Systems, AFIS, are some of the most widely used biometric methods since they provide a high degree of success. The accuracy of AFIS is mainly due to some unique characteristics called minutiae, which are points where a curve track finishes, intersects with another curve track, or branches off. During past decades several efficient minutia-based fingerprint recognition algorithms have been proposed which achieve false recognition rates close to 1%, however, their recognition rate may be still improved. To this end, this paper presents a fingerprint recognition method using a combination of the Fast Fourier Transform (FFT) with Gabor filters for image enhancement. Next, fingerprint recognition is carried out using a novel recognition stage based on Local Features and Hu invariant moments for verification.