| Fingerprint identification is a common biometrical method of hman identity verification. It has developed many years and still is the hot field by many reseachers. This paper tries to study and improve the fingerprint identification algorithm at first. A new algorithm utilizing multiple filter bank is applied in fingerprints enhancement in this paper. It uses the OTSU method to obtain ridge orientation by calculating the Fourier Transform of fingerprints. Considering the specific orientation characteristics of the fingerprint ridges, each fixed-coefficient filter of the filter bank is described with separable radial filter and directional filter. Experimental results show that this method can enhance low quality fingerprint images obviously with less computation workload, and is very useful for the application of real-time fingerprint identification system. To improve the speed of minutiae matching, A state-of-the art 3D based matching method is selected specifically. After the algorithm is established, a fingerprint sensor is connected to the computer to set up the fingerprint identification system. Considering the complexity of the proposed method, FFT, a frequently used and computational expensive module is extract to design to Intellectual Property or fingerprint identification chip. This is a beneficial attempt to improve the processing speed of fingerprint identification system. |
