Study On Automatic Fingerprint Identification System Based On Wavelet Analysis

AFIS (Automatic Fingerprint Identification System) is a kind of ideal safe technique. The application fields of AFIS are very extensive and have vast of market. The research interests of this dissertation are the compression of fingerprint image, the characteristic extraction of fingerprint image, and the characteristic match in AFIS.Most of current algorisms adopt Fourier transform. Since Fourier transform uses the sine function as Orthogonal-base function, and discrete Fourier transform uses not-zero-base vectors in the whole domain, which limits the application of Fourier transform. For example, some partial characteristics of an image (such as edge or line) apart from Fourier transform so that Fourier transform couldn't do enough to the part.This dissertation analyzes the antinomy between AFIS image data and transmission, clarifies the requirements of image compress, image process, image analysis and identification in the AFIS. The wavelet analysis is a new method that develops quickly in recent years. The contrast among Fourier transform, Gabor transform and wavelets transform are completed in the paper. An important work is to analyze the basic frame of wavelet transformation code and to apply wavelet analytical zero trees code into a specific fingerprint image compression processing in AFIS. The paper analyzes relation technique issues of fingerprint image analysis, identification, and precise match. It introduces the key method of image match based on the wavelet, the picture of "a trous" calculate, the picture of the low frequency composition, and the technique to apply to the system of AFIS of image match. These methods are efficiency to raise the quantity, stable, fast, accurate of the whole system.