Research On The Pretreatment Of Fingerprint Identification System Based On Wavelet Analysis

Fingerprint has become the most reliable research object of thebiology identification technique for it's sole and unchangeable. Becausefingerprint identification technique has good implement and practicality,it has become one of the most prevalent, convenient and reliable identityauthentication techniques. So the development of automatic fingerprintidentification system has become the insistent need of the modeminformation safety. The pretreatment of fingerprint is the key content offingerprint identification and it is also the key technology of fingerprintidentification system.The development history and status quo both here and abroad ofthe resent fingerprint identification are summarized in this paper firstlyand the principle and general method of the fingerprint identificationsystem are also summed up, which includes the acquisition,pretreatment, feature extraction and matching of the fingerprint image.All of these help us grasping all parts of the fingerprint identification onthe whole. On the basis of this, the paper mainly researches the part of the fingerprint image pretreatment, which includes the filtering of thefingerprint image's noise, binarization and thinning, of which thefiltering of the fingerprint image's noise is the leading content of thepaper.The main research achievements are as followings: 1.Summarizingthe general methods and steps of the denoising of image according to thecharacter of the image's noise and analyzing the noise basing on thewavelet analysis combined with the denoising methods which is themain content of the paper. 2. Emphatically analyzing the application ofthe Gabor filter on the denoising of the fingerprint image. Aiming at thecondition of the method of traditional Gabber filter cut no ice on thetexture of asymmetric fingerprint image, this paper make some progresson the parameter of the Gabor filter. 3. As to the explosive growth of thewavelet filter in the signal filter field, its theory and implement methodare summarized in the paper. According to the noising character offingerprint image, the filter method of wavelet threshold is improved inthe paper, which includes the algorithm research on the filter of waveletthreshold value, the selection of the threshold value function and thedecision of the threshold value. Finally, we present the result andanalysis of the experiment. 4. On the basis of wavelet filter fingerprintimage, the article shows the ordinary binarization method and procedurebased on wavelet analysis and we improved and post-processed the thinning of the fingerprint image. 5. According to the fight simulationresult, in view of the complexity and relative independency of fingerprintimage pretreatment algorithm, we code to implement the fingerprintimage wavelet threshold filter under VC environment.The technique route and algorithm in the paper are all verifiedand coded to implement, which got a good effect. They have theoreticand reference value, especially we made a through on the pre-process ofthe fingerprint image which underlined the consolidate basis for the nextprocess of the fingerprint system.