The Research On Fingerprint Image Enhancement Algorithm

In recent years, fingerprint identification technology has provided important basis forthe reliability of the identification, which is most widely used among a lot of biologicalrecognition technologies. However, in the process of collecting fingerprints, there alwaysappears fuzzy, fracture or other noise in the fingerprint.That is because of the influence ofthe inappropriate finger pressure, the shortcoming of fingerprint or the stains on thesurface of the instrument. To deal with this kind of bad equality fingerprint image willseriously affect the accuracy of the final results of recognition. Therfore, it is necessary toenhance the equality of fingerprint image. So, the enhancement algorithm of fingerprintimage is researched in this paper with Matlab simulation, in order to improve the equalityof fingerprint image furtherly. The concrete reaserch content mainly includes:First, this paper briefly introduces the fingerprint identification technology and theimportance of fingerprint image enhancement, and analyses the related research of thefingerprint image enhancement algorithm in recent years.Second, in view of normalization treatement, the paper analyses the choice ofparameters to make the normalization processing of fingerprint image having better effect.Again, the paper uses the mask method to get the point direction filed of fingerprintimage, and obtains the corresponding local block direction through the statistical method.It carries on the tag to the reliability of the direction field, and then deals with the noise offingerprint image to improve the reliability of the direction filed.Then, the paper uses the projection method to get the frequency information of thelocal areas with parallel ridge, and estimates the frequnce information of other local areaswith singular points or noise on the basis of the theory of Gaussian kernel. Thus, the wholecorresponding frequency field is achieved.Finally, directional filter and Gabor filter are researched in detail. The researchincludes the theory of directional filter and the principle of selecting the suitable filtertemplates, and the theory of Gabor filter and associative parameters. Respectively，thepaper uses directional filter and Gabor filter to deal with fingerprint image, and compares the result of two kinds of filtering.