Extraction Of Fingerprint Pore Feratures Based On Wavelet Image Enhancement

Fingerprint has two properties, namely, uniqueness and permanence, which has made fingerprint identification a more and more widely used tool in security and law enforcement applications. Generally, fingerprint features are categorized into three levels. Level one are patterns of the fingerprint, depicting the global features of fingerprint; level two are local features, such as ridge bifurcations and endings; level three are shape features, including all dimensional attributes of ridge. Of these three levels, level three is the most accurate. However, the extraction of fingerprint features currently are mainly focused on the first and second levels, due to the limitation of low resolution of fingerprint image. That makes a huge amount of fingerprint details ignored. To extract level three features, it's necessary to enhance fingerprint images. Now that wavelet are capable of multi-resolution analysis(MRA) and can analyze in local time and frequency domain. Therefore,this paper introduces one of wavelet's characteristics, and utilizes it to enhance the fingerprint images. Before transforming the fingerprint images by wavelet, we deal with the images using common standardized method, so the different images may have the same gray-scale on average and variance that makes a convenience to match the fingerprint images. And then advances the other improved soft threshold algorithm which is in wavelet domain to enhance the images. Finally, we utilize a modified minimum squared error approach on fingerprint image, extracting the locations of pore , which is one of the level three features. The results of this method can present the fingerprint features in a better way.