Simple effective fingerprint segmentation algorithm for low quality images

Fingerprint segmentation is one of the first and most integral pre-processing steps for any fingerprint verification system, and it determines the results of fingerprint analysis and recognition. In this paper, we have proposed a novel algorithm for fingerprint segmentation. The fingerprint image gradient are discussed and its application to fingerprint segmentation is presented. Firstly, Gauss filter is used to remove noise and then we utilize histogram equalization to enhance the contrast between background and foreground. Secondly, a novel gradient based method is adopted and then obtain the approximate foreground region of the fingerprint images. Finally, we smooth the image to eliminate the isolated pieces and block effect of fingerprints. The method acquires the threshold combining the basic gradient value distributing character. It is simple to implement and the performance of the proposed algorithm is evaluated on FVC2004 DB3 database. Experimental results show the new adaptive algorithm is effective and reliable, especially for fingerprint images of low quality.