Slap fingerprint segmentation using symmetric filters based quality

In this paper, a slap fingerprint segmentation algorithm is proposed which can accurately segment the fingerprints from a given slap-image and classify them as index, middle, ring or little finger of left/right hand. To improve its performance, quality of components is used to eliminate several non-fingerprint components. It is defined by using symmetric filters which measure the uniformity in ridge-valley structure irrespective of linear or curved pattern. Some of the remaining non-fingerprint components are removed by using geometrical locations. For performance evaluation, a challenging database is used which contains 500 slap-images. Experimental results reveal the proposed algorithm can correctly segment the fingerprints from the slap-images with an accuracy of 99.40%, which is found to be better than other existing algorithms.