Effectiveness of assigning confidence levels to classifiers and a novel feature in fingerprint matching

There are different methods and techniques used for matching fingerprints but the most common and popular approach is minutiae based matching. Our approach is based on structural matching and the matching algorithm presented here is the improved and modified form of. In this method, matching is done on the basis of five closest neighbors of one single minutia that is also called a center minutia. An authentication of minutia is based on these surrounding neighbors. The approach we present here is divided in to two stages. First stage performs initial filtration and the second stage includes special matching criteria that incorporate fuzzy logic as well as a novel feature to select final minutiae for matching score calculation. The method of selecting center point for second stage is also adapted. This algorithm is able to perform well for translated, rotated and stretched fingerprints and does not require any process for alignment before matching. Experimental results show that algorithm is efficient and reliable.