A Robust Fingerprint Alignment and Matching using Multiple Features

In this paper, a novel fingerprint alignment and matching is introduced. It uses multiple features to align fingerprints and then matching. The proposed system uses nonminutiae features such as SURF and BRISK features for alignment and minutiae features for matching. Fingerprint alignment has been done with the help of SURF key points for initial alignment followed by BRISK key points to get the final alignment. Minutiae features are used to find the similarity score between two fingerprints which has helped to check the correctness of the alignment. To verify the proposed alignment approach, a novel matching score calculation method has been presented. This approach is different from the traditional fingerprint alignment approaches which are rely heavily on minutiae features. The proposed algorithm is robust, efficient and accurate. Further, it has been compared with well known alignment approaches on FVC2002 dataset. The proposed approach is found to be superior and has the Equal Error Rate (EER) of 0.035%.