Fingerprint Identification Based on Ridge Lines and Graph Matching

摘要

A novel fingerprints matching algorithm which is robust to spurious features and elastic distortion has been developed. The feature that used for fingerprints matching was the ridge line that connects the minutia point with the associated ridge point. The proposed algorithm decomposes the ridge lines matching into two stages. In the first stage, the matching results of input and template local quadrilateral sets were used to estimate the transformations such as translation and rotation between an input and a template fingerprint. In the second stage, a coarse pair set was achieved firstly by matching two aligned ridge line sets in relatively slack matching qualifications, and then a graph matching algorithm was presented to match the coarse pair set again, to reduce the contributions of possibly false matched ridge lines due to the use of the slack matching qualifications. The final matching score was developed based on both the matching score of ridge line sets and the matching score of global orientation fields. Experiments show that the proposed algorithm is effective.