The Application Of MCC And Propagation Algorithm In Fingerprint Cross-matching

The demand The demand of solving the problem of fingerprint cross-matching is derived from the rapid development in the field of biometric recognition in recent years. The popularization of sensor technology makes the types of the fingerprint sensor more and more diverse. Typical types in the market includes optical fingerprint sensor, swiped fingerprint sensor, dabbed fingerprint sensor and so on. As the principles of acquisition are different and standards、specifications of fingerprint sensor are different, even using the same finger, some fingerprint images by different fingerprint sensors have difference, while others have much difference. Traditional method of fingerprint matching is poor for the interoperability of different acquisition instrument, which caused a lot of resources waste including acquisition instrument waste and fingerprint images waste. According to some global or local features less affected by nonlinear deformation and location, direction of minutiae in local area not susceptible to nonlinear deformation in the fingerprint cross-matching, compiled by MATLAB, on the basis of experiment, the present paper proposed a new matching strategy based on MCC(Minutia Cylinder-Code) and propagation algorithm.The key of fingerprint cross-matching is the large nonlinear deformation result from the difference of fingerprint acquisition principle and the resolution. MCC(Minutia Cylinder-Code) and propagation algorithm has resistance to deformation in fingerprint cross-matching. In this paper the results are as follows: 1. The application of MCC which is a new characteristic expression method in the new matching strategy. 2. The application of a new propagation algorithm in the fingerprint cross-matching.3. Reducing EER to 3.48% and speed up to 20 ms per pair. The experimental results show that the proposed algorithm by this paper effective.There are some shortcomings in this paper and something need to be further improved in the subsequent research work. 1.The similarity of MCC just reflects the relationship between the minutiae and minutiae. Such as direction field, frequency field and other important information haven’t been fully taken advantage of. In the subsequent experiments, orientation field and frequency field can be used to build descriptors based on minutiae, further to distinguish the true or false of the minutiae, and obtain the more excellent calculation method of the similarity of the minutiae. 2. The algorithm is only applied on the AES and URU. We can continue experimenting in subsequent work toshow that this algorithm has better interoperability for different acquisition instrument.