摘要

According to fingerprint recognition, the matching algorithms are often unsatisfactory for the incomplete fingerprint due to skin distortions. Therefore, this paper presents a hybrid matching scheme which is based on quality estimation. Some algorithms are used and modified by extracting the quality of fingerprint in this hybrid scheme. Estimating the quality of incomplete fingerprint can choose the better sub-regions and neglect the worse, adaptively. Furthermore, this scheme can combine the advantages of different algorithms and ignore the wrong matching results. Hence, this scheme can match the incomplete fingerprint better. A series of experimental have been conducted on FVC2002 to evaluate the proposed scheme. Experimental results confirm our scheme is effective for fingerprints verification.

全文