A Performance Evaluation of Fingerprint Minutia Descriptors

Minutiae-based matching method is the most widely used fingerprint matching method and most minutiae matching algorithms employ minutia descriptors of certain type. A good minutia descriptor should be distinctive and at the same time robust in difficult situations, i.e., noise, occlusion and plastic distortion. Many different minutia descriptors have been proposed in the literature, which can be coarsely classified into three categories: image-based, texture-based, and minutia-based. These descriptors have not been evaluated systematically. In this paper, seven different descriptors are evaluated according to local matching accuracy for four types of fingerprint pairs: good quality, poor quality, small common region, and large plastic distortion. Experimental results show that the texture-based descriptor and Minutia Cylinder-Code (MCC) are the two most accurate descriptors. In the case of small common region, texture-based descriptor ranks first, while in the rest three cases, MCC is the most accurate one.