Fingerprint twisting degree estimation method and system

Abstract

The invention provides a fingerprint twisting degree estimation method. The method comprises the following steps of selecting an input fingerprint and a reference fingerprint, and calculating the twisting degree of the reference fingerprint; respectively conducting registration on the input fingerprint and the reference fingerprint to obtain a feature vector of the input fingerprint and a feature vector of the reference fingerprint; according to the twisting degree and the feature vector of the reference fingerprint, obtaining a twisting degree estimator; inputting the feature vector of the input fingerprint into the twisting degree estimator so as to obtain the twisting degree of the input fingerprint. The fingerprint twisting degree estimation method has the higher resolution capacity, can accurately estimate the twisting degree of the fingerprint, and is high in applicability, and capable of being well applicable to an existing fingerprint recognition system. The invention further provides a fingerprint twisting degree estimation system.

Description

technical field [0001] The invention relates to the technical field of fingerprint identification, in particular to a fingerprint distortion estimation method and system. Background technique [0002] As the most popular and successful identification feature, fingerprint plays a very important role in biometric identification. In the past few decades, fingerprint recognition technology has achieved rapid development, and has been applied to many fields from criminal investigation to civilian use, and plays an important role in it. Even so, there are still many challenges in fingerprint recognition technology, and the estimation of distorted fingerprints is one of the very difficult problems. The twisting occurs due to the elasticity of the finger skin and the sideways force or moment applied by the reader when taking the fingerprint. When the force or moment on the side is relatively large, the fingerprint of the collector will be severely distorted, which will cause the m...

AI Technical Summary

Problems solved by technology

The disadvantages of this algorithm are: 1) need a special pressure sensor or a fingerprint sensor with video capture function, which increases the cost and size of the hardware, and is not suitable for many applications; 2) cannot handle the distorted fingerprints that have been collected in the past; 3 ) cannot cope with the distorted fingerprints of the skin itself, that is, the fingerprints of fingers distorted by chemical agents or surgical procedures

The disadvantage of this scheme is that the feature dimension is too low, and the features are defined by humans rather than learned from actual data, so it is difficult to accurately estimate the distortion of fingerprints.

Images