A Fingerprint Matching Method Based on Bipartite Graph Best Matching

Abstract

The invention discloses a fingerprint matching method based on bipartite graph optimal matching. First, a series of reference point pairs are obtained by partial matching, and then the obtained reference point pairs are verified at a higher level by constructing a topology structure. Make the matching effect reach the global optimum. The invention includes: calculating the similarity between the input fingerprint minutiae point and any minutiae point pair of the template fingerprint; constructing a bipartite graph; using the KM algorithm to obtain the best matching. It further includes the step of removing wrong matching point pairs by establishing the topology structure of minutiae points. The invention highlights the globality in image matching, further improves the robustness of the algorithm, has strong anti-noise ability, can take both matching precision and operation time into account, and makes the matching effect reach the global optimum.

Description

technical field [0001] The invention belongs to the technical field of image processing, and relates to a fingerprint recognition technology, more specifically, a fingerprint matching method based on bipartite graph optimal matching is designed. Background technique [0002] In the information society, identity authentication is the basis for safe and fast information transmission. However, traditional identity authentication is more and more difficult to meet the needs of the society because it is easy to forge and lose. The most convenient and safe solution is undoubtedly biometrics. technology. Biometric identification technology is the use of human physiological or behavioral characteristics to identify individuals. Not only is it simple and fast, but it is also safer, more reliable, and more accurate to use it for identity identification; at the same time, it is easier to integrate with computers and security, monitoring, and management systems to realize automated man...

AI Technical Summary

Problems solved by technology

However, due to rigid deformation, elastic deformation, partial overlap and other reasons, the false recognition rate will increase, which will reduce the accuracy of detail point matching.

The method based on local matching also obtains reference point pairs between two matching fingerprint images through the local features of minutiae points, and finally calculates the matching score based on the local matching results of all reference point pairs. Good similarity makes the misrecognition rate of this type of method higher

In addition, the fingerprint image is affected by the noise caused by the finger skin is too dry, too wet, the fingerprint path is very shallow, the finger has a wound, the skin is peeled off, the finger is dirty during collection, the imaging characteristics of the collection equipment, and the residue on the collection plane. , will make the feature set unreliable and reduce the matching effect of minutiae points

To sum up, the existing fingerprint matching methods based on point patterns have many defects, and the correct rate is low.

Images