A Fingerprint Registration Point Extraction Method Based on Central Singular Point

Abstract

The invention belongs to the technical field of fingerprint registration point extraction, and discloses a fingerprint registration point extraction method based on a central singular point, which calculates the index value PI of all pixels in the singular point area 1 (i,j) and PI 2 (i,j), according to the index value, find the block where the central singular point appears, and take the geometric center of the block center as the registration point; if there is no central singular point, use the descending iteration method to find the convergence point of the ridge line, and use the convergent point As a registration point; find the initial point of iteration, find four equidistant points around the initial point, find the nearest ridge line of each point, and make a vertical line on the ridge line, get the intersection point of the vertical line, as the next iteration point ; If the Euclidean distance between the initial iteration point and the new iteration point is less than a certain threshold, the iteration ends, and the new iteration point is the reference point; otherwise, the new iteration point is used as the initial point to continue the iteration. The invention can quickly and accurately calculate the coordinates of the fingerprint registration point, and can be used in an automatic fingerprint identification system.

Description

technical field [0001] The invention belongs to the technical field of fingerprint registration point extraction, in particular to a fingerprint registration point extraction method based on a central singular point. Background technique [0002] In the automatic fingerprint identification system, finding the fingerprint registration point is a very critical step. The registration points of fingerprint images are generally used in the registration process of fingerprint images. Fingerprint registration is to unify two fingerprint images into the same absolute coordinate system for comparison through rotation and translation. The speed and accuracy of registration points determine the speed and quality of fingerprint registration. In most fingerprint recognition systems, the fingerprint registration points are either minutiae points or singular points. Both methods have shortcomings: when minutiae points are used as registration points, due to the limitation of minutiae poin...

AI Technical Summary

Problems solved by technology

[0003] To sum up, the problems existing in the prior art are: the registration accuracy of the current fingerprint identification system is low; therefore, an accurate registration algorithm will greatly improve and enhance the overall performance of the fingerprint identification system. The core of the registration algorithm is an accurate registration point positioning

Images