Direction filtering reinforcement method of fingerprint image

Abstract

The invention is a direction filter enhancement method of a fingerprint image. It mainly includes normalization processing of fingerprint image, directional field estimation of fingerprint ridges and directional filter enhancement. The purpose of the normalization process is to eliminate the influence of uneven finger pressing on the fingerprint image, and make full preparations for accurately extracting the direction information of the ridges. The invention uses a mask method to extract the direction information of the ridges, which has high accuracy and lays a solid foundation for subsequent direction filtering enhancement. In the directional filter enhancement of fingerprint ridges, the present invention uses a smoothing filter template to smooth the tangent direction of the ridges, and uses a sharpening filter template to sharpen in the normal direction of the smoothed ridges, which can significantly enhance the fingerprint ridges and valleys. It can improve the contrast of the line, and it can also connect the break of the ridge line caused by the dry finger and remove the adhesion of the ridge line caused by the wet finger, and achieve a good enhancement effect.

Description

Technical field [0001] The invention relates to a fingerprint image preprocessing method before fingerprint image recognition, in particular to a directional filter enhancement method in fingerprint image preprocessing. Background technique [0002] The preprocessing of fingerprint images is the most important part of the automatic fingerprint identification system. The result of preprocessing will directly affect the reliability and accuracy of recognition. Therefore, the preprocessing of fingerprint images has become a hot spot in the field of fingerprint recognition. The fingerprint image preprocessing includes a fingerprint image filter enhancement process. In the prior art, "Acta Electronica Sinica" 2004 issue 1 provides a fingerprint image addition method "based on ridge fingerprint filtering algorithm". The article points out: In the fingerprint collection process, various noises will inevitably be introduced When it is serious, it may even cause wrong information to be ex...

AI Technical Summary

Problems solved by technology

[0008] (1) Using extremum filtering and smoothing filtering to remove noise before enhancing the image will lose a lot of detail information of the fingerprint image, which is not desirable

[0009] (2) Before the directional filtering of the image is enhanced, the original fingerprint image is not gray-scale normalized, which cannot eliminate the impact of uneven finger pressure on the fingerprint image (that is, the central color of the fingerprint is often heavy, and the peripheral color is light) situation), which will adversely affect the accuracy of subsequent direction information extraction

[0010] (3) The algorithm of direction field estimation in this method is to calculate the direction angle of each pixel point first, and then quantize the direction of the pixel point to eight directions according to the direction angle, that is to say, from the obtained In the process of converting the direction angle value into the corresponding quantized direction value, there are uncertain factors in the quantization selection. Therefore, the direction value is difficult to be accurate, which will directly affect the effect of ridge filter enhancement according to the eight quantized directions.

[0011] (4) In this method, only smoothing filter templates are designed in the directional filter enhancement, and no sharpening filter templates are designed, and the templates are small, so it is difficult to accurately reflect the characteristics of filtering along the direction of the ridges, and the local direction changes sharply or there are ridges. Adhesive fingerprint image enhancement is less than ideal

Images