Research On The Sweat Pores Matching Algorithm Of High Resolution Fingerprint

In recent years,optoelectronic technology,fingerprint collection instruments and methods have made great progress,as makes the high-resolution fingerprint recognition is possible.High resolution fingerprint has special third layer sweat pore feature,which is not only unique,but also plentiful,anti-fake,and have different shapes.In the fragmentary fingerprint matching,compare to the secondary feature of the minute point,sweat feature has more information,which is irreplaceable advantage.However,the current high-resolution fingerprints recognition is limited by the influence of plentiful sweat pores,deformation,displacement,incomplete fingerprints,noise,and first two layer features,the fingerprint matching accuracy and matching speed need to be improved.At present,there are many fingerprint matching algorithms,which mostly focus on the fusion matching,which mixed fingerprint ridge feature,minute point feature and sweat pore feature,so,the sweat pore matching performance directly depends on the matching effect of ridge and minute,the matching effect of first layer characteristic ridge line and the second layer characteristic minute points have direct influence on the matching of the third layer sweat pore feature.In this paper,we improve the influence of the alignment of singular point,which is the basement of fusion matching.The matching of ridges and sweat pores are carried out on the basis of fine minute points matching,the experiments get good matching effect.Fingerprint alignment algorithm based singular point:fingerprint singularity extraction technology is mature,the extraction accuracy of singularity point is relatively high,but sometimes,the position of singular point is offset,inaccurate or missing,so we should take into account the poor extraction situation of singular point when we align two fingerprints by their singular points.In this paper,for the above questions,we get the optimal singularity point position by supposing their positions.In this alignment method,we judge the effect by the accuracy of matching minute points,at the same time,we take into account the special situation of the ridge line in fingerprint,that is,the global optimality achieved by point set matching is not necessarily the optimal case of fingerprint matching.This alignment method improves the matching effect and achieves a better alignment effect.After the end of the fingerprint alignment,we can get the best alignment of two fingerprints,which includes fine corresponding minute point pairs and rotation angle of the two matched fingerprints.According to the above information,we extract the ridge lines of two fingerprints for the ridge line matching,we take the average nearest distance of every two corresponding points and improved average Frechet distance between the ridge sampling points curves as the quantification index,in experiment,the ridge line matching has got a good effect.Sweat pores matching includes two methods,one method is named sweat pores matching based minute points,which needs to select sweat pores nearest corresponding minute points;the other is named sweat pores matching based ridge lines,which needs to select sweat pores nearest every ridge line.These two sweat matching methods have acquired good effects.Finally,experiments were conducted at the Hong Kong Polytechnic University's high-resolution fingerprint database.Experiments show that the above algorithms used in this paper have some improvements over the matching effects of other matching algorithms.