Research On Deformation Correction And Feature Extraction Of Finger Vein Images

Finger vein recognition technology is a biometric technology that uses finger veins as personal identification.Because the finger veins are inside the fingers,and the vein patterns of each person are different,it is widely used in identity authentication systems such as online payment,government and enterprise card punching,bank accounts and so on.However,in the non-contact acquisition system of finger vein images,there is no mesure to fix finger posture,and the finger vein images often have deformation problems.So this thesis studies the image deformation correction and feature extraction algorithms,and does the following work.(1)When correcting the image deformation problem,the traditional finger vein image deformation correction algorithm ignores the positioning and correction of the finger joints.So,this thesis proposes Gassian Location Dectection and Correction(GLDC)method to correct the deformation images.In the preprocessing,an Alternating Displacement Correction(ADC)method is proposed to correct the position of the finger in the image.The detection of the displacement of the finger joints and the positioning and correction of the finger joints are added.The simulation results show that The ADC algorithm enhances the uniformity of the image.The GLDC algorithm can recognize the image of finger displacement and rotation in the image when processing the deformed image,and can correct these images.Compared with some existing algorithms,it can better improve the unity between images and have better robustness.(2)In deformed images,the problem of unclear vein patterns often occurs,and traditional feature extraction algorithms are not effective.In response to this problem,this thesis introduces the PBBM extraction algorithm,and proposes a Personal Best Bit Map Based method(PBBMB).When extracting vein features,PBBMB method can obtain more effective vein features.The simulation results show that the PBBMB algorithm,has better results in extracting images with uneven grayscale and unclear lines.Make the system more effectively reduce the EERor matching rate,and further improve the accuracy of finger vein recognition.