Study On Image Matching And Recognition Technology Of Low Quality Finger Vein

In the field of information security,biometrics has gradually become the main method of identity authentication owing to its unique advantages.Biometrics include fingerprint recognition,iris recognition,face recognition,DNA recognition and finger vein recognition.The finger vein recognition technology has gradually been researched by many experts and scholars in this field due to its characteristics like living characteristic,high safety,non-contact,light weight and so on.The finger vein products are also been widely used in bank ATMs,Social Security Bureau,Access control and insurance systems for prisons and office buildings.However,finger vein image acquisition is often interfered by the acquisition environment,the device itself,and improper use of the collectors,which may lead to low-quality acquised images and seriously affect the performance of the product recognition.Low-quality images have become the main factor of the promotion of the finger vein recognition products.How to improve the recognition performance of low-quality finger vein images is the main research object of this dissertation,and the research on the finger vein recognition system has been deeply studied.The specific research contents are as follows:A template image registration mechanism based on finger vein image quality is proposed.The idea of this mechanism is to add an image quality decision mechanism in the image acquisition module,and to evaluate the image quality of the collected images from the aspect of image exposure,venous blur,and venous sparseness.Only the images judged to be normal quality can be registered to the database,thereby the system's recognition rate and control system error rate are improved.Theoretical analysis and experimental results show that the proposed three image evaluation methods can achieve the same effect as the subjective evaluation,and the three-step image quality discrimination and the filtering of low-quality images can improve the recognition performance of the system.A rapid finger vein identification algorithm based on line distance order statistics is proposed.This algorithm has three innovative points: First,the vein thin line is viewed as a recognition object,which gives up the feature points used as traditional point set distance matching.Second,the idea of order statistics is used to the vein thin line point distance matching,which effectively reduces the effect of pseudo-veins of the low-quality fingers on the matching results.Third,a matching method based on neighborhood search is proposed to solve the problem of high computational complexity due to the excessive number of fine-line points.The combination of these three innovations is a rapid identification algorithm for finger vein based on the thin line distance order statistics.The theoretical analysis and experimental results show that the proposed algorithm has better performance than the widely used MHD algorithm based on feature points.