Font Size: a A A

Research On Contactless Palm Vein Recognition

Posted on:2018-05-26Degree:MasterType:Thesis
Country:ChinaCandidate:P Z WangFull Text:PDF
GTID:2348330512479295Subject:Signal and Information Processing
Abstract/Summary:PDF Full Text Request
The palm vein recognition is a biometric identification technique for identifying a person by analyzing his palm vein information.Because of its recognition accuracy,animate authentication and easy to accept,palm vein recognition has been widely studied in recent years.And it has been applied to security,finance and social security funds successfully.Palm vein image is generally collected contactlessly.The quality of collected image is bound to affect the recognition result.Previous studies on palm vein recognition focused on the feature extraction algorithm mostly.They did not think highly of the preprocessing algorithm after palm vein image acquisition.However,for the palm vein image collected contactlessly,the position,orientation of palms and the angle of fingers are different at each acquisition.Therefore,the result of preprocessing will directly affect the accuracy of the final recognition.For resolving the above-mentioned problems,in this thesis,palm vein image preprocessing algorithm has been studied and discussed carefully,and a new method of extracting palm vein feature and matching method are proposed.A good recognition result has been achieved.This thesis mainly includes the following works:(1)The traditional palm vein preprocessing algorithm did not consider various complex situations of obtaining the region of interest(ROI),this thesis studies the preprocessing algorithm systematically and quantitatively.And the contrast experiment based on fixed length and relative length ROI is carried out.The experimental results show relative length could acquire richer information.(2)A new feature extraction method of palm vein based on principal curvature in the palm vein feature extraction phase is presented.Firstly,ROI is normalized to remove the noise.Then the maximum curvature is calculated and the palm vein structure is distinguished from the surrounding human tissue.Finally,the image is segmented and clear palm vein characteristics are extracted.(3)For the palm vein feature extracted by the principal curvature,this thesis proposes template matching for the feature matching.LBP matching is also used for contrast experiments.Experimental results show that when the edge length of ROI is about 1.5 times the distance between the valley point of index finger and middle finger and the valley point of little finger and ring finger,the best recognition result could be achieved.And the palm vein recognition algorithm based on principal curvature combined with template matching can achieve the EER of 1.965%,which has an accurate result.
Keywords/Search Tags:Palm vein recognition, OTSU, Principal curvature, Template matching, LBP
PDF Full Text Request
Related items