Palm Vein Recognition Algorithm Based On Multiply Features

With the progress of society,the whole society becomes the combination of information network,so the security problem is more important,and identity identification is one of basic method.The traditional identification methods include key,password and so on,but these methods are easy to use,embezzlement,loss.Therefore,the identification technology of biometric identity has appeared.Commonly used biometric identification techniques include fingerprints,faces,irises,veins and so on.The palm vein recognition has attracted more and more attention and studied by more and more people because of its stability,characteristics and uniqueness.Palm vein identification is a technique to identify the identity of an unknown user by extracting the features of the palm vein,which generally includes three steps: image pre-processing,feature extraction,feature recognition or verification.This article mainly studies the palm vein recognition algorithm,introduces two commonly used ROI region extraction methods and compares several image denoising methods.It focuses on the extraction and recognition algorithm of the palm vein features,improves the traditional NMRT algorithm and innovatively proposes the fusion of the NBP features of two palms.The main research works and achievements of this article are as follows:1?Describe Two commonly used ROI region extraction methods in detail,and introduce and compare several commonly used image denoising methods.Finally,select the ROI region extraction algorithm based on the palm rectangle and median filter algorithm to achieve the best recognition effect.2?On the basis of palm vein recognition in traditional NMRT algorithm,we increase the sampling processing of ROI area and the judgement of vein points.The recognition rate of the improved NMRT algorithm proposed in this chapter is 99.64%on the PolyU palmprint library.Compared with the traditional NMRT algorithm,it greatly shortens the time consumed and ensures the basic recognition effect.3?A new recognition algorithm is proposed,which combines the NBP features of the two palms and two kinds of two-hand palm feature fusion recognition algorithm based on multi-samples are proposed.The recognition rate of the two-palm fusion recognition algorithm on the PolyU palmprint library is 98.93%,which is obviously higher than the traditional single palm recognition algorithm,and the recognition rate ofthe two multi-samples fusion recognition algorithms is 99.89% and 99.78%.4?Finally,compare the improved NMRT algorithm recognition and the two-palm feature fusion recognition,the improved NMRT algorithm has a better recognition effect but more time-consuming,while the two-palm feature fusion recognition takes a short time,but the recognition effect is not good enough.