首页    期刊浏览 2022年07月07日 星期四
登录注册

文章基本信息

  • 标题:Feature Extraction using Sparse SVD for Biometric Fusion in Multimodal Authentication
  • 本地全文:下载
  • 作者:Pavan Kumar K ; P. E. S. N. Krishna Prasad ; M. V. Ramakrishna
  • 期刊名称:International Journal of Network Security & Its Applications
  • 印刷版ISSN:0975-2307
  • 电子版ISSN:0974-9330
  • 出版年度:2013
  • 卷号:5
  • 期号:4
  • DOI:10.5121/ijnsa.2013.5406
  • 出版社:Academy & Industry Research Collaboration Center (AIRCC)
  • 摘要:Token based security (ID Cards) have been used to restrictaccess to the Securedsystems.The purpose ofBiometricsistoidentify / verifythe correctness of an individualby using certain physiological orbehaviouraltraits associated with the person.Current biometric systems make use of face, fingerprints,iris,hand geometry,retina, signature, palm print, voiceprint and so on to establish a person's identity.Biometrics isone of the primary key concepts of realapplicationdomains such asaadhar card, passport,pancard, etc.In this paper, we consider face andfingerprint patternsforidentification/verification.Usingthis data weproposed a novel model for authentication in multimodal biometricsoften called Context-SensitiveExponentAssociative Memory Model (CSEAM).It provides different stagesof securityforbiometricsfusionpatterns.Instage1,fusion offace and finger patternsusingPrincipal ComponentAnalysis (PCA),in stage 2by applyingSparseSVD decomposition toextract the feature patternsfrom thefusion data and face pattern and thenin stage 3,using CSEAM model,theextracted feature vectorscan beencoded.Thefinal key will be stored in the smart cardsas Associative Memory (M), which is often calledContext-Sensitive Associative Memory (CSAM). In CSEAM model,theCSEAMwill be computed usingexponential kronecker productforencodingand verificationofthe chosen samplesfrom the users.Theexponentialof matrixcan be computed in various ways such as Taylor Series, Pade Approximation andalso using OrdinaryDifferential Equations (O.D.E.). Among these approaches we considered first twomethods for computing exponential of a feature space.The result analysis of SVD and Sparse SVD forfeature extraction process and also authentication/verification process of the proposed systemin terms ofperformance measuresasMean square error rateswill be presented.
  • 关键词:Biometrics;Biometric fusion; Face;Finger;print;Context;-;Sensitive;Exponent;Associative Memory Model;(CS;E;AM);Kronecker Product;Exponential Kronecker Product (eKP);Multimod;a;l Authentication;Singular Value Decomposition (;SVD); Sprase SVD (SSVD)
国家哲学社会科学文献中心版权所有