基于特征级潜在信息的多生物特征识别方法
|
摘要
针对多模态生物特征识别主流集成方法中信息利用不充分的问题,提出利用特征级潜在信息实现得分级集成的思路。挖掘各模态样本的类内聚合度和类间离散度以及模态间相关度等特征级潜在统计信息,将所挖掘的信息利用F-Ratio模型实现集成。实验中将所提方法与当前具有代表性的集成方法进行对比。经实验证明,所提方法在识别准确度上优于同样使用F-Ratio模型,但未利用特征级潜在信息的集成方法,同时优于其他多个具有代表性的集成方法。
Aims at overcoming the insufficient usage of information in the current mainstream fusion methods of multimodal biometrics,the idea of realizing score level fusion by potential information of feature level is proposed. Based on profound analysis, potential statistical information of feature level such as intra-class aggregation and inter-class dispersion of feature samples,correlation among modals is mined and utilized. Fusion is realized by the well theoretical based F-Ratio model. In experiments,the proposed method is compared with currently representative fusion methods.The experimental results indicate that the proposed method advances in recognition accuracy compared with the method of F-Ratio model which is realized without the usage of feature level information,also,is superior to many other representative fusion methods.
Aims at overcoming the insufficient usage of information in the current mainstream fusion methods of multimodal biometrics,the idea of realizing score level fusion by potential information of feature level is proposed. Based on profound analysis, potential statistical information of feature level such as intra-class aggregation and inter-class dispersion of feature samples,correlation among modals is mined and utilized. Fusion is realized by the well theoretical based F-Ratio model. In experiments,the proposed method is compared with currently representative fusion methods.The experimental results indicate that the proposed method advances in recognition accuracy compared with the method of F-Ratio model which is realized without the usage of feature level information,also,is superior to many other representative fusion methods.
引文
[1]Oloyede M O,Hancke G P.Unimodal and Multimodal Biometric Sensing Systems:A Review[J].IEEE Access,2016,4(1):7532-7555.
[2]Rane S,Wang Y,Draper S C,et al.Secure biometrics:Concepts,authentication architectures,and challenges[J].IEEE Signal Processing Magazine,2013,30(5):51-64.
[3]Ross A A,Nandakumar K,Jain A K.Handbook of multibiometrics[M].New York:International Series on Biometrics,Springer,2006.
[4]Haghighat M,Abdel M M,Alhalabi W.Discriminant correlation analysis:Real-time feature level fusion for multimodal biometric recognition[J].IEEE Transactions on Information Forensics and Security,2016,11(9):1984-1996.
[5]Liu Y,Yang Z,Suen C Y,et al.A Study on performance improvement due to linear fusion in biometric authentication tasks[J].IEEE Transactions on Systems,Man,and Cybernetics:Systems,2016,46(9):1252-1264.
[6]He M,Horng S J,Fan P,et al.Performance evaluation of score level fusion in multimodal biometric systems[J].Pattern Recognition,2010,43(5):1789-1800.
[7]Zhang D,Guo Z,Lu G,et al.Online joint palmprint and palmvein verification[J].Expert Systems with Applications,2011,38(3):2621-2631.
[8]Miaoli W.Face and speech recognition fusion method based on penalty coefficient and SVM[C]//Proc 2015 IEEE Advanced Information Technology,Electronic and Automation Control Conference(IAEAC’15).[s.l.]:[s.n.],2015:6-10.
[9]Chen Y,Yang J,Wang C,et al.Multimodal biometrics recognition based on local fusion visual features and variational Bayesian extreme learning machine[J].Expert Systems with Applications,2016,64(1),93-103.
[10]Julian F A,Daniel G R,Javier O G,et al.Bayesian adaptation for user-dependent multimodal biometric authentication[J].Pattern Recognition,2005,38(8):1317-1319.
[11]Nandakumar K,Chen Y,Dass S C,et al.Likelihood ratio based biometric score fusion[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(2):342-247.
[12]Tao Q,Veldhuis R.Robust biometric score fusion by naive likelihood ratio via receiver operating characteristics[J].IEEETransactions on Information forensics and security,2013,8(2):305-313.
[13]Sudhamani M J,Venkatesha M K,Radhika K R.Fusion at decision level in multimodal biometric authentication system using Iris and Finger Vein with novel feature extraction[C]//Proc 2014 Annual IEEE India Conference(INDICON).[s.l.]:[s.n.],2014:1-6
[14]Williams S.Pearson’s correlation coefficient[J].New Zealand Medical Journal,1996,109(1015):38.
[15]Liu L,Yin Y,Qin W,et al.Gait recognition based on outermost contour[J].International Journal of Computational Intelligence Systems,2011,4(5):1090-1099.
[16]Belhumeur P N,Hespanha J P,Kriegman D J.Eigenfaces vs.fisherfaces:recognition using class specific linear projection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1997,19(7):711-720.
[17]Jiang X D,Yau W Y.Unmatched minutiae:useful information to boost fingerprint recognition[J].Neurocomputing,2016,171(1):1401-1403.
[2]Rane S,Wang Y,Draper S C,et al.Secure biometrics:Concepts,authentication architectures,and challenges[J].IEEE Signal Processing Magazine,2013,30(5):51-64.
[3]Ross A A,Nandakumar K,Jain A K.Handbook of multibiometrics[M].New York:International Series on Biometrics,Springer,2006.
[4]Haghighat M,Abdel M M,Alhalabi W.Discriminant correlation analysis:Real-time feature level fusion for multimodal biometric recognition[J].IEEE Transactions on Information Forensics and Security,2016,11(9):1984-1996.
[5]Liu Y,Yang Z,Suen C Y,et al.A Study on performance improvement due to linear fusion in biometric authentication tasks[J].IEEE Transactions on Systems,Man,and Cybernetics:Systems,2016,46(9):1252-1264.
[6]He M,Horng S J,Fan P,et al.Performance evaluation of score level fusion in multimodal biometric systems[J].Pattern Recognition,2010,43(5):1789-1800.
[7]Zhang D,Guo Z,Lu G,et al.Online joint palmprint and palmvein verification[J].Expert Systems with Applications,2011,38(3):2621-2631.
[8]Miaoli W.Face and speech recognition fusion method based on penalty coefficient and SVM[C]//Proc 2015 IEEE Advanced Information Technology,Electronic and Automation Control Conference(IAEAC’15).[s.l.]:[s.n.],2015:6-10.
[9]Chen Y,Yang J,Wang C,et al.Multimodal biometrics recognition based on local fusion visual features and variational Bayesian extreme learning machine[J].Expert Systems with Applications,2016,64(1),93-103.
[10]Julian F A,Daniel G R,Javier O G,et al.Bayesian adaptation for user-dependent multimodal biometric authentication[J].Pattern Recognition,2005,38(8):1317-1319.
[11]Nandakumar K,Chen Y,Dass S C,et al.Likelihood ratio based biometric score fusion[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,2008,30(2):342-247.
[12]Tao Q,Veldhuis R.Robust biometric score fusion by naive likelihood ratio via receiver operating characteristics[J].IEEETransactions on Information forensics and security,2013,8(2):305-313.
[13]Sudhamani M J,Venkatesha M K,Radhika K R.Fusion at decision level in multimodal biometric authentication system using Iris and Finger Vein with novel feature extraction[C]//Proc 2014 Annual IEEE India Conference(INDICON).[s.l.]:[s.n.],2014:1-6
[14]Williams S.Pearson’s correlation coefficient[J].New Zealand Medical Journal,1996,109(1015):38.
[15]Liu L,Yin Y,Qin W,et al.Gait recognition based on outermost contour[J].International Journal of Computational Intelligence Systems,2011,4(5):1090-1099.
[16]Belhumeur P N,Hespanha J P,Kriegman D J.Eigenfaces vs.fisherfaces:recognition using class specific linear projection[J].IEEE Transactions on Pattern Analysis and Machine Intelligence,1997,19(7):711-720.
[17]Jiang X D,Yau W Y.Unmatched minutiae:useful information to boost fingerprint recognition[J].Neurocomputing,2016,171(1):1401-1403.