大数据环境下利用新型FTS的并行细节点指纹匹配通用分解方法
|
摘要
随着指纹识别数据库规模的不断扩大,指纹识别系统的通用性、可靠性亟须提高。为了解决该问题,提出了利用新型指纹拓扑结构(FTS)的并行化细节点指纹匹配分解方法。该方法首先根据k近邻算法设计了叉点和端点新的结构,并提取出细节点特征。然后,将匹配分数的计算过程分解为几个小步骤来执行,在更精细的层次上分割2个指纹的最终匹配,以此定义局部结构子集之间的部分分数。最后,单独计算这些部分分数,将它们合并在一起,构成一个非常灵活的预测值,并允许丢弃部分分数。此外,提出了基于细节点置信度的指纹匹配算法,有助于全局范围上的指纹信息提取,从而确保局部相似细节点的有效匹配。在SFinGe数据库上的实验结果表明:所提出的分解框架可适用于Apache Hadoop、Apache Spark等大数据环境,具有良好的可靠性。将提出的分解方法应用于3种匹配算法中进行实验,结果表明提出的分解方法具有良好的通用性。
With the continuous expansion of the size of the fingerprint recognition database,the flexibility and reliability of the fingerprint identification system need to be improved. In order to solve this problem,a parallel minutiae fingerprint matching scheme combining novel fingerprint topology structure( FTS) is proposed. Firstly,the new structure of the fork point and the end point is designed by the k-nearest neighbor algorithm and the minutiae feature is extracted. Then,the process of calculating the matching score is divided into several small steps to execute,and the final matching of the two fingerprints is divided to be a finer level,thereby defining the partial scores between the partial structure subsets. Finally,these fractional scores can be calculated separately and then merged together to form a very flexible predictor and allow partial fractions to be discarded. In addition,a fingerprint matching algorithm based on the confidence of detail points is proposed,which is helpful to extract fingerprint information in the global scope,thus ensuring the effective matching of local similar detail points. The experimental results on SFinGe database show that the proposed decomposition framework can be applied to Apache Hadoop,Apache Spark and other data environments,and has good reliability. The proposed decomposition method is applied to three matching algorithms and the results show that the proposed decomposition method has good generality.
With the continuous expansion of the size of the fingerprint recognition database,the flexibility and reliability of the fingerprint identification system need to be improved. In order to solve this problem,a parallel minutiae fingerprint matching scheme combining novel fingerprint topology structure( FTS) is proposed. Firstly,the new structure of the fork point and the end point is designed by the k-nearest neighbor algorithm and the minutiae feature is extracted. Then,the process of calculating the matching score is divided into several small steps to execute,and the final matching of the two fingerprints is divided to be a finer level,thereby defining the partial scores between the partial structure subsets. Finally,these fractional scores can be calculated separately and then merged together to form a very flexible predictor and allow partial fractions to be discarded. In addition,a fingerprint matching algorithm based on the confidence of detail points is proposed,which is helpful to extract fingerprint information in the global scope,thus ensuring the effective matching of local similar detail points. The experimental results on SFinGe database show that the proposed decomposition framework can be applied to Apache Hadoop,Apache Spark and other data environments,and has good reliability. The proposed decomposition method is applied to three matching algorithms and the results show that the proposed decomposition method has good generality.
引文
[1]张莉,李甫,吴开腾.无方向的三角形匹配指纹识别[J].中国图象图形学报,2017,22(9):1214-1221.
[2]PERALTA D,GALAR M,TRIGUERO I,et al.A survey on fingerprint minutiae-based local matching for verification and identification[J].Information Sciences,2015,315(C):67-87.
[3]吴建立,刘宏申.一种新的形状描述与识别方法[J].重庆理工大学学报(自然科学),2017,31(2):110-116.
[4]吴震东,王雅妮,章坚武.基于深度学习的污损指纹识别研究[J].电子与信息学报,2017,39(7):1585-1591.
[5]WANG Y,WANG L,CHEUNG Y M,et al.Learning Compact Binary Codes for Hash-Based Fingerprint Indexing[J].IEEE Transactions on Information Forensics&Security,2015,10(8):1603-1616.
[6]PERALTA D,TRIGUERO I,SANCHEZ-REILLO R,et al.Fast fingerprint identification for large databases[J].Pattern Recognition,2014,47(2):588-602.
[7]CASSALES G W,CHARAO A S,PINHEIRO M K,et al.Context-aware Scheduling for Apache Hadoop over Pervasive Environments[J].Procedia Computer Science,2015,52(1):202-209.
[8]ZAHARIA M,XIN R S,WENDELL P,et al.Apache Spark:a unified engine for big data processing[J].Communications of the Acm,2016,59(11):56-65.
[9]KITANO T,SU L.SPOAN:Load Balancing Replica Placement Strategy for Large Scale Biometric Identification Service[C]//IEEE International Congress on Big Data.USA:IEEE,2013:326-333.
[10]SHELLY,RAGHAVA N S.Iris recognition on Hadoop:Abiometrics system implementation on cloud computing[C]//2011 IEEE international conference on cloud computing and intelligence systems.2011:482-485.
[11]JIANG X,YAU W Y.Fingerprint Minutiae Matching Based on the Local and Global Structures[C]//International Conference on Pattern Recognition,2000.
[12]CAPPELLI R,FERRARA M,MALTONI D.Minutia Cylinder-Code:a new representation and matching technique for fingerprint recognition[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2010,32(12):2128-41.
[13]HASAN H,ABDUL-KAREEM S.Fingerprint image enhancement and recognition algorithms:a survey[J].Neural Computing&Applications,2013,23(6):1605-1610.
[14]袁姮,王志宏,姜文涛.基于复合梯度向量的指纹匹配算法[J].电子学报,2017,45(4):912-921.
[15]ZHAO Yun-xiang,ZHANG Wan-xin,LI Dong-sheng,et al.Pegasus:a distributed and load-balancing fingerprint identification system[J].Front Inf Technol Electron,2016,17(8):766-780.
[16]王伟希,袁杰,臧炅,等.基于局部特征的点模式指纹匹配算法[J].南京大学学报(自然科学版),2009,45(1):18-23.
[17]CAPPELLI R,MAIO D,MALTONI D.Synthetic Fingerprint-Database Generation[C]//International Conference on Pattern Recognition,2002.Proceedings.Canada:IEEE,2002:744-747.
[2]PERALTA D,GALAR M,TRIGUERO I,et al.A survey on fingerprint minutiae-based local matching for verification and identification[J].Information Sciences,2015,315(C):67-87.
[3]吴建立,刘宏申.一种新的形状描述与识别方法[J].重庆理工大学学报(自然科学),2017,31(2):110-116.
[4]吴震东,王雅妮,章坚武.基于深度学习的污损指纹识别研究[J].电子与信息学报,2017,39(7):1585-1591.
[5]WANG Y,WANG L,CHEUNG Y M,et al.Learning Compact Binary Codes for Hash-Based Fingerprint Indexing[J].IEEE Transactions on Information Forensics&Security,2015,10(8):1603-1616.
[6]PERALTA D,TRIGUERO I,SANCHEZ-REILLO R,et al.Fast fingerprint identification for large databases[J].Pattern Recognition,2014,47(2):588-602.
[7]CASSALES G W,CHARAO A S,PINHEIRO M K,et al.Context-aware Scheduling for Apache Hadoop over Pervasive Environments[J].Procedia Computer Science,2015,52(1):202-209.
[8]ZAHARIA M,XIN R S,WENDELL P,et al.Apache Spark:a unified engine for big data processing[J].Communications of the Acm,2016,59(11):56-65.
[9]KITANO T,SU L.SPOAN:Load Balancing Replica Placement Strategy for Large Scale Biometric Identification Service[C]//IEEE International Congress on Big Data.USA:IEEE,2013:326-333.
[10]SHELLY,RAGHAVA N S.Iris recognition on Hadoop:Abiometrics system implementation on cloud computing[C]//2011 IEEE international conference on cloud computing and intelligence systems.2011:482-485.
[11]JIANG X,YAU W Y.Fingerprint Minutiae Matching Based on the Local and Global Structures[C]//International Conference on Pattern Recognition,2000.
[12]CAPPELLI R,FERRARA M,MALTONI D.Minutia Cylinder-Code:a new representation and matching technique for fingerprint recognition[J].IEEE Transactions on Pattern Analysis&Machine Intelligence,2010,32(12):2128-41.
[13]HASAN H,ABDUL-KAREEM S.Fingerprint image enhancement and recognition algorithms:a survey[J].Neural Computing&Applications,2013,23(6):1605-1610.
[14]袁姮,王志宏,姜文涛.基于复合梯度向量的指纹匹配算法[J].电子学报,2017,45(4):912-921.
[15]ZHAO Yun-xiang,ZHANG Wan-xin,LI Dong-sheng,et al.Pegasus:a distributed and load-balancing fingerprint identification system[J].Front Inf Technol Electron,2016,17(8):766-780.
[16]王伟希,袁杰,臧炅,等.基于局部特征的点模式指纹匹配算法[J].南京大学学报(自然科学版),2009,45(1):18-23.
[17]CAPPELLI R,MAIO D,MALTONI D.Synthetic Fingerprint-Database Generation[C]//International Conference on Pattern Recognition,2002.Proceedings.Canada:IEEE,2002:744-747.