Long sequence biometric hashing authentication based on 2D-SIMM and CQCC cosine values

doi:10.1007/s11042-021-11708-z

	Long sequence biometric hashing authentication based on 2D-SIMM and CQCC cosine values
	Huang, Yi-bo 1; Hou, Hexiang 1; Chen, Tengfei 1; Li, Hao 1; Zhang, Qiu-yu 2
刊名	MULTIMEDIA TOOLS AND APPLICATIONS
	2021-11
卷号	81 期号:2 页码:2873-2899
关键词	Speech content authentication Biometric security template 2D-SIMM CQCC cosine Discrimination Unidirectional
ISSN号	1380-7501
DOI	10.1007/s11042-021-11708-z
英文摘要	The existing speech authentication algorithms hash extracted speech features directly and saved them to the cloud, which is easy to cause speech feature leakage. In the process of constructing hashing, the utilization efficiency of speech feature is poor, and the short hashing sequence will lead to the lack of discrimination of hashing sequence and the deviation of authentication. In order to solve the above problems, a long sequence biometric hashing authentication algorithm based on two-dimensional Sine ICMI Cmodulation map (2D-SIMM) and constant Q cepstral coefficients (CQCC) cosine was proposed. First, this algorithm extracts the CQCC of the speech signal, then obtains the eigenvalue of the space cosine distance of the adjacent speech frame CQCC, and finally performs projection mapping between the eigenvalue and the pseudorandom matrix generated by 2D-SIMM to construct a biometric hashing sequence. This paper evaluates the proposed robust feature schemes of MFCC and CQCC space cosine distance through experiments. The experimental results show that CQCC spatial distance combined with 2D-SIMM biometrics characteristics can reach 10(-21). when the threshold is 0.35. The BER mean was only 0.0383 for maintaining the robustness of operation for different contents. When the SNR is -5 dB, the matching rate of different noises can reach 45%. At the same time, it also improves the security of the biological template, and the overall performance is greatly improved compared with the existing algorithm.
WOS研究方向	Computer Science ; Engineering
语种	英语
出版者	SPRINGER
WOS记录号	WOS:000715720000001
内容类型	期刊论文
源URL	[http://ir.lut.edu.cn/handle/2XXMBERH/150082]
专题	计算机与通信学院
作者单位	1.Northwest Normal Univ, Coll Phys & Elect Engn, Lanzhou 730070, Peoples R China; 2.Lanzhou Univ Technol, Sch Comp & Commun, Lanzhou 730050, Peoples R China
推荐引用方式 GB/T 7714	Huang, Yi-bo,Hou, Hexiang,Chen, Tengfei,et al. Long sequence biometric hashing authentication based on 2D-SIMM and CQCC cosine values[J]. MULTIMEDIA TOOLS AND APPLICATIONS,2021,81(2):2873-2899.
APA	Huang, Yi-bo,Hou, Hexiang,Chen, Tengfei,Li, Hao,&Zhang, Qiu-yu.(2021).Long sequence biometric hashing authentication based on 2D-SIMM and CQCC cosine values.MULTIMEDIA TOOLS AND APPLICATIONS,81(2),2873-2899.
MLA	Huang, Yi-bo,et al."Long sequence biometric hashing authentication based on 2D-SIMM and CQCC cosine values".MULTIMEDIA TOOLS AND APPLICATIONS 81.2(2021):2873-2899.