隐私保护的心电图身份识别技术研究

Research on Privacy Preserving ECG-based Identification Technology

心电图是与个体紧密相关的生理特征,用于身份认证有着无可比拟的优势。然而,心电图反映了人体的健康状况,属于重要的个人隐私。文章提出了一种隐私保护的心电图身份识别技术,首先在数据的训练阶段和匹配阶段采用一定机制进行心电图隐私保护,然后分别采用欧几里得距离算法和互相关算法对隐私保护后的心电图数据进行识别实验。结果显示:对于公用数据库MIT-BIH Normal Sinus Rhythm Database中的心电图数据,使用欧几里得距离算法和互相关算法的识别率都能达到100%。对于公用数据库MIT-BIH Arrhythmia Database中的心电图数据,使用欧几里得距离算法和互相关算法的识别率都能达到96.77%。

ECG data are physiological characteristics that are closely related to an individual,which has an unparalleled advantage for authentication. However, ECG data reflect the health situationof an individual, which belong to the important personal privacy. This paper proposes a privacypreserving ECG-based identification technology. Firstly, a certain mechanism is adopted to protect theECG data in the data training phase and the data matching phase, and then identification experimentson the protected ECG data are conducted by the Euclidean distance algorithm and the cross-correlationalgorithm. The results show that the ECG data in MIT-BIH Normal Sinus Rhythm Database are 100 %identified by the Euclidean distance algorithm and the cross-correlation algorithm, and the ECG datain MIT-BIH Arrhythmia Database are 96.77% identified by the Euclidean distance algorithm and thecross-correlation algorithm.