基于Farwell范式诱发ERP的身份识别研究  

Research of person identification using ERP based on Farwell paradigm

作　　者：刘晶[1] 白艳茹[1] 许敏鹏[1] 殷涛[2] 何峰[1] 周鹏[1] 綦宏志[1,2] 明东[1] 

机构地区：[1]天津大学精密仪器与光电子工程学院,天津300072 [2]中国医学科学院生物医学工程研究所,天津300192

出　　处：《电子测量与仪器学报》2015年第2期221-226,共6页Journal of Electronic Measurement and Instrumentation

基　　金：国家自然科学基金(61172008;81171423;81222021;30970875;90920015);国家科技支撑计划(2012BA134B02);教育部新世纪优秀人才支持计划(NCET-10-0618)资助项目

摘　　要：人体脑电特征作为一种新型的生物识别特征正受到越来越多的关注,但事件相关电位是否可用于身份识别尚未得到充分论证。本研究采集了含有55名被试者的较大数据集,对Farwell范式诱发的事件相关电位的个体间差异性与个体内一致性进行分析,进而提取ERP波形特征,利用支持向量机进行身份识别。结果显示靶刺激诱发的ERP特征对55位被试的身份识别正确率可达到96.36%,而非靶刺激诱发的ERP特征识别率可达到98.18%。本研究证实了Farwell范式诱发的事件相关电位可以作为一种新型的生物识别特征,在身份识别或认证等领域具有良好的应用潜力。Personal EEG character has been attracted more and more attention as a new type of biometric feature.However,it has not been fully demonstrated whether event related potential（ ERP） can be used in person identification. In this research,a Farwell Paradigm based experiment is executed and large data sets of 55 subjects＇ ERP are collected. Then the inter-individual difference and intra-individual consistency of ERP are analyzed. The wave feature of ERP is extracted and a Support Vector Machine is trained for individual identifying. The result shows that a person identification accuracy of 96. 36% with ERP feature under target stimulus and 98. 18% with ERP feature under non-target stimulus among 55 subjects. This research proves the feasibility of Farwell Paradigm evoked ERP using as a new kind of biometric feature and its potential application in person recognition and authentication.

关 键 词：事件相关电位 身份识别 脑-机接口 Farwell范式 支持向量机 

分 类 号：R138[医药卫生—劳动卫生] TN911.7[医药卫生—公共卫生与预防医学]