Introduction of a Mutual Feature between Electrodes into Support Vector Machine Based Person Verification Using Evoked Electroencephalogram by Ultrasound

In user management, to realize continuous user authentication, we study the use of an electroencephalogram (EEG) evoked by ultrasound as biometrics. In previous studies, using a spectrum and four nonlinear quantities in EEG as individual features and a support vector machine (SVM) as a verification method achieved an equal error rate (EER) of 0 %. However, it required a large number of SVM models, wherein considerable amount of computation regarding learning was consumed. In this study, we introduce a mutual feature between electrodes and confirm its effectiveness in achieving EER = 0 % with a smaller number of SVM models.