Exclusive EEG identity identification method

Abstract

The invention discloses an exclusive EEG identity identification method using an EEG signal for identity identification. The method comprises: collecting EEG data; carrying out pretreatment on the EEG data; extracting an EEG feature parameter by using an AR model; establishing a BP neural network classifier; constructing a classification network connected in series with the BP classifier based on the classifier; and carrying out training on the EEG feature parameter by using the classification network and carrying out identity identification. With the method, brain waves of others can be eliminated with a near 100% of success rate. The method is also suitable for open-set data. According to the method, the action behavior requirement on a tested person during EEG signal collection is low.

Description

technical field [0001] The invention relates to the field of electroencephalogram identification, and specifically refers to an exclusive electroencephalogram identification method. Background technique [0002] EEG recognition is still an emerging technical field. There are still few existing methods of using EEG signals for identification. Basically, the EEG data is analyzed in the frequency domain, and EEG signals in a certain band are extracted. The sequence model (AR, BL) fits the EEG data, extracts the fitted model parameters as the characteristic parameters of the EEG signal, and then undergoes dimensionality reduction processing, directly and simply using a single structure such as support vector machine, neural network, etc. leaning machine. [0003] And there are various deficiencies in the above method: [0004] For the estimation of AR or BL model order, the AIC criterion or empirical estimation is commonly used at present. Although the AIC criterion has many ...

AI Technical Summary

Problems solved by technology

[0006] When the recognition rate is high, the existing technology has high requirements for the collection of EEG signals. For example, in the 2015 paper "Recognition Research Based on ERP Induced by Fawell Paradigm", the highest recognition rate can reach 98%. During the power-up, the subjects need to focus on multiple characters on the screen, and silently count the number of times the target characters flash

[0007] The existing method is closed-set verification when testing the recognition rate. The existing method has a high accuracy rate when testing the closed-set verification, but for the open-set verification, either the error rate is too large, or it is not applicable at all.

Images