Abstract: In order to automatically remove the ocular artifact from Electroencephalogram(EEG) data, this paper presents an effective and robust algorithm. Blind Source Separation(BSS) algorithm is used to separate real EEG signal, and a universal approach together with sample entropy is employed to identify the artifact components. The artifact is eliminated through reconstruction of EEG data. Simulation results are compared with a robust method based on fractal dimension and show the proposed algorithm can eliminate ocular artifact accurately and do little harm to useful EEG signals that for real EEG data of different lengths. Meanwhile, the proposed method does not need any reference channel, and is fully automated, which is suitable for real-time applications.

Key words: Electroencephalogram(EEG) signal, Second-order Blind Identification(SOBI), ocular artifact, sample entropy, fractal dimension

摘要： 为实现眼电伪迹的自动去除，提高算法的有效性和稳健性，提出一种眼电伪迹自动去除算法。采用样本熵和一种通用的伪迹判决方法对眼电伪迹进行自动识别，通过脑电信号的重构实现眼电伪迹的去除。实验结果表明，对于不同长度的真实脑电信号，该算法均能准确地去除眼电伪迹，较好地保留其他的脑电信号成分，且可以完全自动地去除眼电伪迹，适用于实时场合。

关键词: 脑电信号, 二阶盲辨识, 眼电伪迹, 样本熵, 分形维数

CLC Number: 

• TN911.72