基于一种新的融合特征的癫痫性发作自动检测方法研究

发布时间：2018-04-27 02:26

本文选题：癫痫 + 改进Hjorth参数　；参考：《西北大学》2017年硕士论文

【摘要】：癫痫是最常见的大脑神经紊乱疾病之一,因其发作的突发性和反复性,对患者的生理和心理都造成巨大伤害,严重危害人们的正常生活。传统的癫痫检测主要通过有经验的临床医生对脑电图进行视觉检查来进行诊断,但是海量的脑电数据使得传统的检测方法十分耗时,而且主观性强。于是,癫痫性发作的自动检测成为近年来的一个热门问题。而实现自动检测的关键问题则在于设计有效的特征提取方法。基于此,本论文主要对特征提取方法进行研究,提出一种新的癫痫脑电融合特征提取方法,并结合超限学习机与支撑向量机完成自动检测。具体的工作安排如下:第一章系统论述了癫痫性发作的自动检测的研究背景、检测流程以及国内外的研究现状;第二章主要介绍了脑电信号的相关知识和癫痫性发作自动检测中常用的特征提取方法及分类器;第三章基于Hjorth参数和样本熵首先分别提出了改进的Hjorth参数特征和二阶差分样本熵,其次将二者结合提出一种新的融合特征提取方法;第四章将本文提出的新的融合特征应用于德国波恩大学癫痫疾病研究中心的公开数据集中,通过数值实验验证本文所提方法的可行性与有效性。
[Abstract]:Epilepsy is one of the most common neurologic disorders of the brain. Because of its sudden and recurrent seizures, it causes great harm to the patients' physiology and psychology, and seriously endangers people's normal life. The traditional epilepsy detection is mainly through the experienced clinicians to make the diagnosis of EEG, but the massive EEG data make the traditional detection methods very time-consuming and subjective. Therefore, the automatic detection of epileptic seizures has become a hot issue in recent years. The key problem of automatic detection is to design an effective feature extraction method. Based on this, this paper mainly studies the feature extraction method, proposes a new feature extraction method of epileptic EEG fusion, and combines the out-of-limits learning machine and support vector machine to complete the automatic detection. The specific work arrangements are as follows: the first chapter systematically discusses the background of the automatic detection of epileptic seizures, detection process and domestic and foreign research status; The second chapter mainly introduces the related knowledge of EEG and the methods of feature extraction and classifier used in the automatic detection of epileptic seizures. In chapter 3, based on Hjorth parameters and sample entropy, the improved Hjorth parameter feature and the second order differential sample entropy are proposed, and then a new fusion feature extraction method is proposed. In chapter 4, the new fusion features proposed in this paper are applied to the open data set of the Research Center for Epilepsy at the University of Bonn, Germany. The feasibility and effectiveness of the proposed method are verified by numerical experiments.
【学位授予单位】：西北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：R742.1;TN911.6

【参考文献】