癫痫脑电棘波检测.doc

内容介绍

癫痫脑电棘波检测

1.6万字
自己原创的毕业论文，仅在本站独家出售，重复率低，推荐下载使用

摘要癫痫，是一种生活中篅@＜哪圆考膊。现氐匚：α宋颐堑纳硖寰竦慕】怠Ｋ⒆鞯氖焙虮硐治竽云げ慵捌げ阆禄抑释磐蝗还鹊闹馗葱苑诺纾贾虏∪松硖宓某榇ぃ馐兜母谋湟约熬竦囊斐５鹊取Ｋ运担颐窃跹择拆锬缘缧藕沤蟹浅Ｗ既返姆治雠卸暇拖缘煤苤匾簿哂兄匾庖濉Ｄ缘缤�(Electroencephalogram，EEG)产生于大脑神经细胞群在生理过程中自发或诱发的电活动，表现为神经细胞群在头皮表面或者颅内引起的电位信号的变化，它反映了大脑生物电的节律活动规律。它包含了大量的生理信息，疾病信息，是研究癫痫发作特征的重要工具。
本文是基于形态成分分析脑电棘波，脑电棘波检测选用离散余弦变换作为提取EEG的背景信号的字典，选用Curvelet变换作为提取癫痫脑电棘波信号的字典。首先我们要对癫痫脑电信号数据样本进行特征提取，其提取出的特征向量送入支持向量机，进行训练，得到一个训练模型，利用此模型，进行混合脑电信号的癫痫分析判断。对于一组含有癫痫脑电信号和正常脑电信号的数据，用训练出的模型进行分析判断，分出其中癫痫脑电信号和正常脑电信号，与已知结果进行比较，由此得到模型判断的准确率，由此得到离散余弦变换，Curvelet变换分析的准确率。
关键词癫痫脑电棘波特征提取 Curvelet变换离散余弦变换支持向量机分类

Epileptic EEG Spike Detection
Abstract Epilepsy is a common chronic brain diseases，which is serious harm to human health.It is time to attack the performance of the cerebral cortex and subcortical gray matter sudden excessive repetitive discharge，resulting in the patient's body convulsions，changes and abnormal mental awareness，etc.So how to analysis epileptic EEG accurately has important practical significance. EEG (Electroencephalogram，EEG) is a potential signal specific parts of the brain nerve cell populations in spontaneous or induced electrical activity in physiological processes or intracranial specific parts of the surface of the scalp caused，it reflects the rhythm of bioelectric brain activity patterns. It contains a large number of physiological and disease information which is an important tool to study the characteristics of epileptic seizures.
This article is based on morphological analysis of EEG spikes ingredients.EEG Spike Detection chooses discrete cosine transform as the extraction of the background EEG signal dictionary，and choose Curvelet transform component analysis as a form of a dictionary to extract the signal spikes.Samples for feature extraction，which extracts the feature vector into the support vector machines，training，get a training model，the use of this model for the epileptic EEG analysis and judgment.That is，a set of data containing epileptic EEG and normal EEG，with training models to analyze the judgment，in which the separation of epileptic EEG and normal EEG，the results were compared with the known， whereby the model to determine accuracy，thereby obtaining a discrete cosine transform，Curvelet transform accuracy of the analysis.
Keywords Epileptic EEG Spikes feature extraction Curvelet transform discrete cosine transform Support vector machine classification

目录
第一章绪论 1
1.1 癫痫脑电研究背景 1
1.2 癫痫脑电研究目的及意义 2
1.3 本文研究内容： 3
第二章癫痫脑电研究相关方法概述 4
2.1 癫痫脑电棘波检测方法 4
2.2 形态成分分析的研究现状 7
第三章癫痫脑电形态成分分析原理 9
3.1 形态成分分析 9
3.1.1信号的稀疏表示 9
3.1.2形态成分分析的基本概念以及算法 10
3.1.3快速隐式变换 10
3.1.4形态成分分解 11
3.2 字典的选择 12
3.3 离散余弦变换 13
3.3.1 离散余弦变换(DCT)简介 13
3.3.2 离散余弦变换(DCT)变换式 13
3.4 Curvelet变换 14
3.4.1 Curvelet变换简介 14
3.4.2 Curvelet变换的基本实验步骤 15
3.5 支持向量机 16
3.5.1 支持向量机概述 16
3.5.2 核函数 17
3.5.3 线性分类器 18
3.6 癫痫脑电分析过程 21
3.6.1 实验数据 21
3.6.2 癫痫脑电分析实验流程图 21
3.6.3 癫痫脑电信号特征提取 22
3.6.4 支持向量机分类及算法 24
第四章结果 27
4.1 基于Curvelet变换和离散余弦变换的癫痫脑电棘波检测 27
4.1.1仿真结果与比较 27
第五章总结 33
致谢 34
参考文献 35