基于隐马尔可夫模型和符号条件熵的异常生理信号研究

发布时间：2018-06-30 09:46

本文选题：癫痫脑电 + 心电信号　；参考：《南京邮电大学》2016年硕士论文

【摘要】：异常生理电信号的收集和分析对于评估病患人体机能的状态具有重要的意义,并且能够帮助医生更好地进行疾病的诊断和治疗。本文在前人工作的基础上,从实际应用和理论分析两个角度分别对异常脑电信号和异常心电信号进行研究,具体的工作有如下几个部分:(1)基于隐马尔可夫模型的癫痫脑电信号识别算法癫痫脑电的采集需要医生长期观测患者的脑电图,并且根据经验做出判断。自动化的癫痫脑电识别方法可以减轻医生工作量,减少主观因素的影响。本文使用AR模型对脑电信号进行特征提取,隐马尔可夫模型(Hidden Markov Model,HMM)作为正常脑电和癫痫脑电的分类工具,可以准确区分正常脑电和癫痫脑电。仿真结果表明,癫痫脑电的识别率可达90%。使用隐马尔可夫模型建模的方法对正常脑电信号和癫痫脑电的识别率较高,算法可以达到癫痫脑电自动识别的要求。(2)基于符号条件熵的心电信号冗余信息分析论文从混沌理论普适性的角度出发,提出了一种使用符号条件熵的冗余信息计算方法,分析正常ECG和心肌梗塞异常ECG中含有的冗余信息,数值计算证明心肌梗塞异常ECG中含有的冗余信息相比于正常ECG来说更大,这表明冗余信息可以作为区分心电信号是否异常的一个指标。(3)基于Java平台的生理电信号分析系统的设计与实现论文将上述两种脑电、心电生理电信号的分析算法通过java语言进行了编程实现。首先,分析系统可以对癫痫异常脑电信号和正常脑电信号进行区分,并且显示出识别率;其次,分析系统可以计算输入的心电信号的冗余信息。该系统在脑机接口设备的开发和临床心功能评估中有一定的参考价值和辅助作用。
[Abstract]:The collection and analysis of abnormal physiological signals are of great significance in evaluating the state of the patient's human function and can help doctors to better diagnose and treat diseases. In this paper, based on the previous work, the abnormal EEG and abnormal ECG are studied from the perspective of practical application and theoretical analysis. The specific work is as follows: (1) EEEG recognition algorithm based on Hidden Markov Model (hmm) requires doctors to observe EEG of patients for a long time and make a judgment based on experience. Automatic EEG recognition method can reduce the workload of doctors and reduce the influence of subjective factors. In this paper, AR model is used to extract the features of EEG. Hidden Markov Model (hmm) is used as a classification tool for normal EEG and epileptic EEG, which can accurately distinguish normal EEG from epileptic EEG. The simulation results show that the recognition rate of epileptic EEG can reach 90%. The recognition rate of normal EEG and epileptic EEG is higher by using hidden Markov model. The algorithm can meet the requirements of EEG automatic recognition. (2) based on symbol conditional entropy, the redundant information analysis of ECG signal is presented. From the point of view of the universality of chaos theory, a method of computing redundant information using symbol conditional entropy is proposed. The redundant information contained in normal ECG and abnormal ECG of myocardial infarction is analyzed. Numerical calculation shows that the redundant information in abnormal ECG of myocardial infarction is greater than that in normal ECG. This indicates that redundant information can be used as an index to distinguish whether ECG signal is abnormal. (3) the design and implementation of physiological signal analysis system based on Java platform. The algorithm of electrophysiological signal analysis is programmed by java language. Firstly, the analysis system can distinguish the abnormal EEG signal from the normal EEG signal and show the recognition rate. Secondly, the system can calculate the redundant information of the input ECG signal. The system has certain reference value and auxiliary function in the development of brain-computer interface equipment and the evaluation of clinical cardiac function.
【学位授予单位】：南京邮电大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：R318;TN911.7

【相似文献】