基于Android系统的心电智能诊断终端算法设计与软件实现

发布时间：2018-03-07 20:58

  本文选题：Android　切入点：心电图　出处：《杭州电子科技大学》2017年硕士论文　论文类型：学位论文

【摘要】：心血管疾病突发性高,且致死致残率也极高,每年死于心脏病突发的人数占了死亡人数的1/3。突发性疾病最好的控制方法是长期监测,做到早期预防。患者若选择去医院做长期的检查,不仅就医过程繁琐,而且成本高,普通家庭难以承受。目前市场已有的家用式心电监测仪,则存在体积庞大,不能进行本地诊断等不足之处。而基于移动平台的心电监测系统,不仅降低了设备的成本,缩小了设备体积,而且能实现本地的心电分析以及远程信息通信,将成为移动医疗产品的设计趋势,因此文中选择在Android系统上实现心电智能算法和终端软件的开发。根据课题需求,首先设计整个系统的方案,分析本次开发的主要内容和需要解决的问题。文中具体介绍了心电智能检测算法的设计和终端软件的开发过程。心电智能检测算法主要包括信号预处理、特征提取和分类,首先采用小波变换结合形态学算法对信号进行预处理,去除噪声干扰,得到相对纯净的信号。然后通过K-means聚类算法提取QRS波群等特征参数,根据这些参数建立正常窦性心律和心律异常的正样本和负样本,最后结合极限学习机(Extreme Learning Machine,ELM)分类器对样本进行训练和匹配。文中以MIT-BIH心律异常数据库中的数据作为分析对象,实验结果证明文中提出的算法能准确诊断出室性早博(Premature Ventricular Contraction,PVC)和房性早搏(Atrial Premature Contraction,APC)。最终室性早博的阳性检测率P+达到94.20%,检测灵敏度Se达到96.30%,房性早搏的阳性检测率P+达到98.02%,检测灵敏度Se达到99%。算法经过测试验证后,植入Android客户端实现实时分析处理。Android客户端软件的功能设计包括蓝牙接收、实时绘图、用户管理和界面设计等。客户端将分析得到的诊断结果,通过互联网上传至Web服务器。Web服务器实现实时响应客户端的请求,将客户端上传的数据存入数据库,或是读取数据发送到客户端,实现数据的管理维护。本系统最后经测试,操作简单,运行稳定,可扩展性好,后续可以按需扩展到网络互传,远程诊断,这对心血管疾病的防治有积极的意义。
[Abstract]:Cardiovascular disease is sudden, and the rate of death and disability is extremely high. One third of the deaths from heart attacks occur every year. The best way to control sudden disease is to monitor it for a long time. To achieve early prevention. If patients choose to go to hospital for long-term examination, not only the process of seeking medical treatment is cumbersome, but also the cost is high, and it is difficult for ordinary families to bear it. At present, there is a huge volume of household ECG monitors that are available in the market. The ECG monitoring system based on mobile platform not only reduces the cost and volume of equipment, but also realizes local ECG analysis and remote information communication. It will become the design trend of mobile medical products, so we choose to develop ECG intelligent algorithm and terminal software on Android system. According to the demand of the subject, we first design the scheme of the whole system. This paper introduces the design of ECG intelligent detection algorithm and the development process of terminal software. ECG intelligent detection algorithm mainly includes signal preprocessing, feature extraction and classification. Firstly, wavelet transform combined with morphological algorithm is used to pre-process the signal to remove noise interference and get the relatively pure signal. Then K-means clustering algorithm is used to extract the characteristic parameters such as QRS wave group. According to these parameters, positive and negative samples of normal sinus rhythm and arrhythmia were established. Finally, the samples were trained and matched with extreme Learning machine classifier. The experimental results show that the proposed algorithm can accurately diagnose premature Ventricular PVCs and atrial premature beats. The positive detection rate of ventricular premature beats (P = 94.20), the sensitivity of se (96.30%) and the positive rate of atrial premature beats (P. The detection sensitivity is up to 99%. The algorithm has been tested and verified. The functional design of Android client software includes Bluetooth receiving, real-time drawing, user management and interface design. Through the Internet upload to the Web server. The web server can respond to the request of the client in real time, save the data uploaded by the client into the database, or read the data to the client to realize the management and maintenance of the data. Finally, the system is tested. The operation is simple, the operation is stable, and the expansibility is good. The follow-up can be extended to network transmission and remote diagnosis as needed, which has positive significance for the prevention and treatment of cardiovascular diseases.
【学位授予单位】：杭州电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP311.52;TP316
，

本文编号：1580925