基于深度学习的ECG疾病识别研究材料资料.docxVIP

摘 要 人体中最重要的器官就是心脏，它为流淌在全身的血液提供动力，假如心脏出现了某些问题，即产生了心血管疾病症状，那么血液循环的动力系统就出现了问题，就会影响我们的身体健康，严重的甚至会引发死亡。例如心血管疾病之一的心肌梗塞，它是一种常见的病症，经济的快速发展使得人们的生活水平快速提升，相对应的工作等压力也日益剧增，致使很多人处于亚健康的身体状态，长此以往就极有可能会引发心绞痛——心肌梗塞的前期症状，进而可能发展成为心肌梗塞，更严重的引发死亡。因此能够及时、准确的诊断出病症，做出更加有效的病理控制，对于减少心肌梗塞的死亡率至关重要。 心血管类疾病的诊断大多数都是由临床医生根据心电图进行，但是这样进行诊断的效果还不是很理想，疾病进一步的确诊还需要由专业医师进行。所以，在人工智能、大数据时代，将计算机算法的快速反应以及通过改进的计算机算法相结合进行疾病的诊断的步伐应该加快。 本文从Physiobank数据库中PTB心电数据库中获得心肌梗塞患者以及健康人群的ECG数据信号，通过基于小波变化的去噪方式对提取的信号进行预处理，使用长短期记忆法（即LSTM）对预处理之后的ECG信号进行特征提取，建立心肌梗塞的支持向量机模型（SVM）以及线性回归模型，分析心肌梗塞模型特征，由此归纳总结心肌梗塞患者的ECG信号特有标志，从而为更加高效、准确的对疾病做出诊断提供可能，为患者争取更多的治疗机会。 关键词：心电信号；心肌梗塞；小波去噪；提取；识别 Abstract The?most?important?organ?in?the?human?body?is?the?heart,?which?provides?power?for?the?flow?of?blood?throughout?the?body.?If?there?is?something?wrong?with?the?heart,?that?is,?there?are?symptoms?of?cardiovascular?disease,?so?there?is?something?wrong?with?the?power?system?of?blood?circulation,?which?will?affect?our?health,?and?even?lead?to?death.?Such?as?cardiovascular?disease,?one?of?myocardial?infarction,?it?is?a?common?disease,?the?rapid?development?of?economy?makes?peoples?life?level?is?rapidly?increasing,?the?corresponding?work?pressure?is?also?growing,?such?as?causing?a?lot?of?people?in?the?sub-health?state?of?the?body,?in?the?long?term?it?is?associated?with?a?greater?risk?of?early?symptoms?of?angina?pectoris,?myocardial?infarction,?which?may?become?a?myocardial?infarction,?a?more?serious?cause?death.?Therefore,?timely?and?accurate?diagnosis?of?the?disease?and?more?effective?pathological?control?are?crucial?to?reduce?the?death?rate?of?myocardial?infarction. Most of the diagnosis of cardiovascular diseases is carried out by clinicians according to the electrocardiogram, but the effect of such diagnosis is not ideal, further diagnosis of the disease still needs to be carried out by professional doctors. Therefore, in the era of artificial intelligence and big data, the rapid response of computer algorithms and the combination of improved compute