Research On Location And Detection Method Of Malignant Heart Disease Lesion Based On Multi-lead ECG Signal

Cardiovascular diseases(CVDs)are one of the most important public health problems worldwide,and their morbidity and mortality account for a large proportion of the world.CVDs are a group of diseases that affect the heart and blood vessels,including coronary heart disease,heart failure,stroke,and peripheral artery disease,among others.Among cardiovascular diseases,myocardial infarction and malignant arrhythmia are two common types of malignant heart disease,and the mortality rate of patients remains high.If these malignant heart diseases can be predicted early,the complications and mortality caused by these malignant heart diseases will be greatly reduced.At present,the detection of myocardial infarction has received extensive attention and in-depth research by many scholars at home and abroad,while the related research on the detection of malignant arrhythmia is relatively limited,mainly due to the limitation of the amount of available data.There is room for further improvement in the performance and efficiency of existing myocardial infarction and malignant arrhythmia detection algorithms.Therefore,this paper uses machine learning technology to focus on the research on the above-mentioned malignant heart disease focus location and detection method based on multi-lead ECG signals.The main research content of this paper is as follows:(1)In myocardial infarction detection and localization research.Inspired by the Universitat Politecnica de Catalunya-Universidad Central de Venezuela(UPC-UCV),this paper proposes for the first time the application of UPC-UCV to detect myocardial infarction disease.In order to make UPC-UCV suitable for myocardial infarction detection,we improved the moving step size and order of its filter,and named the improved UPC-UCV as UPC-UCV-MI.The experimental results showed that under the intra-patient scheme,the detection accuracy,recal,specificity,precision and F1 score of UPC-UCV-MI for myocardial infarction reached 99.96%,99.95%,99.95%,99.97%and 99.96%,respectively;Under the inter-patient scheme,the above indicators were95.69%,97.74%,93.62%,94.02% and 95.32%,respectively.Compared with other related methods,our method achieved better results under both scenarios.In terms of myocardial infarction localization research,this paper proposes a hybrid network of UPC-UCV-MI combined with Transformer Encoder(named MI-Hybrid)to locate the site of myocardial infarction.The ECG signals of 9 relevant leads were selected as the input of MI-Hybrid according to the prior knowledge of medicine.The experimental results show that MI-Hybrid can identify 5 kinds of myocardial infarction sites.Under the intra-patient scheme,the accuracy,recall,specificity,accuracy,and F1 score of MIHybrid in myocardial infarction location reached 99.67%,99.66%,99.93%,99.67%,and 99.66%,respectively;Under the inter-patient scheme,the above indicators were73.33%,72.56%,94.63%,74.99% and 73.55%,respectively.Although the effect of the inter-patient scheme is not as good as that of the intra-patient scheme,MI-Hybrid still outperforms other related methods with the characteristics of high efficiency and accuracy.(2)Research on intelligent identification of common malignant arrhythmias.For malignant arrhythmias,including Asystole,Extreme Bradycardia,Extreme Tachycardia,Ventricular Tachycardia,and Ventricular flutter/fibrillation and other five common malignant arrhythmias,this study uses biomedical signal processing technology to construct intelligent identification method.Experimental results show that the method proposed in this paper can detect these malignant arrhythmias more accurately.Specifically,the accuracy,recal,and specificity of our method for detecting Asystole,Extreme Bradycardia,Extreme Tachycardia,Ventricular Tachycardia,and Ventricular flutter/fibrillation reached(90.16%,81.82%,92%),(86.51%,97.83%,74.42%),(97.86%,99.24%,77.78%),(82.7%,83.15%,82.54%),(98.28%,100%,98.08%).This shows that our method has a strong ability to identify malignant arrhythmia,and provides an effective method for the intelligent identification of malignant arrhythmia.