Attention-Based Residual Dense Shrinkage Network for ECG Denoising

Electrocardiogram(ECG)signal is one of the noninvasive physiological measurement techniques commonly usedin cardiac diagnosis.However,in real scenarios,the ECGsignal is susceptible to various noise erosion,which affectsthe subsequent pathological analysis.Therefore,the effective removal of the noise from ECG signals has becomea top priority in cardiac diagnostic research.Aiming at the problem of incomplete signal shape retention andlow signal-to-noise ratio(SNR)after denoising,a novel ECG denoising network,named attention-based residualdense shrinkage network(ARDSN),is proposed in this paper.Firstly,the shallow ECG characteristics are extractedby a shallow feature extraction network(SFEN).Then,the residual dense shrinkage attention block(RDSAB)isused for adaptive noise suppression.Finally,feature fusion representation(FFR)is performed on the hierarchicalfeatures extracted by a series of RDSABs to reconstruct the de-noised ECG signal.Experiments on the MIT-BIHarrhythmia database and MIT-BIH noise stress test database indicate that the proposed scheme can effectively resistthe interference of different sources of noise on the ECG signal.

Preliminary abnormal electrocardiogram segment screening method for Holter data based on long short-term memory networks

Holter usually monitors electrocardiogram(ECG)signals for more than 24 hours to capture short-lived cardiac abnormalities.In view of the large amount of Holter data and the fact that the normal part accounts for the majority,it is reasonable to design an algorithm that can automatically eliminate normal data segments as much as possible without missing any abnormal data segments,and then take the left segments to the doctors or the computer programs for further diagnosis.In this paper,we propose a preliminary abnormal segment screening method for Holter data.Based on long short-term memory(LSTM)networks,the prediction model is established and trained with the normal data of a monitored object.Then,on the basis of kernel density estimation,we learn the distribution law of prediction errors after applying the trained LSTM model to the regular data.Based on these,the preliminary abnormal ECG segment screening analysis is carried out without R wave detection.Experiments on the MIT-BIH arrhythmia database show that,under the condition of ensuring that no abnormal point is missed,53.89% of normal segments can be effectively obviated.This work can greatly reduce the workload of subsequent further processing.

An Approach to High-Order Cumulants Used to Detect Multifrequency Signalsin Telephone Networks

In accordance with the detecting process of multi-frequency signals between the offices in telephone networks, and in contrast with the autocorrelation method used to handle the multi-frequency signals, a fast, inexpensive and unbiased of cumulants estimation method is adopted in detecting signals. This detecting method is better for resisting noise performance and more practical than the autocorrelation method.

Detection of healthy and pathological heartbeat dynamics in ECG signals using multivariate recurrence networks with multiple scale factors

The electrocardiogram(ECG)is one of the physiological signals applied in medical clinics to determine health status.The physiological complexity of the cardiac system is related to age,disease,etc.For the investigation of the effects of age and cardiovascular disease on the cardiac system,we then construct multivariate recurrence networks with multiple scale factors from multivariate time series.We propose a new concept of cross-clustering coefficient entropy to construct a weighted network,and calculate the average weighted path length and the graph energy of the weighted network to quantitatively probe the topological properties.The obtained results suggest that these two network measures show distinct changes between different subjects.This is because,with aging or cardiovascular disease,a reduction in the conductivity or structural changes in the myocardium of the heart contributes to a reduction in the complexity of the cardiac system.Consequently,the complexity of the cardiac system is reduced.After that,the support vector machine(SVM)classifier is adopted to evaluate the performance of the proposed approach.Accuracy of 94.1%and 95.58%between healthy and myocardial infarction is achieved on two datasets.Therefore,this method can be adopted for the development of a noninvasive and low-cost clinical prognostic system to identify heart-related diseases and detect hidden state changes in the cardiac system.

Optimization of Electrocardiogram Classification Using Dipper Throated Algorithm and Differential Evolution

Electrocardiogram(ECG)signal is a measure of the heart’s electrical activity.Recently,ECG detection and classification have benefited from the use of computer-aided systems by cardiologists.The goal of this paper is to improve the accuracy of ECG classification by combining the Dipper Throated Optimization(DTO)and Differential Evolution Algorithm(DEA)into a unified algorithm to optimize the hyperparameters of neural network(NN)for boosting the ECG classification accuracy.In addition,we proposed a new feature selection method for selecting the significant feature that can improve the overall performance.To prove the superiority of the proposed approach,several experimentswere conducted to compare the results achieved by the proposed approach and other competing approaches.Moreover,statistical analysis is performed to study the significance and stability of the proposed approach using Wilcoxon and ANOVA tests.Experimental results confirmed the superiority and effectiveness of the proposed approach.The classification accuracy achieved by the proposed approach is(99.98%).

A Novel Radial Basis Function Neural Network Approach for ECG Signal Classification

ions in the ECG signal.The cardiologist and medical specialistfind numerous difficulties in the process of traditional approaches.The specified restrictions are eliminated in the proposed classifier.The fundamental aim of this work is tofind the R-R interval.To analyze the blockage,different approaches are implemented,which make the computation as facile with high accuracy.The information are recovered from the MIT-BIH dataset.The retrieved data contain normal and pathological ECG signals.To obtain a noiseless signal,Gaborfilter is employed and to compute the amplitude of the signal,DCT-DOST(Discrete cosine based Discrete orthogonal stock well transform)is implemented.The amplitude is computed to detect the cardiac abnormality.The R peak of the underlying ECG signal is noted and the segment length of the ECG cycle is identified.The Genetic algorithm(GA)retrieves the primary highlights and the classifier integrates the data with the chosen attributes to optimize the identification.In addition,the GA helps in performing hereditary calculations to reduce the problem of multi-target enhancement.Finally,the RBFNN(Radial basis function neural network)is applied,which diminishes the local minima present in the signal.It shows enhancement in characterizing the ordinary and anomalous ECG signals.

一种改进U-Net网络的心电图分类算法研究

基于CPSC-2018十二导联数据,提出了一种U-Net网络和注意力机制结合的心电图分类算法。首先,针对数据集数据长度长短不一的问题,对数据进行等长处理和归一化处理。然后,利用U-Net网络中跳层连接和编码解码方式,对预处理后较长的数据进行处理。在U-Net网络解码的最后一层加入注意力机制对抗噪声,提升模型的有效信息关注度和准确性。最后,利用CPSC-2018数据集进行验证。实验结果表明:所提模型能够取得较好的分类效果,识别房颤(AF)和右束支传导阻滞(RBBB)心律失常的精准率、召回率、F1值都可以达到90%以上,平均F1值可以达到82.5%。

基于多特征分支卷积神经网络的心电图分类算法

我国心血管疾病发病率、病死率呈逐年上升趋势。但由于心电图数据规模大且繁杂,临床医护人员在心电图筛查时,工作负担大且容易出现误诊或者漏诊的情况。基于此,利用CPSC-201812导联数据,提出了一种基于多特征分支卷积神经网络的多导联心电信号的智能分类与分析。首先,将CPSC-201812导联数据分为9个类别,基于12导联推导出8导联心电信号并分别提取局部特征。然后,通过双向GRU编码和注意力机制计算出不同类别的注意力权重向量,并将特征信息串联融合成特征向量,从而实现多导联心电图分类。实验结果表明:在验证集上取得了较好的分类效果,正常类别的F1值达到81.2%,平均F1值达到84.2%。特别地,在识别房颤(AF)和右束支传导阻滞(RBBB)这两类别心律失常时F1值分别达到95.1%和93.1%。

基于CNN‑LSTM‑SE的心电图分类算法研究

心血管疾病是我国死亡率较高的疾病之一,通过观察心电图来判断心电信号是否出现异常能够对心血管疾病进行预防和筛查。由于心电图数据规模大且繁杂,临床医护人员在心电图筛查时,工作负担大且容易出现误诊或漏诊的情况。为了提高心电图的筛查效率、减少医护人员的压力,提出了一种基于卷积神经网络、长短期记忆神经网络和SE网络的心电图分类算法模型(CNN-LSTM-SE),该模型将心电图分成5种不同的类别。主要研究内容包括:选用MIT-BIH心律失常数据集作为心电信号的数据来源,使用巴特沃斯带通滤波器对心电信号进行去噪处理,通过Z-score方法对心电信号进行标准化处理,利用独热编码方法对心电信号标签进行编码,最后使用处理后的心电数据对所提算法模型进行训练和测试。实验结果表明:所提模型相较于其它模型,能够有效提高心电图分类的准确性,在实验数据集上的分类准确率达到99.1%。

室上性心动过速机制的智能分类模型:基于十二导联穿戴式心电设备

目的基于十二导联穿戴式心电设备,探索室上性心动过速(SVT)机制鉴别的智能分类模型。方法选取356份SVT的穿戴式心电图,通过五折交叉验证的方式随机分为训练集、验证集建立智能分类模型,选取2021年10月~2023年3月诊断为SVT并行电生理检查及射频消融术的患者共101例作为测试集。对比心动过速诱发前后的心电图参数改变,基于多尺度深度神经网络,并加入窦性心律对比图增强训练,建立SVT机制分类的智能分类模型并验证诊断效能。进一步提取II,III,V1三导联心电信号建立分类模型,并对比其与十二导联智能分类模型的效能。结果101例测试集中68例为房室结折返性心动过速,33例为房室折返性心动过速。预训练模型在验证集中识别房室结折返性心动过速的最高精确率-召回率曲线下面积达到0.9492,F1评分为0.8195。最终II导联,III导联,V1导联,三导联与十二导联智能分类模型于测试集中的总F1评分分别为0.5597,0.6061,0.3419,0.6003与0.6136。对比十二导联,III导联的净重新分类指数与综合判别改善指数分别为-0.029(P=0.714)与-0.005(P=0.817)。结论基于多尺度深度神经网络,初步建立了穿戴式心电图对SVT机制分类的智能分类模型,并具有一定的准确性。

基于CA-MobileNetV2的心肌梗死定位算法研究

为实现临床医疗设备快速辅助诊断心肌梗死(MI)发生的部位。在轻量化卷积神经网络MobileNetV2的基础上结合协调注意力(CA)机制设计出了一种高准确率的MI部位定位算法。从PTB数据集中筛选正常和MI病例的12导联心电图(ECG)样本,将ECG信号进行去噪处理。使用差分阈值法检测出ECG信号的R峰,根据R峰分割出心拍样本,使用心拍数据对所设计模型进行训练和测试。使用准确率、精度、灵敏度、特异性和混淆矩阵对模型的分类性能进行了评估。将训练集迭代60轮后,测试集的准确率达到了99.91%。结果表明,融合CA模块的MobileNetV2模型对于MI部位的定位具有很好的效果,有助于医疗设备实现MI的快速辅助诊断。

多导联心电图识别的稳定步长ResNet深度网络

针对经典的ResNet深度神经网络对一维多导联心电图图像进行识别、分类时,因原始图像的维度较高导致提取到的深度特征维度高,造成全连接层训练出现收敛速度慢和过拟合的问题,在ResNet的全连接层提出一种稳定步长动量训练算法,通过引入近似二阶梯度信息增强动量法的寻优能力和加速收敛速度;利用连续2次迭代的参数变化量和梯度信息自适应调整步长,构造边界函数对步长的大小进行限制,以防止步长过大或过小而影响收敛稳定性,使用动量项对参数的更新方向进行修正。在CPSC2018心电图数据集上的实验结果表明:所提算法训练的ResNet取得的F 1分数、准确率、精确度分别达到0.859、97.4%、87.9%,收敛速度和整体分类指标值优于其他相比较的方法。

基于连续小波变换和残差神经网络的房颤预测研究

心房颤动(AF)是一种最常见的心律失常类型,为了提高房颤预测的准确率和可靠性,提出了一种基于连续小波变换和残差神经网络的房颤预测方法。首先,采用软阈值小波去噪方法去除心电图信号的噪声干扰;其次,通过连续小波变换生成二维时频图;最后,使用带下采样的残差神经网络进行房颤预测。为了全面评估所提方法的性能,新建立了一个包含2160条心电图(ECG)记录的综合数据集,并在此数据集上进行了实验。实验结果表明,该方法在新数据集和公开数据集(AFPDB)上分别得到92.4%和96.1%的精确度,相较于当前的深度学习方法,实现了显著提升。

基于卷积神经网络的心律失常分类研究

心律失常是引起心肌梗塞、突发性心脏死亡等严重疾病的重要原因,常借助心电图进行早期诊断。然而,传统的心电信号分类方法有着复杂的特征提取任务,计算量大、费时费力。为此通过在经典卷积神经网络的基础上加入Dropout层,设计了一种改进的卷积神经网络模型,该模型可以进一步提升模型的泛化能力、提高准确率。其利用CNN自动提取特征,将经过预处理的心电信号直接作为模型的输入,自动识别5种不同类型的心拍。在MIT-BIH心律不齐数据库上进行实验,准确率达到99.68%,特异性为98.94%,灵敏度为99.76%。实验结果表明,相较于经典卷积神经网络,本文提出的方法能够精确、高效地识别不同类型的心律失常疾病。

基于卷积神经网络的隐匿性旁路预测模型

隐匿性旁路(CAP)是一种引起心跳突然加速、心悸和胸闷的心脏疾病。针对目前临床医师尚无法通过窦性心律心电图(ECG)对隐匿性旁路进行诊断的现状,基于临床病例建立包含隐匿性旁路患者术前窦性心律心电图及健康对照人群心电图的数据集,并提出一种以ResNet26为基线网络的利用窦性心律心电图自动识别预测隐匿性旁路患者的卷积神经网络CAPNet。创建初始模块(IB),提升模型非线性表达能力。引入非对称卷积以改进瓶颈残差模块,更好地捕捉心电特征的水平和垂直方向信息,丰富特征空间。使用注意力机制,加强模型对心电图中重点波段区域的关注。实验结果表明,CAPNet模型的预测性能优于对比的经典卷积神经网络模型,与ResNet26相比,F1值、准确率、灵敏度和精确率分别提升了2.41、0.89、4.34和0.47个百分点。上述实验结果验证了CAPNet模型的有效性与优越性。

基于互信息的多导联心电图排序方法

基于卷积神经网络的心电图(Electrocardiograph,ECG)自动分类研究从默认12导联顺序的心电图中提取特征,未考虑导联顺序对卷积网络特征提取的影响。为解决该问题,文中提出了一种基于互信息的两端递增排序方法,使用互信息衡量导联之间的相关性,并根据导联之间的相关性以及二维卷积的特点将关系密切的导联相邻排序。实验结果表明,多导联心电图排序方法在3个数据库和3个卷积网络分类模型上取得了显著效果,F1、正确率、召回率、精确率以及杰卡德系数数值分别提升了0.011、0.009、0.007、0.014和0.013,汉明损失值减低了0.002。

网络心电图诊断临床应用价值研究

目的分析网络心电图在心血管疾病临床诊断中的价值。方法选取2021年3月至2023年3月于医院接受网络心电图检查的1000例心血管疾病患者作为网络心电图组,另取2018年1月至2021年1月于医院接受常规心电图检查的1000例心血管疾病患者作为常规心电图组。比较两组的诊断效果。结果网络心电图组的检查所需时间、检查报告完成时间及就诊等待时间均短于常规心电图组,差异具有统计学意义(P<0.05)。网络心电图组的异常心电图、心律失常及心肌缺血检出率分别为65.30%、41.20%、12.00%,均高于常规心电图组的39.30%、20.50%、2.10%,差异具有统计学意义(P<0.05)。两组患者的Q-T间期、T时限、QRS时限、P-R间期及P波时限比较,差异无统计学意义(P>0.05)。网络心电图组的就医满意度为98.80%,高于常规心电图组的90.20%,差异具有统计学意义(P<0.05)。结论在心血管疾病诊断过程中,选用网络心电图检查可获得明显优于常规心电图检查的效果,其能够为患者争取更多诊治时间,从而促进患者早日康复,提升就医满意度。

基于位置注意力机制的混合神经网络心电信号分类算法

心电信号分类是医疗保健领域的重要研究内容。心电信号数据是类不平衡数据,不同类别的心律失常依赖于心电图的长期变化特征,局部变化特征及其相对位置。针对大多数方法不能较好地解决数据类不平衡,且未考虑特定波形重要性等问题,提出一种基于位置注意力机制的混合神经网络心电信号分类(DCLB)算法。首先,利用深度卷积生成对抗网络扩充数量少的类别样本,从而解决类不平衡问题;其次,利用二维卷积神经网络和双向长短期记忆网络进行特征提取,从而获得心电信号的局部变化特征和长期变化特征;然后,在每个二维卷积神经网络后嵌入位置注意力机制,从而提高关键位置特征的重要程度;最后,利用全连接网络输出分类结果。对MIT-BIH心律失常数据集中的30584个样本的实验结果表明,DCLB算法的平均准确率为98.79%,敏感性为94.21%,特异性为98.98%,阳性预测值为93.70%。该模型可以有效提取心电信号特征,适用于监测系统中心律失常疾病的诊断。

GPON技术在铁路基层生产机构的应用

随着通信技术的迅速发展,铁路通信、信息业务类型不断增加,各类业务接入需求也同步增长。文中在研究各种接入解决方案的基础上,结合GPON技术速率快、容量大、易扩容、造价低的特点,提出了GPON技术是解决铁路基层生产机构业务接入需求的最佳方案。

基于卷积注意力的单导联心电图房颤检测方法

随着可穿戴心电设备的普及,从单导联心电图中自动检测房颤的方法越来越重要。针对可穿戴心电设备采集的单导联心电图中存在噪声干扰的问题,提出一种基于卷积注意力的残差神经网络模型Resnet34-CAB。通过融合卷积注意力块(CAB),在模型复杂度少量增加的情况下,选择性地关注心电图的关键特征,自适应地抑制噪声,提高了模型的检测性能。在公开数据集上的实验结果表明,Resnet34-CAB模型优于Resnet34、Resnet34-Transformer模型,验证了融合CAB的有效性。