心律失常是常见的心血管疾病之一,目前很多方法通过计算机辅助系统对心电图进行分析以识别心律失常,但由于大多数心律失常数据样本较少,计算机辅助系统识别心律失常效果不佳.本文提出了一种基于混合时频域分析特征提取的卷积神经网络方...心律失常是常见的心血管疾病之一,目前很多方法通过计算机辅助系统对心电图进行分析以识别心律失常,但由于大多数心律失常数据样本较少,计算机辅助系统识别心律失常效果不佳.本文提出了一种基于混合时频域分析特征提取的卷积神经网络方法,该方法提取心电图的RR间期时域特征、希尔伯特-黄变换提取的频域特征和连续小波变换提取的时频域联合特征,经过特征融合后输入卷积神经网络训练分类模型,并采用Focal Loss作为网路的损失函数,实现对心律失常的分类.本文使用MIT-BIH(Massachusetts Institute of Technology-Boston’s Beth Israel Hospital)心律失常数据库验证本文提出方法对4类心电数据分类的结果,实验结果表明,与现有的分类算法相比,本文所提出的混合时频域特征方法能有效提升心律失常分类的准确性.展开更多
. In this paper, an approach to the design of hemispherical breast RF coil array is proposed. The target field method is applied to find the current density distribution on the hemisphere surface, which induces a homo.... In this paper, an approach to the design of hemispherical breast RF coil array is proposed. The target field method is applied to find the current density distribution on the hemisphere surface, which induces a homogeneous magnetic field in a hemispherical volume. The components of current density are expanded into Fourier series, and the highly ill-conditioned character of the linear equation related to Fourier coefficients is solved using the Tikhonov regularization method with a minimum curvature penalty function. The winding pattern was acquired using the stream function technique. The results indicate that a simple winding pattern with homogeneous magnetic field can be obtained through manually selecting the penalty factor.展开更多
Atrial fibrillation(AF)is a common cardiac rhythm disturbance that increases in prevalence with advancing age.To get the real time location of the related key lesions is of great significance to treat and abolish AF.T...Atrial fibrillation(AF)is a common cardiac rhythm disturbance that increases in prevalence with advancing age.To get the real time location of the related key lesions is of great significance to treat and abolish AF.Two-dimensional finite element method was adopted to simulate the lead field.Based on the acoustoelectric effect and the reciprocal theory,the source density of the current field was calculated,thus determining the precise location of the current source and sink.Results show that the simulated positioning precision of current source and sink is within submillimetre only using a pair of recording electrode.This work implies that it is promising to locate the key lesions from messy fibrillation using the acoustoelectric effect method.展开更多
文摘心律失常是常见的心血管疾病之一,目前很多方法通过计算机辅助系统对心电图进行分析以识别心律失常,但由于大多数心律失常数据样本较少,计算机辅助系统识别心律失常效果不佳.本文提出了一种基于混合时频域分析特征提取的卷积神经网络方法,该方法提取心电图的RR间期时域特征、希尔伯特-黄变换提取的频域特征和连续小波变换提取的时频域联合特征,经过特征融合后输入卷积神经网络训练分类模型,并采用Focal Loss作为网路的损失函数,实现对心律失常的分类.本文使用MIT-BIH(Massachusetts Institute of Technology-Boston’s Beth Israel Hospital)心律失常数据库验证本文提出方法对4类心电数据分类的结果,实验结果表明,与现有的分类算法相比,本文所提出的混合时频域特征方法能有效提升心律失常分类的准确性.
文摘. In this paper, an approach to the design of hemispherical breast RF coil array is proposed. The target field method is applied to find the current density distribution on the hemisphere surface, which induces a homogeneous magnetic field in a hemispherical volume. The components of current density are expanded into Fourier series, and the highly ill-conditioned character of the linear equation related to Fourier coefficients is solved using the Tikhonov regularization method with a minimum curvature penalty function. The winding pattern was acquired using the stream function technique. The results indicate that a simple winding pattern with homogeneous magnetic field can be obtained through manually selecting the penalty factor.
基金Science Technology Department Program of Zhejiang Provincegrant number:2014C37074
文摘Atrial fibrillation(AF)is a common cardiac rhythm disturbance that increases in prevalence with advancing age.To get the real time location of the related key lesions is of great significance to treat and abolish AF.Two-dimensional finite element method was adopted to simulate the lead field.Based on the acoustoelectric effect and the reciprocal theory,the source density of the current field was calculated,thus determining the precise location of the current source and sink.Results show that the simulated positioning precision of current source and sink is within submillimetre only using a pair of recording electrode.This work implies that it is promising to locate the key lesions from messy fibrillation using the acoustoelectric effect method.
