摘要
心磁图是根据人体心脏微弱磁场测量信号计算得到的医学图像。它反映了人体心脏的电活动,可以给医生提供诊断心脏疾病的信息。为了提高心磁图成像精度,通常需要对心磁检测信号进行插值预处理。本文提出了一种基于RBF神经网络模型的心磁数据插值方法,数值仿真的结果证明,RBF神经网络插值方法比线性插值、BP神经网络插值的精度高,接近三次样条插值的结果。
Magnetocardiogram（MCG） is a kind of medical image based on measuring human heart. It reflects the electromagnetic activity in human heart, which can provide the the magnetic field of diagnosis information of heart diseases. In order to raise the precision of MCG image, the MCGs are obtained through interpolation pre -processing. A new method to solve the interpolation problem in MCG-RBF（ Radial Basis Function） ANN（ Artificial Neural Network） interpolation is presented in this article. The results of numerical value simulation prove that RBF ANN interpolation is better than linear interpolation and BP ANN interpolation. Its result is close to the cubic spline interpolation.
出处
《现代科学仪器》
2007年第2期43-45,共3页
Modern Scientific Instruments
基金
上海市科学发展基金资助(项目编号:054407061)
关键词
信号处理
RBF网络
心脏磁场
插值
Signal procession
RBF ANN
bioelectromagnetism
interpolation
