Holter系统的技术要求及神经网络方法

为了长时间、准确的记录分分析体表ＥＣＧ，从各个方面满足基础研究、临床、康复以及特殊生理需要，对Ｈｏｌｔｅｒ系统提出了愈来愈高的要求。作者从技术实现的角度讨论了Ｈｏｌｔｅｒ的主要技术要求，包括记录部分和分析部分；还介绍了几种用神经网络模型满足Ｈｏｌｔｅｒ系统技术要求的原理，包括用多层感知器实现数据压缩及分类，用高阶神经网络实现ＥＣＧ的分类，以及用ＡＲＴ模型识别ＥＣＧ中的Ｐ波与Ｔ波。

Technical Requirements for Holter System and Method of Neural Network

It was found recently that the need of ECG in several fields,such as clinic,fundamental research,rehabilitation and special applications in physiology can be met by long-time recording and analyzing of ECG.So the leval of technical requirements for Holter system becomes higher and higher.In this paper, we discuss main requirements for Holter system,including recording part and analyzing part,in order to meet those requirements.We also introduce the principle of using several neural network models to meet technical requirements for Holter system,including:the use of multilayer perceptron model in data compression and classification;the use of high order neural network model in classifying ECG,and the use of ART model in recognizing T-wave and P-wave of ECG.