摘要
首先介绍了传统人工脉象识别的缺点,而后介绍了现代脉象分类的优点和当今主要特征提取方法,针对其提取特征的特点提出了利用分形理论提取脉搏信号的分数维,以此作为波形的重要特征,用于脉象分类识别。利用这些特征矢量在BP神经网络系统中对实测的脉象数据进行了分类,取得了令人满意的分类正确率。
Introduced the meaning of the pulse＇s classify and the conventional classified methods,thereafter advanced using the fractal which are distilled by fractal theory and the multiscale energy distribution special features which is based on wavelet as the characters for the classify.Then using these feature vectors for BP neural networks,acquired famous preciseness.Throw compared the two methods demonstrate that the fractal more easy.
出处
《计算机与数字工程》
2011年第3期24-25,40,共3页
Computer & Digital Engineering
关键词
脉象
识别
分形
神经网络
pulse
identify
fractal
neural networks
作者简介
陈雷，男，助教，研究方向：信号处理及计算机控制。
