摘要
本文基于小波分形技术,用于神经系统的细胞外多神经元活性的小波分形识别研究。首先,我们应用多尺度小波变换对原始信号进行分解,得到4个子信号,计算子信号的分形维数,然后用KNN分类器进行分类。经过加噪信号的实验研究,结果表明该算法具有很高的分类准确率,为临床应用提供了有力的工具。
A new wavelet - fractal technique for the automatic online clustering of extracellular muhin euron recordings from the nervous system was presented. First, we employ the wavelet transform to obtain the sub - patterns of the original signal. Further from the resulting non - self- intersecting curves, the divider fractal dimensions are readily computed. In terms of the efficiency and accuracy, this method was compared with the other proposed techniques. And the resuits show that it is a powerful tool in this discrimination problem.
出处
《激光杂志》
CAS
CSCD
北大核心
2006年第4期62-63,共2页
Laser Journal
基金
国家自然科学基金资助项目(批准号:60371034)
关键词
电生理波
小波分解
分形维数
electrophysiological spike
wavelet transform
fractal
