Abstract: In order to study meridian characteristics, a feature extraction method of ElectroCardioGraph (ECG) signal on the meridian based on wavelet packet analysis and energy entropy was proposed. A meridian measuring experiment was firstly built to complete the acquisition of meridian data. Then meridian ECG signals were decomposed by a three layer wavelet packet decomposition. Energy entropy features of meridian ECG signals were extracted according to the results of signal reconstruction. After that, both K-means and Fuzzy C-Means (FCM) clustering techniques realized the effective partition of acupoints and non-acupoints. The derived clustering results indicate that the energy entropy values of ECG signals on the acupoints are obviously higher than those on the non-meridian points. It can be used as a powerful scientific basis for the discrimination of acupoints and non-acupoints.

Key words: meridian acupoint, ElectroCardioGraph (ECG) signal, wavelet packet, entropy, feature extraction, clustering

摘要： 为研究人体经络特征，提出了基于小波包分析的经络穴位心电信号熵特征提取的方法。首先通过建立经络检测实验采集了经络测试点的心电信号，然后采用小波包对经络心电信号进行三层分解，并根据重构后的心电信号提取经络穴位的熵特征。同时采用了K-means和模糊C均值聚类方法实现了穴位点和非穴位点的有效分类。研究结果表明经络上穴位点心电信号的能量熵明显大于非穴位点的熵值，并且这一特征可以作为区分经络穴位点和非穴位点的有力科学依据。

关键词: 经络穴位, 心电信号, 小波包, 熵, 特征提取, 聚类