【摘要】 提出了一种对于构成针电极肌电信号(NEMG)的不同形状的运动单位动作电位(MUAP)进行分类的新方法.该方法采用小波变换(WT)的多尺度分析提取表征MUAP特性的时 频特征,再用无导师 有导师混合模式识别网络完成对输入样本的学习和分类.实验结果表明,这种方法对不同波形的MUAP具有较强大的分类能力,整个分类过程完全人工干预,且所需的时间不多,可以应用于临床实际,为神经肌肉疾病的诊断提供实时、有效的信息.更多还原

【Abstract】 A new method, which adopts the timefrequency feature extraction technology of motor unit action potential (MUAP) by means of wavelet transform and the classification technology based on supervisedunsupervised artificial neural network pattern recognition, of classifying various MUAPs of the NEMG signal is presented in this paper. The results of the experiments show that the method has a very strong ability at clustering different MUAPs and can be used in clinical applications to supply immediate and effective information for the diagnosis of some muscular diseases.更多还原