中文摘要：

为了实现脑电图分析定量化、智能化以便为医学分析诊断提供客观有效的辅助信息,探讨了应用谐波小波包变换分析脑电波.通过谐波小波包变换提取用于临床诊断的,δ,θ,α,β等4种基本节律的波形,并引入反映基本节律变化的特征参量即基本节律的频带能量比例进行定量分析.经过对正常人和癫痫病人的脑电波进行分析验证,其分析结果与先验知识和确诊病症吻合得很好,可以提取得到精确的、量化的、直观的特征参数作为诊断依据.实验结果表明,应用谐波小波包变换分析脑电波是一种有效的方法,它为脑电图机实现智能化和临床分析诊断提供了有益的参考.

英文摘要：

In order to implement intelligent electroencephalogram（EEG） analysis and provide some objective and effective information for medical diagnosis,a new EEG analysis method based on discrete harmonic wavelet packet transform is proposed.First,the four basic rhythms,i.e.,δ,θ,α,and β wave,which are very useful for medical diagnosis,are extracted accurately by discrete harmonic wavelet packet transform.Next,based on the transformation,quantitative parameter such as frequency band energy ratio of each basic rhythm from single electrode is introduced to analyze the basic rhythms.The experimental results on healthy person＇s EEG and epileptic＇s EEG are provided,and the results agree well with prior knowledge and proved diagnosis conclusions.And accurate,quantitative and intuitive feature parameters are obtained.The experimental results demonstrate that the proposed method is an effective way to analyze EEG.It provides a useful reference for intelligent EEG analysis and medical diagnosis.