采用小波变换的方法对采集自同一被试的不同情感数据样本进行分析,从小波系数中提取心电图信号的情感特征.对同一天采集自同一被试的4种情感的特征进行比较分析,得出大小关系一致的特征作为情感识别依据.选取的特征在归一化之后对高兴和悲伤2类情感分类效果较好,最高可以达到92%. The wavelet transform was used to analyze the data of different emotion samples collected from the same subject, and the emotion features of ECG signal were extracted from the wavelet coefficients. The comparative analysis of the four emotions collected from the same subject on the same day showed that The characteristics of the same size relationship are used as the basis of emotion recognition.The selected features have good effect on happy and sad category 2 emotion classification after the normalization, up to 92%.