【摘要】 语音情感识别是用计算机去识别输入语音信号的特征，计算分析得出情感状态。文章建立的语音情感识别系统，用支持向量机SVM去识别输入的语音模型，对输入的语音信号进行预加重处理、端点检测、mfcc情感特征提取，与已训练语音提取的情感特征进行一定的特征比对，分析情感状态。通过对语音数据库模型的训练和测试，针对高兴、悲伤、愤怒等三种情感，最终对语音数据库里的语料达到了71.3%的识别率。最后根据识别的情绪进一步进行相关的心理辅导。更多还原

【Abstract】 Speech emotion recognition is to use the computer to recognize the features of the input speech signal, and calculate and analyze the emotional state. In the speech emotion recognition system built in this paper, SVM is used to recognize the input speech model. The system performs preemphasis processing, endpoint detection, MFCC emoion feature extraction on the input speech signal, and compares the emotion feature with the emotion feaure extracted by the trained speech to get the existing emotion.Through the training and testing of the speech database mod-el, aiming at the three emotions of happiness, sadness and anger, the recognition rate of the corpus in the speech database reaches 71.3%. And according to the identified emotions for further psychological counseling.更多还原