基于PCA和SVM的普通话语音情感识别

在语音情感识别中,情感特征的选取与抽取是重要环节。目前,还没有非常有效的语音情感特征被提出。因此,在包含6种情感的普通话情感语料库中,根据普通话不同于西方语种的特点,选取了一些有效的情感特征,包含Mel频率倒谱系数、基频、短时能量、短时平均过零率和第一共振峰等,进行提取并计算得到不同的统计量；接着采用主成分分析(PCA)进行抽取；最后利用基于支持向量机(SVM)的语音情感识别系统进行分类。实验结果表明, 与其他一些重要的研究结果相比,该方法得到了较高的平均情感识别率, 且情感特征的选取、抽取及建模是合理、有效的。

Speech Emotion Recognition in Mandarin Based on PCA and SVM

Feature selection and extraction play a vital role in speech emotion recognition.At present,no effective speech emotion features are proposed.For these reasons,according to the characteristic of Mandarin which is different from western languages,some effective emotional features and related statistics,including Mel-frequency cepstral coefficients,pitch frequency,short-time energy,short-time average zerocrossing rate,the first formant and so on,were analyzed on a Mandarin emotional corpus which contains 6 kinds of emotions.Then we chose principal component analysis (PCA) for extraction and presented a speech emotion recognition method based on support vector machine (SVM) for classification.The experimental results show that the proposed method achieves high emotion recognition accuracy compared with several significant methods,and the emotion extraction and modeling are reasonable and effective.