| 声学参数和听觉参数结合的语音情感交叉识别 |
| |
| 引用本文： | 乔冠楠,胡剑凌,刘鹏.声学参数和听觉参数结合的语音情感交叉识别[J].电声技术,2009,33(6). |
| |
| 作者姓名： | 乔冠楠 胡剑凌 刘鹏 |
| |
| 作者单位： | 1. 上海交通大学,图像通信与信息处理研究所,电子工程系,上海,200240
2. 上海交通大学,图像通信与信息处理研究所,电子工程系,上海,200240;苏州大学,电子信息学院,江苏,苏州,215021 |
| |
| 摘 要： | 语音情感识别中,情感特征信息的提取和选择、情感识别模型的选择是2个重要部分.结合语音信号的声学特征参数和听觉特征参数进行情感识别,针对两类不同情感之间的差别选择最优的特征集,并设计了一个基于神经网络的情感交叉识别,与听觉特征参数结合,经过分类器得到识别情感,达到平均92%识别率.
|
| 关 键 词： | 语音 神经网络 特征选择 情感交叉识别 |
Speech Emotion Recognition Combining Acoustic Features and Aural Features |
| |
| QIAO Guan-nan,HU Jian-ling,LIU Peng.Speech Emotion Recognition Combining Acoustic Features and Aural Features[J].Audio Engineering,2009,33(6). |
| |
| Authors: | QIAO Guan-nan HU Jian-ling LIU Peng |
| |
| Affiliation: | 1. Institute of Image Communication and Information Processing;Department of Electronic Engineering;Shanghai Jiao Tong University;Shanghai 200240;China;2. School of Electronic & Information;Suzhou University;Suzhou Jiangsu 215021;China |
| |
| Abstract: | The features selection and the recognition model construction are both the most important parts in emotion recognition of speech. In this paper,the acoustic features and aural features of speech are combined to recognize four kinds of emotion. The most excellent features are selected based on the discrimination of each two-group emotions. Besides,a recognition system of crossed ANN model combined with aural feature recognition to archive the result is designed. These methods improve the performance of the e... |
| |
| Keywords: | speech ANN feature selection crossed emotion recognition |
| 本文献已被 CNKI 维普 万方数据 等数据库收录！ |
|
