Cross-language speech emotion recognition in German and Chinese

被引：0

作者：

Huang, Chengwei ^{[1
]}

Han, Dong ^{[2
]}

Bao, Yongqiang ^{[3
]}

Yu, Hua ^{[1
]}

Zhao, Li ^{[1
]}

机构：

[1] School of Information Science and Engineering, Southeast University, No. 2, Si Pai Lou, Nanjing 210096, China

[2] Department of Computer Science, University of Houston, No. 4800, Calhoun Road, Houston, TX 77004, United States

[3] School of Communication Engineering, Nanjing Institute of Technology, No. 1, Hongjing Ave., Jiangning Science Park, Nanjing 211167, China

来源：

ICIC Express Letters, Part B: Applications | 2012年 / 6卷 / 08期

关键词：

Speech recognition - Emotion Recognition - Linguistics;

D O I：

暂无

中图分类号：

学科分类号：

摘要：

Cross-language speech emotion recognition is an unsolved challenge in automatic emotion recognition by computers. We experimentally studied this problem in two languages: German and Chinese. Cross-language features are selected across the German and Chinese speech emotion databases. Acoustic models are trained in one language and tested in the other language based on the Gaussian mixture model. Angry emotion is successfully recognized across German and Chinese. The universality of vocal emotional behavior is an essential character to understand human emotions. The experimental results support the existence of a cross-language emotion model on acoustic features. © 2012 ICIC International.

引用

页码：2141 / 2146