A new approach of audio emotion recognition

作者：

Highlights：

• This paper presented a new architecture of intelligent audio emotion recognition system.

• The proposed architecture is designed to focus on the 6 universal emotions.

• Both prosodic and spectral features are fully utilized in the design of this parallel architecture.

• Simulation examples and comparison on databases have revealed good performance of the proposed architecture.

摘要

•This paper presented a new architecture of intelligent audio emotion recognition system.•The proposed architecture is designed to focus on the 6 universal emotions.•Both prosodic and spectral features are fully utilized in the design of this parallel architecture.•Simulation examples and comparison on databases have revealed good performance of the proposed architecture.

论文关键词：Audio emotion recognition,RBF neural network,Prosodic features,MFCC feature

论文评审过程：Available online 22 March 2014.

论文官网地址：https://doi.org/10.1016/j.eswa.2014.03.026