Recognizing low/high anger in speech for call centers

Authors:
Fu-Ming Lee;Li-Hua Li;Ru-Yi Huang
Affiliations:
Department of Information Management, Chaoyang University of Technology, Taichung County, Taiwan, R.O.C.;Department of Information Management, Chaoyang University of Technology, Taichung County, Taiwan, R.O.C.;Department of Information Management, Chaoyang University of Technology, Taichung County, Taiwan, R.O.C.
Venue:
ISPRA'08 Proceedings of the 7th WSEAS International Conference on Signal Processing, Robotics and Automation
Year:
2008

Citing 4
Cited 2

Emotion Detection from Speech to Enrich Multimedia Content

PCM '01 Proceedings of the Second IEEE Pacific Rim Conference on Multimedia: Advances in Multimedia Information Processing
2005 Special Issue: Challenges in real-life emotion annotation and machine learning based detection

Neural Networks - Special issue: Emotion and brain
Ensemble methods for spoken emotion recognition in call-centres

Speech Communication
An evaluation of the robustness of existing supervised machine learning approaches to the classification of emotions in speech

Speech Communication

A comparison of neural networks for real-time emotionrecognition from speech signals A comparison of neural networks for real-time emotionrecognition from speech signals

WSEAS Transactions on Signal Processing
Anger recognition in speech using acoustic and linguistic cues

Speech Communication

Quantified Score

Hi-index	0.00

Visualization

Abstract

Automatic multi-level anger recognition in speech is an important factor to enhance user satisfaction for call centers. In this research, three emotional states, namely, neutral, low anger, and high anger of acted corpora with telephone quality are specified for emotion recognition. The corpora are collected from amateur actors and, thereafter, verified by the actors themselves. The emotion recognizer is implemented by using Back-propagation Network (BPN) with acoustic features of speech examples. Due to the variation of expression methods by different person, the feature values of the training examples used are too complex to make the BPN model convergent. To overcome the problem, a codified method is developed to simplify the feature values. With the codified inputs, the BPN model and a comparative Decision Tree C5.0 give quite satisfactory test performances for anger recognition. Therefore, they can be used as a part of a decision support system for proper applications in call centers.