A Computational Approach to Emotion Recognition in Intelligent Agent

Tee Connie
Faculty of Information Science and Technology
Multimedia University

Goh Kah Ong
Multimedia University


The ability to express emotions in an intelligent agent is a very important factor in creating an interesting and compelling interaction between human and the agent. In this paper, a computational approach is used to identify the emotions in the conversation between the human and the agent. With this, the agent is able to display different facial expressions throughout the conversation and make the interaction process more "real" and natural. Natural language is the main interaction channel in this research. An emotion recognition engine will be developed to analyze the emotion in the speech and instruct the agent to display the appropriate expressions.

© ATCM, Inc. 2002.