International Symposium on Chinese Spoken Language Processing (ISCSLP 2002)

Taipei, Taiwan
August 23-24, 2002

Emotion Recognition via Acoustic Features and Semantic Contents in Speech

Ze-Jing Chuang, Chung-Hsien Wu

National Cheng Kung University, Tainan, Taiwan

Recent researches into human-machine communication make more emphasis on the recognition of nonverbal information, especially on the topic of emotional reaction. Many kinds of physiological characteristics are used to extract emotions, such as voice, facial expression, hand gesture, body movement, even heartbeat and blood pressure. In this paper, based on the idea that humans are capable of detecting human emotions through speech input without other visual or physiological information, an emotion recognition system that can detect the emotion from acoustic features and semantic contents in speech is proposed. In this approach, the acoustic features are extracted for feature-based emotion extraction. On the other hand, the speech signal is also fed to a speech recognizer and the recognized contents are then used for content-based emotion extraction. Finally, the integration of the results from acoustic features and semantic contents is used to determine the final emotion.

Full Paper

Bibliographic reference.  Chuang, Ze-Jing / Wu, Chung-Hsien (2002): "Emotion recognition via acoustic features and semantic contents in speech", In ISCSLP 2002, paper 45.