Abstract

Abstract. Human language carries various kinds of information. In human computer interaction the detection of the emotional state of a speaker as re ected in his or her utterances is crucial. In this investigation we will explore how acoustic and prosodic information can be used to detect the emotional state of a speaker. We will show how prosodic information can be combined and integrated with acoustic information within a hidden Markov model architecture, which allows one to make observations at a rate appropriate for the phenomena to be modeled. Using this architecture, we will demonstrate that prosodic information adds discriminative power to the overall system.

Citations