Fuzzy emotion recognition in natural speech dialogue

摘要

This paper describes the realization of a natural speech dialogue for the robot head MEXI with focus on its emotion recognition. Specific for MEXI is that it can recognize emotions from natural speech and also produce natural speech output with emotional prosody. For recognizing emotions from the prosody of natural speech we use a fuzzy rule based approach. Since MEXI often communicates with well known persons but also with unknown humans, for instance at exhibitions, we realized a speaker-dependent mode as well as a speaker-independent mode in the prosody based emotion recognition. A key point of our approach is that it automatically selects the most significant features from a set of twenty analyzed features based on a training data base of speech samples. This is important according to our results, since the set of significant features differs considerably between the distinguished emotions. With our approach we reached average recognition rates of 84% in speaker-dependent mode and 60% in speaker-independent mode.