Age driven automatic speech emotion recognition system

With the advent of technology, speech recognition is no longer just the capability of the humans. Voice based interfaces can turn most favorable for human computer interaction if computers respond according to its users emotional state. Emotion recognition from speech is a challenging problem as the system has to interact with diverse user utterances. This paper presents an age driven speech emotion recognition system. The performed emotion recognition experiment reveals that, the classifier model trained according to its user's age group show improved accuracy over the model based on simply acoustic features. The age and emotion detection method adopted employs extraction of basic prosodic and spectral feature from the emotional speech corpuses and uses Support Vector Machine (SVM) algorithm for classification.