Method and Apparatus for Recognizing Human Emotion Expressions based on Speech Signal

A method and apparatus for recognizing human emotion expressions based on speech signal is disclosed. In one embodiment, a method includes filtering an input speech signal corresponding to a voiced acoustic expression of an individual and determining glottal closure instance (GCI) locations in the speech signal using the filtered signal. Further, the method includes computing instantaneous values of a set of acoustic features of the speech signal using the estimated GCI locations in the speech signal. The instantaneous values of the set of acoustic features are normalized with reference to a same set of acoustic features of a neutral speech signal associated with the individual. A multi–dimensional joint distribution is generated based on the normalized instantaneous values of acoustic features and matched with a pre-defined emotion specific template. The Moreover, the method includes determining a human emotion expression associated with the individual by calculating the number of matched found.