摘要 |
PURPOSE: A feeling recognition method using a speech is provided to extract and analyze various parameters showing feelings of human beings from spoken voice signals, thereby efficiently estimating the feelings of the human beings. CONSTITUTION: In a speech sound inputting step, a voice signal in accordance with a speaker's speech is inputted(100). In a non-verbal parameter extracting step, at least one non-verbal parameter is extracted from a spoken sound(110). In a verbal parameter extracting step, at least one verbal parameter in accordance with a meaning of the spoken sound is extracted by recognizing the voice signal(120). In a feeling calculating step, a feeling of the speaker is finally calculated by synthesizing the verbal parameters with the non-verbal parameters(130). |