A speech emotion recognition model generation method using a Max-margin framework incorporating a loss function based on the Watson-Tellegen's Emotion Model

摘要

The present invention is based on the WTM (Watson-Tellegen Emotional Model) and in the method of building a model for recognizing the emotions carried in the voice through the training of the emotion of the voice through the HMM (Hidden Markov Model) and training data, WTM The second and second steps of calculating the value of the loss function based on the values set in the first step and the first step of quantifying the difference between each emotion using the geometric distance between the emotion groups of A speech emotion recognition model construction method using a loss function and a maximum margin method based on a WTM comprising a third step of obtaining a parameter of each speech emotion model based on the loss function obtained in the step. The emotion recognition model can be expected to improve the performance of speech emotion recognition.;Speech emotion recognition, max-margin, loss function, watson-tellegen model