Speech emotion identifying method based on supporting vector machine

Abstract

A method for identifying voice emotion based on support vector computer includes characteristic picking up-analyzing to collate characteristic parameter selection of global structure and sex, as well as to collate characteristic parameter selection of time sequence structure and to collate sex and vowel number; support vector computer training to carry out identification on five emotions of happy, angry, sad, fear and surprise.

Description

technical field [0001] The invention relates to a speech recognition method, in particular to a speech emotion recognition system and method. Background technique [0002] The automatic speech emotion recognition technology mainly includes two problems: one is to use the characteristics in the speech signal as emotion recognition, that is, the problem of emotional feature extraction, and the other is how to classify specific speech data, that is, the problem of pattern recognition . [0003] Before the present invention, the main emotional features currently used are prosodic features and their derived parameters, such as duration, speech rate, amplitude, pitch frequency, formant, etc. Among them, the pitch frequency and the formant frequency are important emotional parameters, but due to the individual differences between people (variability of the vocal tract, vocal tract characteristics, word pronunciation pitch, etc.), it is difficult to realize the pitch It is difficu...

AI Technical Summary

Problems solved by technology

In the past literature, these gender differences in acoustic parameters without considering emotional factors have been fully studied. However, there are few literatures that specifically study the influence of characteristic parameters of different genders on emotional states.

Among the various existing recognition methods, specifically, the vector segmentation type Mahalanobis distance judgment method and the principal component analysis method are too simple to obtain a good recognition rate; although the neural network method has a high degree of nonlinearity and strong Classification ability, but with the increase of the network, the required learning time increases rapidly, and the local minimum problem is also a shortcoming; the hidden Markov method (HMM) takes a long time to establish and train, and it is applied in practice. It is also necessary to solve the problem of excessive computational complexity

In addition, although the above method has a certain recognition rate in the experimental environment, it is mainly based on the existing speech library samples to optimize the performance of the training set, and the actual environment often has a large difference in the training set.

Images