Voice emotion feature fusion method and system based on main and auxiliary networks
A technology of emotional features and fusion methods, applied in speech analysis, instruments, etc., can solve problems such as unreasonable correspondence, failure to achieve the desired accuracy, and achieve the effect of improving the accuracy.
Active Publication Date: 2020-05-12
TAIYUAN UNIV OF TECH
View PDF4 Cites 1 Cited by
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
However, due to the different dimensions and dimensions of different features, as well as the difference in the actual physical meaning of each type of feature, different types of features affect each other, and the accuracy rate fails to achieve the desired effect.
[0005]Second, the correspondence between network output and actual labeling is unreasonable
Method used
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
View moreImage
Smart Image Click on the blue labels to locate them in the text.
Smart ImageViewing Examples
Examples
Experimental program
Comparison scheme
Effect test
Embodiment approach
[0125] As an implementation manner, the set determination module of the present invention specifically includes:
[0126] A speech emotion database determining unit is used to determine a speech emotion database; the speech emotion database includes 363 pieces of speech emotion data;
[0127] A standard database determining unit, configured to determine a standard database according to the speech emotion database;
[0128] Feature extraction unit, for carrying out feature extraction to each described speech emotion data in described standard database, obtains a plurality of speech MFCC segment features and a plurality of speech global features;
[0129] A standardization processing unit is used to standardize a plurality of the speech MFCC segment features corresponding to each of the speech emotion data and the speech global feature, respectively, to obtain the first feature and the second feature respectively;
[0130] The set determining unit is used to select the first fe...
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More PUM
Login to View More Abstract
The invention provides a voice emotion feature fusion method and a system based on main and auxiliary networks. The voice emotion feature fusion method comprises the following steps: respectively inputting a plurality of first features and second features corresponding to each piece of voice emotion data in a test set into the lower half part of a main network model with parameters and an auxiliary network model to obtain main network high-level features and auxiliary network high-level features corresponding to each piece of voice emotion data; performing feature fusion on the main network high-level features, auxiliary parameters and the auxiliary network high-level features, and determining main and auxiliary network fusion features corresponding to the voice emotion data; and inputtingthe main and auxiliary network fusion features corresponding to the voice emotion data into the upper half part of the main network model with parameters to obtain fusion features. According to the method, multiple types of features are effectively fused, and the accuracy of voice emotion fusion is improved.
Description
technical field [0001] The invention relates to the technical field of emotional feature fusion, in particular to a voice emotional feature fusion method and system based on primary and secondary networks. Background technique [0002] Emotional state is an important factor for effective communication between humans and machines. In order to realize the natural communication between human and machine, it is the goal that the field of artificial intelligence has been pursuing to make the machine have the ability of human speech, thinking and emotion at the same time. Research on Speech Emotion Recognition (SER) will promote the realization of this goal, and its results can be widely used in fields such as human-computer interaction, telemedicine, electronic education, criminal investigation and emotional guidance. Therefore, the development of speech emotion Identification research is of great significance and practical value. [0003] The main structure of the current spee...
Claims
the structure of the environmentally friendly knitted fabric provided by the present invention; figure 2 Flow chart of the yarn wrapping machine for environmentally friendly knitted fabrics and storage devices; image 3 Is the parameter map of the yarn covering machine
Login to View More Application Information
Patent Timeline
Login to View More Patent Type & AuthorityApplications(China)
IPC IPC(8): G10L25/63G10L25/27G10L25/03G10L25/24
CPCG10L25/63G10L25/27G10L25/03G10L25/24
Inventor张雪英胡德生张静黄丽霞牛溥华李凤莲
OwnerTAIYUAN UNIV OF TECH