Attention mechanism based voice emotion identification method

A technology of speech emotion recognition and emotion recognition, which is applied in speech recognition, speech analysis, mechanical equipment, etc., can solve the problems of emotional information weakening, achieve the effect of optimizing emotion recognition performance, good application prospects, and improving performance

Active Publication Date: 2019-01-29
SOUTHEAST UNIV
View PDF3 Cites 27 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, for the speech emotion recognition task, the speech is mostly silent at the end and contains almost no emotional information. Therefore, the emotional information contained in the output corresponding to the last moment of the LSTM will be weakened.

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 more

Image

Smart Image Click on the blue labels to locate them in the text.
Viewing Examples
Smart Image
  • Attention mechanism based voice emotion identification method
  • Attention mechanism based voice emotion identification method
  • Attention mechanism based voice emotion identification method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0057] The present invention will be further described below in conjunction with the accompanying drawings.

[0058] Such as figure 1 Shown, the speech emotion recognition method based on attention mechanism of the present invention, comprises the following steps,

[0059] Step (A), extracting the speech features with timing information from the original speech data, wherein, extracting the speech features with timing information is to retain the timing information in the original speech data through the sequence relationship between speech frames, and this has timing The dimension of the voice features of the information varies with the actual length of the original voice data. The detailed voice feature set is shown in Table 1 below.

[0060] Table 1 Detailed speech feature set table

[0061]

[0062]

[0063] Step (B), establishing an LSTM model capable of processing variable-length data, the specific implementation process is as follows in the calculation method of...

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

No PUM Login to view more

Abstract

The invention discloses an attention mechanism based voice emotion identification method. The method comprises the following steps that voice characteristics with sequential information is extracted from original voice data; an LSTM model capable of processing length variable data is established; a forgetting door calculation manner in the LSTM model is optimized via the attention mechanism; the optimized LSTM model is output, and attention in the time and characteristic dimensions is weighted; a full connection layer and a soft maximal layer are added to the LSTM model to form a complete emotion identification network model; and the emotion identification network model is trained, and an identification performance of the emotion identification network model is evaluated. The attention mechanism based voice emotion identification method can be used to improve the voice emotion identification performance, the method is ingenious and novel, and the method has good application prospects.

Description

technical field [0001] The invention relates to the technical field of speech emotion recognition, in particular to a speech emotion recognition method based on an attention mechanism. Background technique [0002] Speech emotion recognition has important application value in human-computer interaction. In order to realize the automatic recognition of speech emotion, many scholars have done a lot of research work on machine learning algorithms, such as support vector machines, Bayesian classifiers and K-nearest neighbor algorithms. In recent years, with the development of deep learning, its application in automatic speech emotion recognition is also increasing. Deng (scholar) did semi-supervised learning with autoencoders and a small amount of emotional label data, and Neumann (scholar) applied convolutional neural networks to speech emotion recognition. [0003] Although the above algorithms have been successfully applied in emotion recognition, both traditional machine l...

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
no application Login to view more
Patent Type & Authority Applications(China)
IPC IPC(8): G10L25/63G10L25/27G10L15/06
CPCG10L15/06G10L15/063G10L25/27G10L25/63Y02T10/40
Inventor 谢跃梁瑞宇梁镇麟郭如雪
Owner SOUTHEAST UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Try Eureka
PatSnap group products

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap