Unlock instant, AI-driven research and patent intelligence for your innovation.

Speech emotion recognition method based on deep learning

A technology of speech emotion recognition and deep learning, applied in speech analysis, instruments, etc., to achieve the effect of benefiting performance, deepening understanding, and avoiding information loss

Pending Publication Date: 2022-04-22
HEFEI UNIV OF TECH +1
View PDF0 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Existing research has achieved rich results, but there is still a broad research space for speech emotion recognition based on acoustic features

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
  • Speech emotion recognition method based on deep learning
  • Speech emotion recognition method based on deep learning
  • Speech emotion recognition method based on deep learning

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0069] In the present embodiment, a kind of speech emotion recognition method based on deep learning is to carry out as follows:

[0070] Step 1: Extract LLDs sequence W l and spectrogram W g ;

[0071] Acquire speech samples [x(t), y], t=1, 2,..., N, where x(t) is the signal of the tth sampling point, y is the label of the speech sample, and N is the total number of sampling points of the speech sample ;

[0072] And perform framing processing on x(t), get the framing sequence of x(t): [x 1 (t),x 2 (t),...,x i (t),...,x n (t)], where x i (t) represents the ith voice frame of the t sampling point voice x(t), and n represents the number of voice frames;

[0073] Use formula (1) to perform windowing processing on the framed sequence to obtain the windowed sequence [x' 1 (t), x' 2 (t),...,x' i (t),...,x' n (t)], t=1,2,...,N:

[0074] x' i (t) = ω(t) x i (t) (1)

[0075] In formula (1), ω(t) represents the Hanning window function; x' i (t) represents the i-th wind...

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 a voice emotion recognition method based on deep learning. The method comprises the following steps: 1, obtaining a session voice set, and carrying out framing and windowing processing; 2, extracting an LLDs feature sequence and a speech spectrogram of the speech frame; 3, extracting the representation of the LLDs sequence; and 4, extracting the representation of the spectrogram. And 5, obtaining a final emotion prediction result by using a feature fusion and result fusion strategy. According to the invention, the emotion category of the voice can be effectively predicted, and the prediction accuracy can be improved.

Description

technical field [0001] The invention belongs to the field of speech data analysis and processing, in particular to a speech emotion recognition method based on deep learning. Background technique [0002] With the widespread application of intelligent service robots in e-commerce, hotels, shopping malls and other fields, voice-based emotion recognition has become an effective means for companies to understand consumer needs and analyze consumer satisfaction. Speech acoustic information has become the main basis for speech emotion recognition due to its advantages of low extraction complexity and less interference information. How to construct speech emotion recognition method based on acoustic information has important theoretical and practical value. The emotion recognition method based on acoustic features is the mainstream direction of speech emotion recognition. Existing research has achieved rich results, but there is still a broad research space for speech emotion re...

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
IPC IPC(8): G10L25/63G10L25/45G10L25/30G10L25/24
CPCG10L25/63G10L25/45G10L25/30G10L25/24
Inventor 姜元春葛鸿飞朱波穆利吴铭刘业政袁昆孙见山柴一栋钱洋
Owner HEFEI UNIV OF TECH