Voice enhancement method based on deep neural network

A deep neural network and speech enhancement technology, which is applied in the field of speech enhancement based on deep neural network, can solve the problems of poor speech enhancement effect of hearing aids and affecting user experience, etc.

Inactive Publication Date: 2019-04-19
CHONGQING UNIV OF POSTS & TELECOMM
View PDF12 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

It is also an important application to improve and improve the voice enhancement function of hearing aids. Most people with hearing impairments do not choose to wear hearing aids. One of the main reasons is that the voice enhancement effect of hearing aids is not good. The noise is amplified at the same time, which seriously affects the user experience, and this technology can effectively improve the quality of speech while filtering the noise, so it is very suitable for application in hearing aids

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
  • Voice enhancement method based on deep neural network
  • Voice enhancement method based on deep neural network
  • Voice enhancement method based on deep neural network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0042] In order to express the object, technical solution and advantages of the present invention more clearly, the present invention will be further described in detail below in conjunction with the accompanying drawings and specific implementation cases.

[0043] figure 1 The present invention proposes and uses the speech enhancement model schematic diagram based on deep learning, comprises the following steps:

[0044] ① Data set: provide training data set and test data set;

[0045] ② Model building and training: build and train a feature mapping deep neural network model based on DNAT-DSAT-DNN;

[0046] ③ Perform model decoding on the test noisy speech signal to obtain enhanced speech logarithmic power spectrum features;

[0047] details as follows:

[0048] First collect and organize the data sets, provide the noisy speech signal and pure speech signal data set pairs required for model training, build a feature mapping network model based on DNAT-DSAT-DNN, and then pr...

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 enhancement method based on a deep neural network in order to overcome the shortcomings that a traditional DNN (deep neural network) voice enhancement method based on characteristic mapping is poor in noise-robustness and inaccurate in characteristic mapping, enhanced voice has spectral distortion, and the like. The voice enhancement method includes the steps: firstly, proposing and building a characteristic mapping deep neural network model (DNAT-DSAT-DNN) based on dynamic noise and voice joint sensory training, and learning the characteristic mapping relationship between a noisy voice signal and a pure voice signal to obtain the log power spectrum characteristic value of an enhanced voice signal; secondly, acquiring phase information of the enhanced voicesignal according to the geometrical relationship among the noisy voice signal, the pure voice signal and the noise signal; finally, restoring time domain expression of the enhanced voice signal by theaid of an overlap-add principle.

Description

technical field [0001] The invention relates to the fields of speech enhancement and digital speech signal processing, in particular to a speech enhancement method based on a deep neural network. Background technique [0002] The transmission of information through voice is the most important, effective and commonly used form of exchanging information for human beings. Language is a unique function of human beings, and sound is the most commonly used tool for human beings. With the continuous development of computer technology, people are increasingly demanding to get rid of the shackles of their hands and replace it with voice as the information entrance of intelligent equipment, so as to realize the dream of man-machine dialogue; however, the acoustic environment of human life is extremely complex, usually Interferenced by various noises, it is necessary to implement the speech enhancement function in the front-end module of speech signal processing in order to carry out ...

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): G10L21/0216G10L21/02G10L25/30G10L25/03
CPCG10L21/02G10L21/0216G10L25/03G10L25/30
Inventor 李湑李秋俊陈毅彭鑫黄胜
Owner CHONGQING UNIV OF POSTS & TELECOMM
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