Looking for breakthrough ideas for innovation challenges? Try Patsnap Eureka!

Generative Adversarial Network Speech Enhancement Method Based on Deep Fully Convolutional Neural Network

A technology of convolutional neural network and network speech, which is applied in the field of speech enhancement based on deep fully convolutional neural network, can solve the problems of poor speech signal quality, and achieve the effect of reducing influence and enhancing speech signal

Active Publication Date: 2022-02-11
NORTHWESTERN POLYTECHNICAL UNIV
View PDF8 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] The purpose of the present invention is to provide a method for speech enhancement based on deep fully convolutional neural networks to solve the problem of poor speech signal quality in existing high-noise environments

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
  • Generative Adversarial Network Speech Enhancement Method Based on Deep Fully Convolutional Neural Network
  • Generative Adversarial Network Speech Enhancement Method Based on Deep Fully Convolutional Neural Network
  • Generative Adversarial Network Speech Enhancement Method Based on Deep Fully Convolutional Neural Network

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0027] The present invention will be described in detail below in conjunction with the accompanying drawings and specific embodiments.

[0028] The present invention provides a method for generating adversarial network speech enhancement based on deep fully convolutional neural networks, such as figure 1 As shown, the specific implementation is as follows:

[0029] Step 1. Construct a data set, including a noisy speech signal, and a pure speech signal corresponding to the noisy speech signal; obtain the spectrogram of the noisy speech signal, and use the spectrogram as the Input to generator G. Among them, the method of obtaining the spectrogram is as follows: for processing the noisy speech signal, first divide into frames, and then perform Fourier transform to obtain the graph of the speech spectrum changing with time, that is, the spectrogram of the noisy speech.

[0030] Step 2. The generator G of the generative adversarial network model based on the deep fully convolut...

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 generation confrontation network speech enhancement method based on a deep fully convolutional neural network, which solves the problem of poor speech signal quality in an existing high-noise environment. The method comprises the following steps: step 1, constructing a data set, including a noisy speech signal and a pure speech signal; obtaining the spectrogram of the noisy speech signal, using the spectrogram as the input of the generator G in the generation confrontation network ; Step two, the generator G processes the spectrogram generated in step one, constructs the generator as an encoder-decoder structure, obtains the latent vector z through the encoder part, and then passes the latent vector z through the decoder part Obtain the output signal of the speech signal; step 3, set the structure of the discriminator D, then use the output signal in the step 2 and the pure speech signal in the step 1 as the input of the discriminator D, the two constantly interact and confront, Train until equilibrium is reached to obtain enhanced speech signals.

Description

【Technical field】 [0001] The invention belongs to the technical field of speech recognition, and in particular relates to a speech enhancement method based on a deep fully convolutional neural network. 【Background technique】 [0002] In the speech signal, there must be a variety of interfering noises. The purpose of speech enhancement is to remove the unnecessary noise contained in the signal to the greatest extent, improve the quality of noisy speech, make the listener happy to accept it, and increase the reliability of the speech at the same time. Comprehension makes it easy for the listener to understand. The enhanced speech signal should be infinitely close to the pure speech signal on the waveform as much as possible to facilitate subsequent signal processing. Nowadays, speech enhancement is used in a wide range of applications, such as military communications, eavesdropping technology and speech recognition. However, due to the randomness, diversity and instability o...

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 Patents(China)
IPC IPC(8): G10L21/02G10L21/0208G10L25/30G10L25/60G06N3/04
CPCG10L21/0208G10L25/30G10L25/60G10L21/0364G06N3/045
Inventor 李立欣程倩倩李旭程岳
Owner NORTHWESTERN POLYTECHNICAL UNIV
Who we serve
  • R&D Engineer
  • R&D Manager
  • IP Professional
Why Patsnap Eureka
  • Industry Leading Data Capabilities
  • Powerful AI technology
  • Patent DNA Extraction
Social media
Patsnap Eureka Blog
Learn More
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com