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

Speech enhancement method, system, computer equipment and storage medium

A voice enhancement and target voice technology, applied in computer equipment and storage media, in the field of voice enhancement based on acoustic vector sensors and deep neural networks, can solve the problems of large acquisition system, high computational complexity, and high hardware cost, and achieve algorithmic Low complexity, low hardware cost, effect to remove residual noise and reverberation

Active Publication Date: 2022-04-19
PEKING UNIV SHENZHEN GRADUATE SCHOOL
View PDF7 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

According to different audio collection devices, speech enhancement technology can be divided into single-channel and multi-channel speech enhancement. Among them, multi-channel speech enhancement has the advantages of more effective suppression of environmental noise and reverberation, but has the advantages of high hardware cost, large acquisition system volume, High computational complexity and other limitations

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 enhancement method, system, computer equipment and storage medium
  • Speech enhancement method, system, computer equipment and storage medium
  • Speech enhancement method, system, computer equipment and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0077] The technical solutions of the present invention will be clearly and completely described below in conjunction with the embodiments. Obviously, the described embodiments are part of the embodiments of the present invention, but not all of them. Based on the embodiments of the present invention, all other embodiments obtained by persons of ordinary skill in the art without making creative efforts belong to the protection scope of the present invention.

[0078] The present embodiment provides a method for speech enhancement based on an acoustic vector sensor and a deep neural network. The method uses an acoustic vector sensor (Acoustic Vector Sensor, AVS) to collect audio signals. In remote speech applications, four sensors of the AVS are considered to have In the same spatial position, the four sensors of AVS collect and output four-channel voice signals synchronously. Usually AVS selects the pressure sensor as the omnidirectional sensor, the particle velocity sensor an...

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 provides a speech enhancement method, system, computer equipment and storage medium, and relates to the technical field of human-computer speech interaction, including collecting multi-channel acoustic signals through an acoustic vector sensor, preprocessing the multi-channel acoustic signals and obtaining time spectrum , filter the time-spectrum and output the signal spectrum; mask the signal spectrum through a nonlinear mask and output the enhanced single-channel spectrogram; input the single-channel spectrogram into the deep neural network mask estimation model and Output the mask spectrogram, and perform time-frequency masking enhancement processing on the signal spectrum through the mask spectrogram to obtain the enhanced amplitude spectrogram; output the enhanced target speech signal through the reconstruction of the enhanced amplitude spectrogram, which solves the problem of multi-channel speech enhancement It has the technical problems of high hardware cost, large volume of the acquisition system, and high computational complexity. It can obtain excellent technical effects of speech enhancement effects under different interference noise types, strengths, and room reverberation conditions.

Description

technical field [0001] The invention relates to the technical field of human-computer voice interaction, in particular to a voice enhancement method, system, computer equipment and storage medium based on an acoustic vector sensor and a deep neural network. Background technique [0002] Speech enhancement technology is an important research direction of speech signal processing and one of the core technologies of speech processing systems. It has a wide range of applications in mobile phones, hearing aids, service robots and smart homes. The purpose of speech enhancement is to suppress non-target speech and noise interference signals in the collected multi-channel acoustic signal, and at the same time enhance the target speech signal, thereby improving the intelligibility of speech and improving the performance of the speech recognition system. According to different audio collection devices, speech enhancement technology can be divided into single-channel and multi-channel ...

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): G10L21/0208G10L21/0216G10L25/30G10L25/03
CPCG10L21/0208G10L21/0216G10L25/30G10L25/03G10L2021/02082G10L2021/02166
Inventor 邹月娴刘钊祎张皓然
Owner PEKING UNIV SHENZHEN GRADUATE SCHOOL
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