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

Edge device-oriented voice recognition-synthesis joint modeling method

A speech recognition and edge device technology, applied in the field of speech recognition-synthesis combined modeling, can solve the problems of lack of model performance, operational process loopholes, poor input data characteristics, etc., to achieve rich processing capabilities, robust performance, model Robustness guaranteed effect

Active Publication Date: 2020-02-21
NORTHWEST UNIV(CN)
View PDF4 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The reasons are mostly poor input data characteristics, lack of performance of the model itself, and loopholes in the operation process.

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
  • Edge device-oriented voice recognition-synthesis joint modeling method
  • Edge device-oriented voice recognition-synthesis joint modeling method
  • Edge device-oriented voice recognition-synthesis joint modeling method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0029] Below in conjunction with accompanying drawing and embodiment the present invention is further described, but the present invention is not limited to following embodiment:

[0030] Such as figure 1 , 2 , 3, an edge device-oriented speech recognition-synthesis joint modeling method, including the following steps:

[0031] 1) Collect a dataset sample. Divided into a. Clean audio in a quiet environment b. Different types of noise audio (specifically involved: white noise, pink noise, speech babble, etc., refer to the noise noise library for classification) All audio data are sampling rate 16k, storage format pcm ( Shaanxi, Minnan, Changsha, Sichuan, Hebei, Shanghai dialects);

[0032] 2) Perform data processing. First do noise fusion processing, add noise to clean audio, package and assemble into clean audio and corresponding noise-added audio;

[0033] 3) Build an edge server, do audio front-end processing on this layer of equipment to perform reverberation, noise re...

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 edge device-oriented voice recognition-synthesis joint modeling method. Through research on real-time performance calculation, distribution of edge computing strategies andinspiration of an entertainment game 'just repeat', a back end fuses model iteration methods of voice recognition and voice synthesis technologies. A real-time efficient processing module is established based on a voice enhancement function in the field of audio processing, a voice recognition and synthesis iteration model for Chinese dialects is built based on the voice recognition technology andthe voice synthesis technology, the characteristics of the voice technologies are fully utilized to realize a real-time dialect processing model with the characteristics of recognition, synthesis andhigh efficiency, the richer processing capabilities of the edge environment are effectively utilized, the voice recognition and voice synthesis technologies are combined, and a voice model with the richer functions and the robuster performance is designed.

Description

technical field [0001] The invention belongs to the technical fields of edge computing and audio research, and relates to an edge server, speech enhancement, speech recognition, speech synthesis, neural network, and in particular to a joint speech recognition-synthesis modeling method for edge devices. Background technique [0002] After Industry 4.0, the rapid rise of artificial intelligence and the Internet of Things (IoT) has provided great potential for the convenience of human beings in terms of food, clothing, housing and transportation, and many smart products have emerged as the times require. At the same time, with the development of edge computing in recent years, the edge computing strategy can effectively realize the distribution of large task calculations, solve real-time problems, and improve the calculation ability of the model. Therefore, it provides infinite possibilities for continuously strengthening and expanding the functions of smart products. [0003]...

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): G10L13/02G10L13/04G10L15/02G10L15/06G10L15/26G10L25/24G10L25/30
CPCG10L13/02G10L13/04G10L15/02G10L15/063G10L15/26G10L25/24G10L25/30
Inventor 王海秦晨光张晓刘艺赵子鑫高岭任杰郑杰
Owner NORTHWEST UNIV(CN)
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