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

End-to-end BNF feature extraction method, network model, training method and system

A technology of network model and network module, which is applied in speech analysis, speech recognition, instruments, etc., can solve problems such as the inability to guarantee the accuracy of BNF features of the second network, and achieve improved accuracy, good fault tolerance, and better timeliness Effect

Pending Publication Date: 2022-04-12
NANJING SILICON INTELLIGENCE TECH CO LTD
View PDF0 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] In order to solve the technical problem that two networks need to be used to extract BNF features in the prior art, so that the accuracy of the second network output BNF features cannot be guaranteed, this application provides an end-to-end method for extracting BNF features, a network model, Training Method and System

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
  • End-to-end BNF feature extraction method, network model, training method and system
  • End-to-end BNF feature extraction method, network model, training method and system
  • End-to-end BNF feature extraction method, network model, training method and system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0064] The following will clearly and completely describe the technical solutions in the embodiments of the application with reference to the drawings in the embodiments of the application. Apparently, the described embodiments are only some of the embodiments of the application, not all of them. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts belong to the scope of protection of this application.

[0065] First of all, it should be noted that the brief description of the terms in this application is only for the convenience of understanding the implementations described below, and is not intended to limit the implementations of this application. These terms are to be understood according to their ordinary and usual meaning unless otherwise stated.

[0066] The terms "first", "second", and "third" in the description and claims of this application and the above drawings are used to distingu...

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 end-to-end BNF feature extraction method, a network model, a training method and a system, and the network model comprises a loop network module and a coding module: the loop network module is used for inputting Mel frequency cepstral coefficient features of audio of a source speaker and outputting down-sampling features; the coding module is used for inputting the down-sampling feature, obtaining a first feature based on a self-attention algorithm and deep convolutional learning, performing full connection processing on the first feature, and outputting a BNF feature of the audio of the source speaker, the first feature includes global dependencies and local dependencies between frames of the source speaker audio. According to the network model provided by the invention, end-to-end extraction of the BNF features can be realized, the timeliness is better, and intermediate errors are not superposed, so that the accuracy of the extracted BNF features is ensured.

Description

technical field [0001] The application belongs to the technical field of speech recognition, and in particular relates to an end-to-end extraction method of BNF features, a network model, a training method and a system. Background technique [0002] With the development of the Internet and artificial intelligence technology, voice interaction business scenarios are becoming more and more abundant. For example, telephone robots and virtual digital humans are used in various industries. The specific voice technologies involved in voice interaction services may include: ASR (Automatic Speech Recognition, automatic speech recognition), TTS (Text-To-Speech, from text to speech), sound cloning, speech conversion, noise reduction, etc., [0003] Among them, speech conversion is a research branch of speech signal processing, which covers the fields of speaker recognition, speech recognition and speech synthesis. It intends to change the personalized information of speech while keepi...

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): G10L25/24G10L25/30G10L15/26G10L15/02
Inventor 司马华鹏毛志强孙雨泽
Owner NANJING SILICON INTELLIGENCE TECH CO LTD
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