Eureka AIR delivers breakthrough ideas for toughest innovation challenges, trusted by R&D personnel around the world.

Speaker verification method and system based on tensor structure and sparse representation

A technology of speaker confirmation and sparse representation, applied in the field of speaker recognition, can solve problems such as a large amount of memory, slow down the recognition process, limit the number of training samples, etc., to reduce the computational complexity and improve the confirmation efficiency.

Active Publication Date: 2019-07-12
HANGZHOU DIANZI UNIV
View PDF6 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

But due to the over-complete dictionary, the sparse representation of high-dimensional supervectors requires a lot of memory, which limits the number of training samples and may slow down the recognition 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
  • Speaker verification method and system based on tensor structure and sparse representation
  • Speaker verification method and system based on tensor structure and sparse representation
  • Speaker verification method and system based on tensor structure and sparse representation

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0033] This embodiment provides a speaker confirmation method based on tensor structure and sparse representation, such as figure 1 shown, including steps:

[0034] S1. Construct an auditory feature tensor;

[0035] The present invention processes speech signals by simulating the auditory system of the human ear to obtain its power spectrum. In order to obtain robust features based on tensor structure, this embodiment models the obtained power spectra of different speakers as a third-order tensor quantity.

[0036] Specifically, the human ear can easily perform speaker recognition tasks and is insensitive to noise. In our feature extraction framework, we obtain frequency-selective information by mimicking the processes performed by the human ear in the auditory periphery and pathways.

[0037] First, the present invention extracts features by simulating the process of auditory periphery and pathway occurrence, such as outer ear, middle ear, basilar membrane, inner hair cell...

Embodiment 2

[0099] This embodiment provides a speaker confirmation system based on tensor structure and sparse representation, such as figure 2 shown, including:

[0100] Building blocks for constructing auditory feature tensors;

[0101] The present invention processes speech signals by simulating the auditory system of the human ear to obtain its power spectrum. In order to obtain robust features based on tensor structure, this embodiment models the obtained power spectra of different speakers as a third-order tensor quantity.

[0102] Specifically, the human ear can easily perform speaker recognition tasks and is insensitive to noise. In our feature extraction framework, we capture frequency-selective information by mimicking the processes performed by Ren in the auditory periphery and pathways.

[0103]First, the present invention extracts features by simulating the process of auditory periphery and pathway occurrence, such as outer ear, middle ear, basilar membrane, inner hair ce...

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 speaker verification method and system based on a tensor structure and sparse representation. The method comprises the following steps of Step 1, building an auditory featuretensor; Step 2, converting the auditory feature tensor into a spare matrix; Step 3, reducing the dimension of the sparse matrix so as to generate a final feature vector; and Step 4, performing speaker verification on the basis of a sparse representation classifier. The speaker verification method and system have the advantages that on the basis of remaining an internal structure of data, the calculation complexity is reduced; and the speaker verification efficiency is improved.

Description

technical field [0001] The present invention relates to the technical field of speaker recognition, in particular to a speaker confirmation method and system based on tensor structure and sparse representation. Background technique [0002] The task of speaker recognition is to identify the speaker. Speaker recognition can be divided into speaker identification and speaker confirmation. For speaker identification, it is to find the correct speaker from multiple speakers. "One-on-one" questions. Speaker verification is the process of extracting personality characteristics from the interlocutor's voice, establishing a recognition model, and verifying the identity of the interlocutor. It is an important branch of speaker recognition research. In speaker confirmation, the extraction of effective feature parameters and the establishment of high-performance recognition models are the key. [0003] Support vector machine (Support Vector Machine, SVM) and Gaussian Mixture Model ...

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): G10L17/06G10L17/02
CPCG10L17/02G10L17/06
Inventor 简志华郭珊徐剑金易帆
Owner HANGZHOU DIANZI UNIV
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
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