A speaker recognition method and device

A speaker recognition and speaker technology, applied in the field of speaker recognition methods and devices, can solve problems such as poor speaker recognition test performance, and achieve the effect of improving system performance

Active Publication Date: 2017-09-15
SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
View PDF10 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The purpose of the embodiments of the present invention is to provide a speaker recognition method based on JFA speaker supervector non-parametric distinction, to solve the prior art based on the joint factor analysis method of GMM-UBM model, to test the performance of speaker recognition bad question

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
  • A speaker recognition method and device
  • A speaker recognition method and device
  • A speaker recognition method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0028] In order to make the object, technical solution and advantages of the present invention clearer, the present invention will be further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific examples described here are only used to explain the present invention, not to limit the present invention.

[0029] In view of the joint factor analysis JFA theory acting on the Gaussian mixture model-universal background model GMM-UBM system can achieve a good mismatch compensation effect, so this application will apply the mismatch compensation method based on joint factor analysis to the non-parametric Discriminative Analysis Techniques in Speaker Verification Systems. We extract the speaker’s GMM supervector that removes the influence of part of the channel information, that is, the JFA speaker supervector, as a new feature expression representing speaker information. In this way, each speaker’s speech o...

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 present invention is applicable to the field of speaker recognition, and provides a speaker recognition method and device, the method comprising: extracting a JFA speaker supervector, and the JFA speaker supervector is a Gaussian that removes the influence of channel information through a joint factor analysis method Mixed model GMM supervector; segment the extracted JFA speaker supervector, and divide the JFA speaker supervector into multiple subvector sets; according to the multiple subvector sets obtained by segmentation, perform non- Parameter discrimination analysis to establish a subspace speaker model; obtain the reference vector of the speaker to be identified and the reference vector of the training sample speaker according to the subspace speaker model to perform speaker identification. Compared with the prior art, the invention has a small calculation amount; and the test data collection is simple, and the recognition performance is improved.

Description

technical field [0001] The invention belongs to the field of speech recognition, in particular to a speaker recognition method and device. Background technique [0002] With the continuous advancement of computer technology and Internet technology, smart devices have become more and more indispensable in people's lives. As one of the interaction methods between humans and smart devices, voice interaction has also become a hot research field because of its characteristics such as easy collection, easy storage, difficult imitation, and low cost of voice acquisition. [0003] The current intelligent speech processing methods are mainly divided into speech recognition (Speech Recognition), language recognition (Language Recognition) and speaker recognition (Speaker Recognition) according to the difference of the speech information used. Among them, the goal of speech recognition is to determine what kind of semantic information is conveyed in the speech signal; the goal of lang...

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): G10L17/04G10L17/02
Inventor 李志锋李娜乔宇
Owner SHENZHEN INST OF ADVANCED TECH CHINESE ACAD OF SCI
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
Try Eureka
PatSnap group products

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

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap