Method and apparatus for training voiceprint recognition system

A technology of voiceprint recognition and computer system, which is applied in the field of training voiceprint recognition systems, can solve the problems of low accuracy rate and low improvement of voiceprint recognition accuracy rate, and achieve the effect of improving accuracy rate

Active Publication Date: 2017-01-04
TENCENT TECH (SHENZHEN) CO LTD +1
View PDF8 Cites 5 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0005] In order to solve the problem in the prior art that the accuracy of voiceprint recognition is low by using identity vectors processed by LDA, an embodiment of the present invention provides a method and device for training a voiceprint recognition system
[0007] Since the regularization matrix determined by the computer system maximizes the sum of the first numerical values ​​of each category, the identity vectors of the voices of different segments of the same user are improved after being regularized by the regularization matrix, which solves the problem of using LDA in related technologies. The problem of the low degree of improvement in the accuracy of voiceprint recognition for the processed identity vector improves the accuracy of voiceprint recognition

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
  • Method and apparatus for training voiceprint recognition system
  • Method and apparatus for training voiceprint recognition system
  • Method and apparatus for training voiceprint recognition system

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0021] figure 1 It is a flowchart of a method for training a voiceprint recognition system provided in an embodiment of the present invention, such as figure 1 As shown, the method for training the voiceprint recognition system may include the following steps:

[0022] In step 101, the computer system determines the identity vector of each segment of speech in the speech training set, and divides the identity vectors of the same user's speech in the determined identity vector into one category.

[0023] Step 102, the computer system establishes a first function for calculating the first value corresponding to each category, the first value is the first identity vector in the corresponding category after regularization using the regularization matrix and the first identity vector in the corresponding category after regularization using the regularization matrix The sum of similarities between other identity vectors of , the random variable of the first function is a regular ma...

Embodiment 2

[0028] Figure 2A is a flow chart of a method for training a voiceprint recognition system provided in another embodiment of the present invention, as Figure 2A As shown, the method for training the voiceprint recognition system may include the following steps:

[0029] Step 201, the computer system determines the identity vector of each segment of speech in the speech training set, and classifies the identity vectors of the same user's speech in the determined identity vector into one category.

[0030] Generally speaking, for a user, at least two speeches of the user are recorded or collected, and these recorded or collected speeches are added to the speech training set, and the speech training set includes at least two speeches of the user.

[0031] Further, the computer system processes each segment of speech in the speech training set, generates an identity vector indicating the identity information of the person who entered the speech, and divides the identity vector 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 invention discloses a method and apparatus for training voiceprint recognition system, belonging to the technical field of voiceprint recognition. The method comprises a step of determining the identity vector of each section of speech in a speed training set, and dividing the identity vectors of the speech of a same user in the determined identity vectors into a category, a step of establishing a first function for calculating the first value corresponding to each category, wherein the first value is the sum of the similarities of a first identity vector normalized by using a normalization matrix in a corresponding category and other identity vectors normalized by using the normalization matrix in the corresponding category, a step of determining the normalization matrix which allows the sum of the first value of each category to be maximum, and a step of normalizing the identity vector of the speech obtained in the voiceprint recognition system by using the determined normalization matrix. The problem of the low improvement of the accuracy of the voiceprint recognition by using an identity vector which is subjected to linear discrimination analysis processing in the related technology is solved, and the accuracy of the voiceprint recognition is improved.

Description

technical field [0001] The invention relates to the technical field of voiceprint recognition, in particular to a method and device for training a voiceprint recognition system. Background technique [0002] Voiceprint recognition is a kind of biometric recognition technology. By processing the voice, an identity vector indicating the identity information of the voice inputter can be generated. The two voices can be determined by calculating the similarity between the identity vectors of the two voices. Whether the voice inputter is the same user. [0003] Speech is easily disturbed by channel variability and environment variability, resulting in distortion of its identity vector. In related technologies, it is assumed that the distribution of several segments of speech of the same user in space is a multidimensional Gaussian distribution, and the identity vector is processed by using linear discriminant analysis (English: linear discriminant analysis, LDA), which compensat...

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): G10L17/04G10L17/02G10L17/14
CPCG10L17/02G10L17/04G10L17/14G10L17/08G10L17/00G10L17/22
Inventor 李为钱柄桦金星明李科吴富章吴永坚黄飞跃
Owner TENCENT TECH (SHENZHEN) CO LTD
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