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

Speaker identification method base on simple direct tolerance learning algorithm

A metric learning and speaker technology, applied in the field of speaker recognition, can solve the problems of large correlation and redundancy, and achieve the effect of easy acquisition, good recognition effect and fast speed

Inactive Publication Date: 2016-09-07
JIANGXI NORMAL UNIV
View PDF1 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

As the dimension of data increases, there is often a large correlation and redundancy between these high-dimensional data

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 identification method base on simple direct tolerance learning algorithm
  • Speaker identification method base on simple direct tolerance learning algorithm
  • Speaker identification method base on simple direct tolerance learning algorithm

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0043] A speaker recognition method based on a simple direct metric learning algorithm according to an embodiment of the present invention will be described in detail below with reference to the accompanying drawings. refer to figure 1 , figure 1 A flowchart of an embodiment of the method of the present invention is shown, the method includes the following steps:

[0044] In step S110, collect the speech samples of a plurality of speakers, and extract the i-vector in all samples;

[0045] In step S120, LDA or WCCN method is used to perform channel compensation to process the i-vectors in all samples, and perform length regularization to form a training sample set;

[0046] In step S130, construct the i-vector based on the training sample set and the similar sample pair set and the non-similar sample pair set of speaker identity;

[0047]In step S140, the KISS algorithm is used to train the similar sample pair set and the non-similar sample pair set to obtain the metric matr...

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 provides a speaker identification method base on a simple direct tolerance learning algorithm. The method comprises the following steps: acquiring voice samples of multiple speakers, extracting i-vectors of all the samples, performing channel compensation processing by use of an LDA or WCCN method, performing length normalizing, and forming a training sample set; according to the i-vectors of the training sample set and speaker identity, constructing a similar sample pair set and a non-similar sample pair set; by use of a KISS algorithm, obtaining a tolerance matrix by performing training on the similar sample pair set and the non-similar sample pair set; and for two pieces of new voice, their i-vectors are extracted firstly, the channel compensation processing is carried out by use of the LDA or WCCN method, the length normalizing is performed, by use of the previously calculated tolerance matrix, a Mahalanobis distance between the two i-vectors is calculated and compared with a threshold, and thus whether the two pieces of new voice belong to the same speaker is determined. According to the invention, the obtained Mahalanobis distance tolerance matrix can better truly reflect similarities and distinctions of a sample space so as to improve the performance of a speaker identification system.

Description

technical field [0001] The invention is a speaker recognition method based on a simple and direct metric learning algorithm, which can be widely used in speaker recognition, pattern recognition, metric learning, machine learning and other fields. Background technique [0002] Speaker recognition (Speaker Recognition, SR), also known as voiceprint recognition, is a technology that identifies the speaker's identity by processing and analyzing the speaker's voice. How to effectively measure the similarity between speaker speech samples is one of the hot issues in the field of speaker recognition research. In the field of pattern recognition, there are many methods to measure the similarity between samples. The more commonly used methods are distance scoring methods, such as cosine distance scoring and Mahalanobis distance scoring. [0003] The cosine distance scoring method measures the similarity between samples by calculating the cosine value of the included angle in the inn...

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): G10L17/00G10L17/04
CPCG10L17/00G10L17/04
Inventor 雷震春杨印根朱明华
Owner JIANGXI NORMAL UNIV
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