A sound fingerprint recognition method based on the vector quantification

A voiceprint recognition and vector quantization technology, which is used in speech analysis, instruments, user identity/authority verification, etc. , to avoid the cavity phenomenon, improve the accuracy and recognition speed, and improve the quality

Inactive Publication Date: 2007-10-17
EAST CHINA NORMAL UNIVERSITY
View PDF0 Cites 7 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the LBG-based vector quantization algorithm has the following disadvantages: 1) In the optimal division stage of each iteration, searching for the nearest codeword of the training vector from the codebook requires a large amount of storage space and cumbersome calculations
Randomly generated initial codebooks are prone to empty blocks (empty cells)
3) Poor adaptability of the codebook
The fast nearest search algorithm FNS is an improvement of the exhaustive search algorithm FS, it does not have the disadvantage of large amount of calculation

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 sound fingerprint recognition method based on the vector quantification
  • A sound fingerprint recognition method based on the vector quantification

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0024] Example 1 The user of the client is an ordinary user

[0025] The method in this embodiment needs to be implemented in an identity authentication system based on voiceprint features composed of an input device, an operating system, a first processor, a local area network, a second processor, a Matlab server, a workflow engine, and an output device. The input device It is a touch screen, the operating system is the commonly used Windows Mobile system, the first processor and the second processor are PCs, the workflow engine is a management system that can parse the workflow definition and execute the workflow, input devices, operating systems and first The processor constitutes the client, the second processor, the Matlab server, the workflow engine, and the output device constitute the server. The client is connected to the server via a local area network. The characteristic is that the implementation process of the method is described as follows:

[0026] Step 1 Use the to...

Embodiment 2

[0035] Embodiment 2 The user of the client terminal is an ordinary user except "admin".

[0036] Except for the following differences, the rest is exactly the same as in Example 1.

[0037] In step 5, another user enters the username "admin". In step 7, the Matlab server uses the FNS algorithm to calculate the distortion value based on the voice file “admin.wav” input by the user, and judges whether it matches the codebook “admin.txt” in the codebook library. The judgment result is a mismatch, and go to 7.1 Steps, so that users cannot log in to the system, thereby effectively preventing others from embezzling user names and passwords, achieving voiceprint recognition and improving system security.

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

An acoustic figure identifying method based on vector quantization belongs to an acoustic figure identifying technic field. In an acoustic figure identifying system realized on an Matlab server, the acoustic figure is identified by using a genration algorithm ACBI of a self-adaptive original code based on vector quantization, and quality of the acoustic figure identification is improved by using a rapid close searching algorithm FNS based on an inequation during a code searching process. The invention has advantages of acoustic figure identification with high accuracy, rapid identification speed, strong self-adaptive performance of the codes, avoided cavum phenomenon, high quality codes, etc.

Description

Technical field [0001] The invention relates to a voiceprint recognition method based on vector quantization, and belongs to the technical field of voiceprint recognition. Background technique [0002] Among the current voiceprint recognition methods, the vector quantization method is a source coding technology developed in the late 1970s, which models users based on the spatial distribution of the speaker’s speech feature parameters. The LBG algorithm is a classic vector quantization codebook design algorithm. It is based on the two conditions of the best division in the best vector quantizer and the best codebook, and has been widely used in the field of statistical clustering research. This algorithm can not only generate codebooks independently, but also can be used as an additional step in other codebook design techniques. However, the LBG-based vector quantization algorithm has the following shortcomings: 1) In the optimal division stage of each iteration...

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): G10L19/00H04L12/28H04L9/32G10L19/038
Inventor 吕钊顾君忠吴悦李卓辉庄杰郭李华张卫祥王峰
Owner EAST CHINA NORMAL UNIVERSITY
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