Unlock instant, AI-driven research and patent intelligence for your innovation.

Method for identifying speaker based on identification principle of bionic mode

A bionic pattern recognition and speaker recognition technology, applied in the field of speaker recognition

Inactive Publication Date: 2007-07-25
INST OF SEMICONDUCTORS - CHINESE ACAD OF SCI
View PDF3 Cites 8 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0007] The purpose of the present invention is to provide a speaker recognition method based on the principle of bionic pattern recognition for the problems existing in the existing methods, which has an excellent recognition effect and effectively solves the misrecognition caused by untrained speakers participating in the test, and adds The new speaker does not need to retrain the original system, which improves the network performance

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 for identifying speaker based on identification principle of bionic mode
  • Method for identifying speaker based on identification principle of bionic mode
  • Method for identifying speaker based on identification principle of bionic mode

Examples

Experimental program
Comparison scheme
Effect test

Embodiment

[0037] The experimental samples were collected from 25 speakers (14 males and 11 females), and each person said "one, two, three, four" (named as type A sample) and "7890" (named as type B sample) 5 times each , a total of 250 samples, sampling rate 22.05kHz, mono 16-bit precision.

[0038]In the feature extraction process, the present embodiment adopts MFCC (Mel-FrequencyCepstrum Coefficients) feature parameters as the speech frame feature of speaker recognition, which utilizes the decorrelation characteristics of the auditory principle and cepstrum, and also has the ability to correct convolutional channel distortion. ability to compensate. After endpoint detection and pre-emphasis on the voice signal, 25ms is taken as a frame, the relative displacement between two adjacent frames is 12.5ms, and 12 MFCC parameters are extracted for each frame, and then normalized cepstrum lifting is performed again to obtain A 12-dimensional feature vector.

[0039] In the process of const...

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

A method for identifying a speaking person based on bionic mode identification principle includes collecting voice signal then carrying out end-point detection and pre-weighting treatment to obtain voice sample, structuring identification network of speaking person based on bionic mode identification principle by utilizing network matching degree and carrying out identification on sample to be identified according to identification policy.

Description

technical field [0001] The invention belongs to the field of speaker recognition, in particular to a speaker recognition method based on the principle of bionic pattern recognition. Background technique [0002] Automatic speaker recognition (Automatic Speaker Recognition) is to use the speaker's voice features to identify or confirm the speaker's identity. [0003] Due to the excellent complex mapping ability of artificial neural network, it has been successfully applied in speaker recognition. However, the traditional neural network method for speaker recognition uses a multilayer perceptron (MLP, multilayer perceptron) based on the idea of ​​division or a radial basis function (RBF, radial basis function) network based on the idea of ​​template matching. The handling of the sample distribution in space is relatively rough, so there are obvious limitations. For example, the appearance of a new speaker in the test will inevitably lead to misrecognition, and adding a new s...

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): G10L17/00G10L15/00G10L15/02G10L15/04G10L15/06G10L15/08G10L25/87
Inventor 王守觉黄翼曹瑜
Owner INST OF SEMICONDUCTORS - CHINESE ACAD OF SCI