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

Speaker recognition method based on speech feature fusion and GMMs

A speaker recognition and speech feature technology, applied in speech analysis, instruments, etc., can solve the problems of unstable speaker recognition system and low speaker recognition accuracy

Active Publication Date: 2019-11-05
HANGZHOU DIANZI UNIV
View PDF9 Cites 2 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0003] The purpose of the present invention is to provide a speaker recognition method based on speech feature fusion and GMM in order to solve the technical problems that the accuracy rate of speaker recognition using a single speech parameter is not high and the speaker recognition system is unstable in the prior art

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 recognition method based on speech feature fusion and GMMs
  • Speaker recognition method based on speech feature fusion and GMMs
  • Speaker recognition method based on speech feature fusion and GMMs

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0032] The present invention will be described in further detail below in conjunction with the accompanying drawings.

[0033] The present invention proposes a speaker recognition method based on speech feature fusion and GMM, and its overall block diagram is as follows: figure 1 As shown, it specifically includes the following five steps:

[0034] Step (1) Preprocessing of the original speech signal. The specific operation is as follows:

[0035] 1. Voice Segmentation: Use audio processing software to divide the 15-20 second long voice into 8 equal-length short voices. 2. Pre-emphasis: The pre-emphasis of the voice signal is realized by means of a first-order high-pass filter. The high-pass filter can filter out the low-frequency part of the signal and improve the frequency spectrum of the high-frequency part of the signal. Its transfer function is expressed as: H(z)=1-αz -1 , α is the pre-emphasis coefficient, generally between 0.9-1. 3. Frame division and windowing: Si...

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 relates to a speaker recognition method based on speech feature fusion and GMMs (Gaussian Mixture Models). According to the method, first, different types of speech feature parameters are fused to acquire new speech feature parameters, and the new feature parameters are used for training of a recognition model; and in the establishing process of the recognition model, the two GMMs are connected in series through a threshold function, so that the recognition model is optimized. Through the method, the new speech feature parameters are extracted and have respective advantages of multiple types of speech feature parameters, and personalized features of a speaker can be better represented; and moreover, the two GMMs are connected in series through the threshold function for recognition, so that the accuracy of speaker recognition is improved compared with a single GMM.

Description

technical field [0001] The invention belongs to the technical field of speech recognition, in particular to a speaker recognition method based on speech feature fusion and GMM. Background technique [0002] With the advent of the information and digital age, people's demand for highly reliable identity authentication technology is also increasing. Traditional identity authentication technology usually adopts password authentication method, which is seriously insufficient in security. Now, with the rapid development of biotechnology and information technology, speaker authentication technology is a more advanced and convenient information security technology. Began to be widely used in people's life. Due to the different congenital physiological differences of each person's vocal organs (such as the shape of the vocal cords) and acquired behavioral differences such as differences in speaking habits, each person's voice is different. These differences exist in some form in t...

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): G10L17/02G10L17/10G10L17/06G10L17/04
CPCG10L17/02G10L17/04G10L17/06G10L17/10
Inventor 张桦戴美想戴国骏吴以凡沈方瑶商吉利
Owner HANGZHOU DIANZI UNIV