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

A Voiceprint Recognition Method Based on RNN

A technology of voiceprint recognition and subtraction method, which is applied in speech analysis, instruments, etc., can solve the problems of not considering the context correlation of speech data, not using deep learning to extract features, and features not being able to represent the speaker well, etc. To achieve the effect of improving the recognition performance

Active Publication Date: 2021-02-09
王华锋
View PDF14 Cites 0 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0009] The technical problem solved by the present invention is: overcoming the context correlation of voice data that is not considered in the existing voiceprint recognition method, the extracted features cannot represent the speaker well, and the powerful feature extraction ability of deep learning has not been brought into play And other issues

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 Voiceprint Recognition Method Based on RNN
  • A Voiceprint Recognition Method Based on RNN
  • A Voiceprint Recognition Method Based on RNN

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0035] figure 1 The overall processing flow of the present invention is given. The present invention provides a method for voiceprint recognition based on RNN. The main steps are as follows: firstly, denoise the input voice data using spectral subtraction, and extract the MFCC characteristic parameters and first and second order differences of the obtained pure voice data by frame , and then concatenate the MFCC parameters and the first and second order differences of multiple consecutive frames to obtain a two-dimensional feature parameter matrix, which is used as the input of the cyclic neural network. The present invention uses the variant LSTM of the cyclic neural network for training, and the LSTM passes through a unique gate mechanism-a forget gate, an input gate and an output gate ( figure 1 The δ in is the sigmoid activation function, tanh is the hyperbolic tangent activation function) to retain the long-term information of the sequence, and thus complete the training...

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 present invention provides a voiceprint recognition method based on RNN. After obtaining the MFCC feature of the denoising voice data and its first and second order differences, the cyclic neural network is used to extract the advanced features of the speaker in the MFCC feature, and the extracted features are classified using a softmax classifier, and finally a naive Bayesian method is used to identify speakers. Different from the traditional method of silence elimination, this method retains the silent segment in the speech data, and can extract context-related features based on the cyclic neural network, and can extract the advanced features of the speaker's voice such as speaking style and rhythm for the speech data. , making the feature information more complete and more able to represent the speaker. Compared with the existing Gaussian-based voiceprint recognition method, this method has relatively lower requirements for voice data and higher accuracy. Even in the face of large data, the accuracy rate remains at a high level, and the running speed does not decrease.

Description

technical field [0001] The invention provides an RNN-based voiceprint recognition method, which relates to the fields of deep learning, pattern recognition, and voice signal processing. Background technique [0002] With the rapid development of information technology, how to accurately authenticate a person's identity, protect personal privacy and ensure information security has become an urgent problem to be solved. Compared with traditional identity authentication methods, biometric identity authentication technology has the characteristics that it will not be lost, stolen or forgotten during use; it is not only fast, convenient, but also accurate and reliable. As one of the most popular biometric identification technologies at present, voiceprint recognition has unique advantages in remote authentication and other application fields, and has attracted more and more attention. WeChat has enabled the voice lock verification login method, the world's first voice lock Lenov...

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 Patents(China)
IPC IPC(8): G10L17/02G10L17/04G10L17/18G10L17/14G10L25/24G10L25/18G10L21/0208
CPCG10L17/02G10L17/04G10L17/14G10L17/18G10L21/0208G10L25/18G10L25/24
Inventor 冯毅夫王华锋徐雷杜俊逸付明霞马晨南齐一凡潘海侠
Owner 王华锋
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