Voiceprint recognition method and device, electronic equipment and storage medium

A voiceprint recognition and to-be-recognized technology, applied in the field of identity recognition, can solve problems such as noise, low accuracy of recognition results, and impact of recognition on results, and achieve the effect of improving accuracy and avoiding poor feature selection.

Pending Publication Date: 2020-12-08
BEIJING SANKUAI ONLINE TECH CO LTD
View PDF0 Cites 14 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] Since the low-dimensional vector iVector contains both speaker information and channel information, even if PLDA channel compensation is performed, it still contains noise and background sound, which still have a great impact on the recognition results, resulting in low accuracy of recognition results

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
  • Voiceprint recognition method and device, electronic equipment and storage medium
  • Voiceprint recognition method and device, electronic equipment and storage medium
  • Voiceprint recognition method and device, electronic equipment and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0051] A sound pattern recognition method provided in the present embodiment, as figure 1 , The method comprising: a step 110 to step 160.

[0052] Step 110, obtaining spectrum information of the speech to be recognized.

[0053]It may be the speech to be recognized STFT (Short-Time Fourier Transform, STFT) process, obtaining spectrum information of the speech to be recognized; alternatively, also be speech to be recognized corresponding to Mel-frequency cepstral coefficients ( Mel Frequency Cepstrum Coefficient, MFCC) as the spectral information of the speech to be recognized.

[0054] In one embodiment of the present application, the obtaining spectral information to be speech recognition, comprising:

[0055] Speech to be recognized for the short time Fourier transform process, to obtain spectral information of the speech to be recognized; or

[0056] Calculating the speech to be recognized corresponding to Mel-frequency cepstral coefficients as spectral information of the spee...

Embodiment 2

[0086] A sound pattern recognition device provided in the present embodiment, as figure 2 As shown in the voiceprint identification apparatus 200 comprising:

[0087] First spectral information obtaining module 210, configured to obtain spectrum information of the speech to be recognized;

[0088] Speech segment identification module 220 for the speech segment based on the spectrum information, identifying the speech piece to be valid and invalid voice recognition;

[0089] Valid voice stitching module 230, for removing the invalid voice segment, and splice the active voice segment, the effective voice;

[0090] Second spectral information obtaining module 240, configured to obtain the active speech spectral information;

[0091] Feature extraction module 250 for extracting depth-based model by the features of convolutional neural networks, spectrum information of the active speech feature extraction to obtain a feature vector to be recognized voiceprint corresponds to the speech ...

Embodiment 3

[0107] The embodiment of this application also provides an electronic device, such as image 3 , The electronic device 300 may include one or more processors 310 and one or more memories 310 coupled to the processor 320. The electronic device 300 may further include an input interface 330 and an output interface 340 for communicating with another device or system. Program code executed by the processor 310 may be stored in memory 320.

[0108] Call processor 310 stores the electronic device 300 in the program code memory 320 to perform the voiceprint identification method of the above-described embodiments.

[0109] The above elements in the electronic device can be connected to each other, and the bus, an address bus, a control bus, an extended bus, and a partial bus, or any combination thereof.

[0110] Example embodiments of the present application also provides a computer-readable storage medium, having stored thereon a computer program implemented steps as described herein in ...

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

Embodiments of the invention disclose a voiceprint recognition method and device, electronic equipment and a storage medium. The method comprises the steps of obtaining frequency spectrum informationof a to-be-recognized voice; recognizing valid voice segments and invalid voice segments in the to-be-recognized voice according to the frequency spectrum information; removing the invalid voice segments, and splicing the valid voice segments to obtain a valid voice; acquiring frequency spectrum information of the valid voice; performing feature extraction on the frequency spectrum information ofthe valid voice through a feature extraction model based on a deep convolutional neural network to obtain a to-be-recognized voiceprint eigenvector corresponding to the to-be-recognized voice; and performing similarity calculation on the to-be-recognized voiceprint eigenvector and existing voiceprint eigenvectors in a voice feature library, and determining speaker identity information corresponding to the to-be-recognized voice. According to the embodiment of the invention, the invalid voice segments are removed, so that high-quality voice data are provided for the feature extraction model, and the accuracy of a voiceprint recognition result is improved.

Description

Technical field [0001] Example embodiments relate to identification technology, and more particularly, to a voiceprint recognition method, apparatus and storage medium of the electronic apparatus of the present application. Background technique [0002] Voiceprint identification, also known as speaker recognition, according to a speaker's voice characteristics to identify the identity of the speaker's biometric technology can be widely used in security, finance, fraud and other fields. [0003] Currently, the most widely used voiceprint identification method is iVector / PLDA algorithm. Which process: the voice frequency spectrum information obtained by MFCC (Mel-Frequency Cepstral Coefficients, mel-frequency cepstral coefficients); super-Gaussian vector factor analysis, MFCC obtained high dimensional feature information is mapped to low dimensional vectors iVector, the low dimensional vectors iVector comprises a speaker voiceprint information and channel information; using PLDA ...

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): G10L17/02G10L17/18G10L25/24
CPCG10L17/02G10L17/18G10L25/24
Inventor 邹佳宏梁延峰
Owner BEIJING SANKUAI ONLINE TECH CO LTD
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

Browse by: Latest US Patents, China's latest patents, Technical Efficacy Thesaurus, Application Domain, Technology Topic.

© 2024 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap