Voice change detection method and system, mobile terminal and storage medium

A detection method and technology for speech detection, applied in speech analysis, instruments, etc., can solve problems such as low detection efficiency and poor detection accuracy, and achieve the effects of improving accuracy, reducing computational complexity, and improving resolution

Inactive Publication Date: 2020-02-14
XIAMEN KUAISHANGTONG TECH CORP LTD
View PDF8 Cites 9 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of the embodiments of the present invention is to provide a voice change detection method, system, mobile terminal and storage medium, aiming to solve the problems cau

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
  • Voice change detection method and system, mobile terminal and storage medium
  • Voice change detection method and system, mobile terminal and storage medium
  • Voice change detection method and system, mobile terminal and storage medium

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0040] see figure 1 , is a flow chart of the speech change detection method provided by the first embodiment of the present invention, including steps:

[0041] Step S10, acquiring sample speech data, and performing feature extraction on the sample speech data to obtain cqt speech features;

[0042] Wherein, the sample voice data includes positive sample data and negative sample data, specifically, the positive sample data is mainly the voice data of a real person, and the negative sample data is mainly voice change data, recording playback data and synthetic audio data, etc.;

[0043] Preferably, the voice-changing data can be collected by some mainstream voice-changing apps, and the sound in the audio can be converted into the voice of a specific person through a relevant conversion algorithm. The recording playback data can be collected by some recording equipment, In addition, the synthesized voice data can also be generated through any voice interface;

[0044] Step S20...

Embodiment 2

[0050] see figure 2, is a flow chart of the speech change detection method provided by the second embodiment of the present invention, including steps:

[0051] Step S11, obtaining sample speech data, and performing feature extraction on the sample speech data to obtain cqt speech features;

[0052] Wherein, the sample voice data includes positive sample data and negative sample data. Preferably, since human voices are mainly concentrated in low frequencies, they have higher resolution for low frequencies and lower resolution for high frequencies. Therefore, this step Through the extraction based on cqt features, the model obtained after subsequent training can better distinguish the difference between the altered voice and the normal voice, and can also reduce the amount of data calculation;

[0053] Step S21, performing rate-spectrum conversion on the cqt speech features to obtain a speech power spectrum, and obtaining the logarithm of the speech power spectrum;

[0054] ...

Embodiment 3

[0069] see image 3 , is a schematic structural diagram of the speech change detection system 100 provided by the third embodiment of the present invention, including: a feature extraction module 10, a model training module 11 and a speech detection module 12, wherein:

[0070] The feature extraction module 10 is configured to acquire sample speech data, and perform feature extraction on the sample speech data to obtain cqt speech features, and the sample speech data includes positive sample data and negative sample data.

[0071] The model training module 11 is used to optimize the cqt speech features to obtain the cqcc speech features, and input the cqcc speech features to a preset convolutional neural network for model training to obtain a speech detection model, wherein, The preset convolutional neural network includes three convolutional layers and two fully connected layers.

[0072] Wherein, the model training module 11 is also used to: control the preset convolutional...

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 is applicable to the technical field of automatic speaker verification, and provides a voice change detection method and system, a mobile terminal and a storage medium. The method comprises the following steps of acquiring sample voice data, and carrying out feature extraction on the sample voice data to obtain a cqt voice feature; carrying out optimization processing on the cqt voice feature to obtain a cqcc voice feature, and inputting the cqcc voice feature into a preset convolutional neural network for model training in order to obtain a voice detection model; and acquiring to-be-detected voice, inputting the to-be-detected voice into the voice detection model for voice analysis, and carrying out voice change judgement on the to-be-detected voice according to an analysisresult of the voice detection model. According to the voice change detection method and system, the mobile terminal and the storage medium, the manual feature selection is not needed, the model training is carried out by adopting a convolutional neural network based mode, the accuracy of subsequent voice change detection for the to-be-detected voice is improved, and the resolution of the voice detection model is improved through extraction and optimization based on the cqt feature.

Description

technical field [0001] The invention belongs to the technical field of automatic speaker verification, and in particular relates to a voice change detection method, system, mobile terminal and storage medium. Background technique [0002] Over the years, Automatic Speaker Verification (ASV) technology has matured as a low-cost, reliable method of authentication and identification. However, like all biometric modes, the technology can be vulnerable to some spoofed voice attacks, such as replayed voices, voice changers, synthetic voice attacks, etc. The purpose of using these types of voices is to impersonate other registrants and then break through the verification system, and then perform some illegal operations. Therefore, in the process of using ASV technology, the step of pronunciation change detection of the voice to be tested is particularly important. [0003] The existing voice change detection methods all need to manually select the sound wave features, and then use...

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): G10L17/02G10L17/04G10L17/18G10L17/00
CPCG10L17/02G10L17/04G10L17/18G10L17/00
Inventor 陈文敏肖龙源李稀敏蔡振华刘晓葳王静
Owner XIAMEN KUAISHANGTONG TECH CORP 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