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

Voiceprint recognition model training method, storage medium and computer equipment

A voiceprint recognition and training method technology, applied in neural learning methods, biological neural network models, speech analysis, etc., can solve problems such as complex channel differences, achieve good learning effects, and improve recall and accuracy.

Active Publication Date: 2021-01-22
SOUTHWEST UNIVERSITY OF POLITICAL SCIENCE AND LAW +1
View PDF6 Cites 11 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0008] Of course, due to the personal information of the speaker such as gender and dialect contained in the voice, it appears as a different frequency distribution on the spectrum, and the channel difference is mainly reflected in the change in the frequency domain, so information such as gender and dialect will make the channel difference more complicated.

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 model training method, storage medium and computer equipment

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0041]Such asfigure 1As shown, a method for training a voiceprint recognition model of the present invention specifically includes the following steps:

[0042]Step 1. Collect the set of speech samples to be trained

[0043]Collect 100,000 speech samples of the two channels to be trained for voiceprint recognition and comparison. One of the channels collects 1 speech sample for each sample object, totaling 100,000 people. Each speech sample collected for this channel needs to be based on the sample Objects are labeled with features such as gender and dialect; the other channel also collects 1 voice sample per sample object, a total of 100,000 people, the collected voice samples do not need to be labeled with the above-mentioned labels, and the sample objects of the channel that need to be labeled are in When selecting, it is necessary to consider the uniform distribution of gender, dialect and other characteristics as much as possible. For channels that do not need to be labeled, when sel...

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

According to the voiceprint recognition model training method, the storage medium and the computer equipment, the linguistic features containing the identity information of the speaker are extracted as the input features, multi-task training is performed by using the gender and other tags of the speaker, and the cross-channel problem is solved in combination with the adversarial training method. Finally, stable features reflecting the identity essence of the speaker are extracted. According to the method, the linguistic characteristics and the deep neural network are combined to simulate the learning mechanism of the human brain, so that the extraction capability, stability and interpretability of the identity essential characteristics of the speaker are improved, and finally, the accuracyand recall rate of automatic voiceprint recognition are improved.

Description

Technical field[0001]The invention relates to the field of automatic voiceprint recognition, in particular to a training method, storage medium and computer equipment of a voiceprint recognition model oriented to judicial voice evidence evaluation mode.Background technique[0002]In the task of speaker identity identification in the field of forensic speech, the current mainstream identification methods in China are based on several dimensions such as watching, listening, and testing, and rely on the personal experience of voiceprint experts. This method is time-consuming, labor-intensive, and contains the subjective judgment of appraisal experts, and cannot be quickly promoted in a larger group of practitioners. In addition, limited by the characteristics of this type of method, it can only be suitable for small-scale inspection materials and sample scenarios. When the inspection materials and samples to be compared are hundreds, thousands or more, voiceprint identification experts a...

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): G10L17/00G10L17/02G10L17/04G10L17/14G10L17/18G10L25/24G06N3/04G06N3/08
CPCG10L17/02G10L17/04G10L17/14G10L17/18G10L25/24G06N3/049G06N3/08G06N3/045
Inventor 张翠玲谭铁君李稀敏杨东升叶志坚肖龙源
Owner SOUTHWEST UNIVERSITY OF POLITICAL SCIENCE AND LAW
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