Method and system for training voiceprint recognition model

A technology for voiceprint recognition and recognition models, which is applied in the field of training voiceprint models, can solve problems such as time-consuming and economical costs, and achieve the effect of improving performance

Active Publication Date: 2018-01-19
BAIDU ONLINE NETWORK TECH (BEIJIBG) CO LTD
View PDF5 Cites 31 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

In order to develop a voiceprint authentication algorithm and products for a characteristic scene, an end-to-end voiceprint authentication algorithm often needs to record a large amount of data to ensure the performance of the algorithm, which requires a lot of time and economic cost

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
  • Method and system for training voiceprint recognition model
  • Method and system for training voiceprint recognition model
  • Method and system for training voiceprint recognition model

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach

[0104] According to an implementation manner of an embodiment of the present invention, the voiceprint registration includes:

[0105] S1, receiving the voiceprint registration request sent by the user,

[0106] For example, the voiceprint registration request includes the user identification and the voice provided in the form of English, text or digital string-related voice.

[0107] S2. Obtain the posterior probability corresponding to the voice in the voiceprint registration request according to the second voiceprint recognition model.

[0108] S3. Extract a feature vector of the speech according to the posterior probability.

[0109] For example, each posterior probability is normalized according to the unified background model, and the feature vector of the speech is extracted according to the normalized posterior probability by using a feature vector extraction model.

[0110] S4. Acquire a voiceprint registration model of the user according to the feature vector, wher...

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 provides a method and system for training a voiceprint recognition model, and the method comprises the steps: training a depth neural network through first training data, and generatinga first voiceprint recognition model; carrying out the transfer learning of the first voiceprint recognition model through second training data, and generating a second voiceprint recognition model, wherein the second voiceprint recognition model is used for voiceprint registration or authentication. The method can solve a problem in the prior art that a large amount of time cost and economic costis needed for guaranteeing the performances of an algorithm because a large amount of data needs to be recorded for the end-to-end voiceprint authentication algorithm in order to develop the voiceprint authentication algorithm and a product for a characteristic scene. The method can make the most of the whole training data, and improves the voiceprint recognition and authentication performances of the specific data through the transfer learning.

Description

【Technical field】 [0001] The present application relates to the field of artificial intelligence, in particular to a method and system for training a voiceprint model. 【Background technique】 [0002] Artificial Intelligence (AI) is a new technical science that studies and develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence. Artificial intelligence is a branch of computer science that attempts to understand the essence of intelligence and produce a new intelligent machine that responds in a manner similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing and expert systems, etc. [0003] With the continuous development of artificial intelligence, voice interaction is increasingly promoted as the most natural way of interaction, and people's demand for voice recognition services is increasing. According to the user's voice ...

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): G10L17/04G10L17/18G06N3/08
Inventor 马啸空李超蒋兵
Owner BAIDU ONLINE NETWORK TECH (BEIJIBG) 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