End-to-end voice intention recognition method

A recognition method and voice technology, applied in voice recognition, voice analysis, instruments, etc., can solve problems such as loss of user intentions in text information transmission, and achieve the effects of avoiding processing difficulties, improving accuracy, and simplifying the construction process

Pending Publication Date: 2020-04-28
NANJING SILICON INTELLIGENCE TECH CO LTD
View PDF6 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

When the correct rate of speech-to-text information is high, the accuracy of intent recognition is high; when the recognition rate of text information is low, a large amount of useful information is discarded by speech recognition during the recognition process, resulting in the user's intent of text information transmission also varies. lost

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
  • End-to-end voice intention recognition method
  • End-to-end voice intention recognition method
  • End-to-end voice intention recognition method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0031] Reference figure 1 The invention discloses an end-to-end speech intention recognition method, which includes the following steps:

[0032] S1. Input the voice to be recognized, perform noise reduction and feature extraction processing on it, and convert the voice to be recognized into a feature vector containing sound information;

[0033] S2. Input the feature vector into the speech intention recognition model, and output the speech intention.

[0034] In step S1, the noise reduction and feature extraction processing of the input speech includes a preprocessing process and a feature extraction process. The preprocessing process first cuts off the mute at the beginning and the end to reduce the interference caused to subsequent steps. The operation of mute cut is generally called Voice Activity Detection (Voice Activity Detection, VAD). Then the sound is divided into frames, that is, the sound is cut into small segments, and each small segment is called a frame, which is real...

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 discloses an end-to-end voice intention recognition method, and relates to the technical field of voice intention recognition. Most of existing voice intention recognition applications firstly acquire texts through voice recognition and then perform intention recognition, the accuracy of the intention recognition mode based on the texts seriously depends on the accuracy of translating the texts through voice recognition, and texts and pictures with inaccurate voice intention recognition exist. In order to solve the problem, the key points of the technical scheme are that the method comprises the steps: inputting a to-be-recognized voice, carrying out the noise reduction and feature extraction of the to-be-recognized voice, converting the to-be-recognized voice into a featurevector containing sound information, inputting the feature vector into a voice intention recognition model, and outputting a voice intention. The voice intention recognition model adopts a pre-training model thought of a deep learning network. According to the invention, the effects of reducing information loss caused by voice recognition and improving the voice intention recognition accuracy areachieved.

Description

Technical field [0001] The present invention relates to the technical field of speech intention recognition, and in particular to an end-to-end speech intention recognition method. Background technique [0002] With the rapid development of artificial intelligence technology in academia and its widespread use in life, voice interaction has become an important bridge for communication between humans and machines. The robot system needs to talk to the user and complete specific tasks. One of the core technologies is the determination of voice intent, that is, after the robot system receives the user's voice, it can determine the user's intent through the voice. [0003] Voice intent recognition technology refers to the recognition of the corresponding intent or feature of the input voice (the intent here includes single intent, multiple intent, slot value, emotion and other types of questions), and provides effective support for specific back-end service goals , High-performance voi...

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): G10L15/02G10L15/06G10L15/16G10L15/26G10L21/02
CPCG10L15/02G10L15/063G10L15/16G10L15/26G10L21/02G10L2015/025
Inventor 司马华鹏汤毅平
Owner NANJING SILICON INTELLIGENCE 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