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

Voice recognition result screening method and device

A technology for speech recognition and screening methods, applied in speech recognition, speech analysis, instruments, etc., can solve problems such as low efficiency and rely on human labor, and achieve the effect of solving low efficiency

Active Publication Date: 2016-11-23
ZHEJIANG TMALL TECH CO LTD
View PDF7 Cites 3 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, the current process of recognizing voice data and selecting and labeling data is an activity that relies on human labor, and the efficiency is very low.

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 recognition result screening method and device
  • Voice recognition result screening method and device
  • Voice recognition result screening method and device

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0018] Embodiment 1 provides a method for screening speech recognition results, which is used to improve the efficiency of screening speech recognition results. The specific flow chart of the method is shown in figure 1 shown, including the following steps:

[0019] Step 11, using the speech recognition system to recognize the speech data, and obtain text information corresponding to the speech data;

[0020] Among them, in the speech data recognition process, one or more non-local recognition systems in different fields can be used according to different needs and purposes to obtain the recognition result, that is, the text information corresponding to the speech data. For example, if it is necessary to expand the labeling database of medicine in the local acoustic model, it is necessary to obtain the voice data containing the name of the medicine first, and then use one or more voice recognition systems with the ability to recognize medical terminology to recognize the voic...

Embodiment 2

[0042] Based on the same inventive concept, Embodiment 2 provides a device for screening speech recognition results, which is used to improve the efficiency of screening speech recognition results. like figure 2 As shown, the screening device of the speech recognition result comprises:

[0043] The recognition unit 21 can be used to recognize the voice data by using a voice recognition system to obtain text information corresponding to the voice data;

[0044] The calculation unit 22 can be used to determine the confidence level of the information unit contained in the text information;

[0045] The selection unit 23 may be configured to select information units satisfying the confidence condition according to the confidence degree.

[0046] In one embodiment, the device also includes:

[0047] The determining unit 24 may be configured to determine the selected information unit as the annotation data corresponding to the speech unit included in the speech data.

[0048] I...

Embodiment 3

[0067] Based on the same inventive concept, Embodiment 3 provides a method for supplementing the annotation data of the local acoustic model, which is used to supplement the annotation data of residential quarters in the Beijing area, and provide navigation software with a better service for identifying the user's destination. A schematic diagram of the method is shown in image 3 shown, including the following steps:

[0068] Step 31, acquiring the voice data of the navigation software client.

[0069] For example, “yīng’tè’gōng’yù”, “dōng’fēng’jiā’yuán”.

[0070] Step 32, using a speech recognition system capable of recognizing the names of places and features, to recognize the acquired speech data and obtain text information.

[0071] For example, get the text information: "Intel Apartment", "Dongfeng Home".

[0072] Step 33, using the voice recognition system to calculate the confidence level of the text information, the speech recognition system explains the confidence...

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 a voice recognition result screening method and device used for improving voice recognition result screening efficiency. The method includes utilizing a voice recognition system for recognizing voice data and obtaining text information corresponding to the voice data; determining the confidence coefficient of information unit contained in the text information; according to the confidence coefficient, selecting the information unit meeting confidence conditions. The invention also discloses a voice recognition result screening device.

Description

technical field [0001] The present application relates to the technical field of speech recognition, in particular to a method and device for screening speech recognition results. Background technique [0002] A speech recognition system usually consists of two parts: an acoustic model and a language model. The establishment of the acoustic model needs to rely on a large amount of speech data and the correct text information corresponding to the speech data, so as to obtain the statistical correspondence between speech and text, and accumulate speech data and speech. The process of correct text information corresponding to the data is also called model training. The correct text information corresponding to the voice data is defined as the labeled data corresponding to the voice data. For example, the labeled data corresponding to the voice data "bǐ'jì" is " Notes" and "handwriting"; and "Biji" does not have the meaning of words, so it is not labeled data. The more labeled ...

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): G10L25/48G10L15/00
Inventor 陈昊李宏言钱霄
Owner ZHEJIANG TMALL TECH CO LTD
Features
  • Generate Ideas
  • Intellectual Property
  • Life Sciences
  • Materials
  • Tech Scout
Why Patsnap Eureka
  • Unparalleled Data Quality
  • Higher Quality Content
  • 60% Fewer Hallucinations
Social media
Patsnap Eureka Blog
Learn More

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

© 2025 PatSnap. All rights reserved.Legal|Privacy policy|Modern Slavery Act Transparency Statement|Sitemap|About US| Contact US: help@patsnap.com