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

Bidirectional speech recognition processing system and method

A speech recognition and processing system technology, applied in speech recognition, speech analysis, instruments, etc., can solve the problems of long speech recognizer cycle and many resources

Inactive Publication Date: 2010-02-10
INST OF ACOUSTICS CHINESE ACAD OF SCI +1
View PDF0 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

The recognition errors of speech recognizers are usually caused by model errors and decoding errors. The multi-knowledge source speech recognizer fusion method can reduce model errors to a certain extent, but the construction of speech recognizers based on multi-knowledge sources takes a long time. , requiring more resources

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
  • Bidirectional speech recognition processing system and method
  • Bidirectional speech recognition processing system and method
  • Bidirectional speech recognition processing system and method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0063] The present invention will be described in detail below in conjunction with specific embodiments and accompanying drawings.

[0064] This embodiment is based on figure 2 A two-way speech recognition system construction block diagram provides a two-way speech recognition processing system, including: (as image 3 )

[0065] A feature extractor is used to sequentially extract the feature information of the speech signal input by the user from front to back, and send the feature information of the speech signal to the forward speech recognizer and the backward speech recognizer frame by frame.

[0066] A forward speech recognizer, which is used to decode the feature information of the speech signal frame by frame according to the forward speech model to obtain the forward recognition text and the forward recognition word map, and the forward recognition text and the forward recognition word map Send it to the recognition result fuser.

[0067] A backward speech recogni...

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 relates to a bidirectional speech recognition processing system and a method. The system comprises a characteristic extractor, a forward speech recognizer, a backward speech recognizer,a recognition result reverser and a recognition result merger. The system and the method finish a framework by bidirectional speech recognition decoding based on the same knowledge source, respectively obtain corresponding speech request recognition text in a forward speech recognition process and a backward speech recognition process and merge and process speech request recognition results in different decoding directions to obtain a text request at the front end of a speech search system. The bidirectional speech recognition processing system and the method merge bidirectional decoding results under the prior limited knowledge source to obtain more accurate recognition text. Compared with the prior forward decoding method, the method effectively improves the speech recognition rate and enables a search result of the speech search system to be more accurate and effective.

Description

technical field [0001] The invention relates to the field of voice search systems, in particular to a voice recognition processing system and method. Background technique [0002] Today's world is in the era of information explosion, and information is increasing at an exponential rate. The continuous development of Internet technology and the emergence of efficient text search engines, such as Google and Baidu, provide search services so that people can obtain the information they need most in the shortest time. In the absence of keyboards, handwriting boards and other text input means, in order to provide people with information query services more conveniently, the voice search system enables people to obtain the information they need to query only through voice input for information request services. [0003] The main framework of the voice search system in the prior art is as figure 1 As shown, the main technologies that affect the performance of the voice search syst...

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): G10L15/28G10L15/30
Inventor 颜永红黎塔赵庆卫潘接林
Owner INST OF ACOUSTICS CHINESE ACAD OF SCI
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