Phonetic segmentation-based isolate word recognition method

A recognition method and speech segmentation technology, applied in speech recognition, speech analysis, instruments, etc., can solve the problem of short response time of speech recognition performance, and achieve the effect of improving the recognition response time and improving the recognition performance.

Inactive Publication Date: 2010-09-01
NO 709 RES INST OF CHINA SHIPBUILDING IND CORP
View PDF0 Cites 10 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0004] The purpose of this invention is to develop a kind of isolated word recognition method based on speech segmentation that can effectively solve the defects in the existing isolated word speech recognition method, high speech recognition performance, and short recognition response time

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
  • Phonetic segmentation-based isolate word recognition method
  • Phonetic segmentation-based isolate word recognition method
  • Phonetic segmentation-based isolate word recognition method

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0020] according to figure 1 As shown, a method for identifying isolated words based on speech segmentation, the continuous HMM model obtained from speech training uses syllables or semi-syllables as primitives, and the trained primitive models are divided into syllables or semi-syllables of isolated words in the vocabulary The whole word model is obtained by splicing half-syllable sequences, and the Viterbi algorithm is used for recognition;

[0021] The specific steps are as follows:

[0022] (1) separate each syllable or semi-syllable of all isolated words in the vocabulary as a recognition unit;

[0023] (2) Record the average probability of continuous n frames of speech feature parameters to stay at the end of the first syllable or semi-syllable of each isolated word, reject isolated words whose probability value is less than the specified threshold, and select those probability values ​​for comparison Large isolated words are used as the next matching object;

[0024]...

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 phonetic segmentation-based isolate word recognition method. In the method, a continuous HMM model obtained by voice training takes syllable or semi-syllable as a unit set, trained unit set models are spliced into whole word models according to the syllables or semi-syllables of isolate words in a word list, and a Viterbi algorithm is adopted in recognition. The method has the advantages of improving recognition performance because each segment of HMM model in recognition results (accurately recognized isolate words) can be better matched with each segment of characteristic parameters of voices to be recognized, and shortening recognition response time because the recognition of the isolate words with relatively smaller probability values can be directly refused every time recognition operation is performed on the tail state of one syllable or semi-syllable.

Description

technical field [0001] The invention relates to the technical field of automatic speech recognition, in particular to an isolated word recognition method based on speech segmentation. Background technique [0002] The non-specific isolated word speech recognition methods currently used are all based on the hidden Markov model (Hidden Markov Model, HMM). HMM is used for matching processing, and the maximum probability value is calculated as the recognition result. [0003] The traditional non-specific isolated word speech recognition method is to sequentially calculate the output probability that the speech feature parameters to be recognized stay in the last syllable or semi-syllable (that is, the last state) in each isolated word, and the isolated word with the highest probability is used as recognition result. This method of calculating the total probability of the isolated word from the speech to be recognized at one time will inevitably mistakenly identify some isolate...

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): G10L15/14
Inventor 廖广锐刘萍汤磊
Owner NO 709 RES INST OF CHINA SHIPBUILDING IND CORP
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