Field-based method and system for feeding back text error correction after speech recognition

A speech recognition and text error correction technology, applied in speech recognition, speech analysis, instruments, etc., can solve problems such as difficulty in correcting errors of homophones

Active Publication Date: 2016-08-17
CHONGQING UNIV
View PDF8 Cites 33 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

However, this method requires a large amount of corpus for training. Error correction after speech recognition in a specific field is difficult due to the limitation of the corpus, and it is difficult to correct some infrequent words, especially homonyms.

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
  • Field-based method and system for feeding back text error correction after speech recognition
  • Field-based method and system for feeding back text error correction after speech recognition
  • Field-based method and system for feeding back text error correction after speech recognition

Examples

Experimental program
Comparison scheme
Effect test

Embodiment 1

[0052] As shown in the figure, a field-based text error correction method after speech recognition with feedback provided by this embodiment includes the following steps:

[0053] S1. Perform part-of-speech tagging on the text sentence after speech recognition according to the controlled dictionary, and identify pause errors according to the structure of the Chinese sentence;

[0054] S2. Convert the text sentence into a phoneme string, and match it with the phoneme string in the corrected text library. If the match is successful, go to step S4; if the match is unsuccessful, go to step S3;

[0055] S3. Perform matching according to the ontology and the controlled dictionary, if the matching is unsuccessful, then end the recognition process; if the matching is successful, then enter the next step;

[0056] S4. Outputting one or more matching results;

[0057] S5. Add the successfully recognized text sentence selected by the user and the corresponding phoneme string of the origin...

Embodiment 2

[0086] The object of the present invention is to provide a kind of field-based text error correction method with feedback speech recognition, comprising the following steps:

[0087] S1. Perform part-of-speech tagging on the text after speech recognition, and judge whether there is a pause error in the sentence. If there is a clause caused by the pause, merge the two sentences.

[0088] S2. Convert the text sentence into a phoneme string, set a threshold, and check whether the sentence has been recognized in the corpus. If the recognition is successful, go to step S4, otherwise go to step S3.

[0089] S3. Correct the words in the text sentence according to the controlled dictionary and ontology.

[0090] S4. Output the error correction result to the front page for the user to choose, and add the user's correct recognition result and the original phoneme string to the corpus.

[0091] After speech recognition in the step S1, the sentence pause error judgment is made up of the ...

Embodiment 3

[0103]This field is set as the stock field in the present embodiment, and the first sentence of the input voice is "find the rise and fall of the electronics industry industry", and the text recognized by the speech engine is "find the electronics industry, the rise and fall of the industry", through the controlled dictionary The result of part-of-speech tagging is that "search" is a verb, "electronics industry" is a noun, "industry" is a noun, "Zhang" does not exist in the controlled dictionary, it is marked as a noun, and "fall" is marked as a noun. Through the part-of-speech matching of the sentence pattern template, it can be known that "find the electronics industry" matches the sentence pattern, but "Zhang Shuai" does not match the sentence pattern, so the two sentences are combined. Then convert the sentence into a phoneme string. Because the corpus is empty, it cannot be matched through the corpus. It is necessary to match words that do not exist in the controlled dicti...

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 field-based method for feeding back text error correction after speech recognition, and belongs to the speech recognition field. Text sentences after speech recognition are analyzed based on errors of speed pauses of Chinese sentence structures. The method is characterized by detecting whether structures before and after a sentence separator meet the sentence pattern rules of Chinese language, finding pause errors, calculating and dividing sentences based on phoneme string similarity and converting into pinyin, converting pinyin into phoneme strings according to a phoneme table, finding sentences corresponding to the phoneme strings similar to the strings in a corpus, establishing a body based on a body controlled word query module through the controlled word table of the field, correcting the errors related to the field in the text after speech recognition through the body, outputting the matching result by a feedback module, and adding the correct identification result selected by a user and the original phoneme strings in the corpus. According to the method and system, the originally correct result of speech recognition may not be affected, and the speech recognition accuracy can be better determined through a body and feedback mechanism.

Description

technical field [0001] The invention relates to the technical field of speech recognition, in particular to a text error correction method after speech recognition with feedback based on the field. Background technique [0002] With the rapid development of computing science and technology, the opportunities for human beings to communicate with computers and machines are increasing and becoming more and more important. How to make it "understand" human language and make appropriate responses is on the research agenda. The goal of the research is to make machines "understand" human language. [0003] The speech recognition process can also be divided into a pre-processing process and a post-processing process. The pre-processing process is mainly to analyze the input speech signal extraction parameters, and its focus is on speech signal processing. Then the processing is mainly to complete the conversion of syllables to Chinese characters, that is, to convert voice informat...

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): G10L15/183G10L15/26G10L15/06
CPCG10L15/183G10L15/26G10L2015/0635
Inventor 钟将崔磊时待吾何隆
Owner CHONGQING UNIV
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