Speech recognizer, speech recognition method, and speech recognition program

A sound recognition and sound technology, applied in speech recognition, speech analysis, instruments, etc., can solve problems such as misrecognition and sound recognition devices that cannot recognize sounds

Inactive Publication Date: 2010-03-31
FUJITSU LTD
View PDF4 Cites 4 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

Therefore, there is such a problem that even in the case where the person utters the sound of the recognized word, the voice recognition device sometimes cannot recognize the utterance of the person.
However, in the above-mentioned conventional voice recognition device, when a person utters a voice to recognize a word "tomotomi" other than the word "toyotomi", the speaking voice "tomotomi" and the word model "tootomi" are separated at each time. Sometimes the similarity of is above the threshold, and in this case, the voice "tomotomi" is misrecognized as "toyotomi"

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
  • Speech recognizer, speech recognition method, and speech recognition program
  • Speech recognizer, speech recognition method, and speech recognition program
  • Speech recognizer, speech recognition method, and speech recognition program

Examples

Experimental program
Comparison scheme
Effect test

Embodiment approach 1

[0049] figure 1 It is a block diagram showing a schematic configuration of the voice recognition device 1 according to the present embodiment. figure 1 The illustrated voice recognition device 1 is used, for example, as a voice recognition engine that transmits the user's spoken voice from a host program such as a voice dialogue application, and returns the recognition result to the host program. Furthermore, the voice recognition device 1 is constituted by a general-purpose computer such as a personal computer and a server, for example. In addition, the voice recognition device 1 may be constituted by a computer incorporated in an electronic device such as an in-vehicle information terminal, a mobile phone, or a home electric appliance, for example.

[0050] That is, the voice recognition device 1 according to this embodiment has: a voice analysis unit 11, a recognized word storage unit 12, a conversion rule storage unit 13, a phoneme string conversion unit 14, a phoneme mod...

Embodiment approach 2

[0081] Figure 9 It is a block diagram showing a schematic configuration of the voice recognition device 2 according to this embodiment. That is, the voice recognition device 2 according to this embodiment has conversion rule storage units 21 to 23 instead of figure 1 The conversion rule storage unit 13 is shown. In addition, in Figure 9 In the figure, three conversion rule storage units 21 to 23 are shown for simplicity of description, but the number of conversion rule storage units constituting the voice recognition device 2 is arbitrary. In addition, the voice recognition device 2 according to the present embodiment has a phoneme string conversion unit 24 instead of figure 1 The phoneme string conversion unit 14 is shown. In addition, in Figure 9 , for those with figure 1 Structures with the same functions are assigned the same reference numerals, and detailed description thereof will be omitted.

[0082] Conversion rule storage parts 21-23 and figure 1 Similarly,...

Embodiment approach 3

[0087] Figure 11 It is a block diagram showing a schematic configuration of the voice recognition device 3 according to the present embodiment. That is, the voice recognition device 3 according to this embodiment not only has figure 1 The voice recognition device 1 shown further includes a conversion rule counting unit 31 , a usage frequency calculating unit 32 , and a first threshold condition updating unit 34 . Furthermore, the voice recognition device 3 according to this embodiment has a conversion rule storage unit 33 instead of figure 1 The conversion rule storage unit 13 is shown. In addition, the above-mentioned conversion rule counting unit 31 , use frequency calculating unit 32 and first threshold value condition updating unit 34 can also be realized by the CPU of a computer operating in accordance with a program realizing the functions. In addition, in Figure 11 , for those with figure 1 Structures with the same functions are assigned the same reference numera...

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 provides a speech recognizer, a speech recognition method and a speech recognition program. The speech recognizer (1) comprises a speech collating section (17) for calculating similarities at each time between the amount of a feature converted by a speech analysis section (11) and word models generated by a word model generating section (16). The speech collating section (17) extracts the word model in which the minimum similarity in the similarities at each time or the entire similarity obtained from the similarities at each time out of the word models generated by the word model generating section (16) satisfies a second threshold condition, and in which similarities at each time in the section corresponding to the phoneme or the phoneme string associated with the first threshold condition out of the speech sections of utterance speech satisfy the first threshold condition and outputs the recognized word corresponding to the extracted word model as the result of the recognition.

Description

technical field [0001] The present invention relates to a voice recognition device, a voice recognition method, and a voice recognition program: convert the pronunciation of a recognized word into a phoneme string according to a conversion rule, and generate a word model as a standard pattern string based on the converted phoneme string, thereby recognizing people's speaking voice. Background technique [0002] Generally, a voice recognition device has a function of converting the pronunciation of a recognized word stored in a recognized word storage unit into a phoneme string, and generating a word model as a standard pattern string from the converted phoneme string to recognize human speech. Specifically, the voice recognition device converts the pronunciation of the recognized word into a phoneme string according to the conversion rule between the pronunciation and the phoneme or the conversion rule between the pronunciation and the phoneme string. The voice recognition ...

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): G10L15/06
CPCG10L15/10G10L2015/025
Inventor 原田将治
Owner FUJITSU 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