Unlock instant, AI-driven research and patent intelligence for your innovation.

Multi-language Family Voice Identifying Device And Method Thereof

A speech recognition, multilingual technology, applied in the field of speech recognition, can solve problems such as the ineffectiveness of the deviation correction method, and achieve the effects of eliminating the multilingual deviation phenomenon, reducing the cost, and avoiding the computational burden.

Active Publication Date: 2016-09-07
CHUNGHWA TELECOM CO LTD
View PDF3 Cites 1 Cited by
  • Summary
  • Abstract
  • Description
  • Claims
  • Application Information

AI Technical Summary

Problems solved by technology

[0006] In view of the above-mentioned problems in the prior art, the purpose of the present invention is to provide a multilingual speech recognition device and its method to solve the problem of poor effect of the deviation correction method in the prior art

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
  • Multi-language Family Voice Identifying Device And Method Thereof
  • Multi-language Family Voice Identifying Device And Method Thereof
  • Multi-language Family Voice Identifying Device And Method Thereof

Examples

Experimental program
Comparison scheme
Effect test

Embodiment Construction

[0065] A multilingual speech recognition device and its method provided by the present invention will be described in more detail below with reference to the drawings and embodiments.

[0066] see figure 2 , is a block diagram of a multilingual speech recognition device provided by an embodiment of the present invention. As shown in the figure, the multilingual speech recognition device 2 proposed by the embodiment of the present invention includes a receiving module 20 and a plurality of speech models 21 . in:

[0067] The receiving module 20 is configured to receive the sound frame VF.

[0068] Speech model 21, the speech model 21 is a speech model trained based on corpus of different language families, and includes a plurality of speech states, and the speech model 21 is used to generate a plurality of speech sounds corresponding to the sound frame VF received by the receiving module 20 A plurality of speech state scores 211 of the state, a plurality of correction eleme...

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 embodiment of the present invention provides a multi-language family voice identifying device and a method thereof. The device comprises a receiving module and a plurality of voice models of different language families, wherein the receiving module is used to receive sound frames, the voice models are trained based on language materials of different language families, comprise a plurality of voice states, and are used to generate a plurality of voice state scores corresponding to the plurality of voice states according to the sound frames received by the receiving module. A plurality of correction factors are selected from the voice state scores of the voice models, and the correction values are generated according to the plurality of correction factors. The multi-language family voice identifying device can eliminate a multi-language family deviation phenomenon.

Description

technical field [0001] The invention relates to the technical field of speech recognition, in particular to a multilingual speech recognition device and method. Background technique [0002] Traditionally, the speech model is generally based on the Hidden Markov Model (HMM), in which different speeches in the model will have different numbers of speech states (State), and the model will generate according to the change of the sound frame (Frame). The likelihood (Likelihood) value of each speech state is used as the speech state score. see figure 1 , is a schematic diagram of a multilingual speech recognition device in the prior art. As shown in the figure, the multilingual speech recognition device 1 of the prior art receives the sound frame VF through the receiving model 10, which contains speech models of three languages, wherein the first speech model 11A has 440 first speech state scores 111A, The second speech model 11B has 650 second speech state scores 111B, and th...

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
CPCG10L15/144G10L15/146
Inventor 林心鹏陈建宏陈奕丞林薰苑
Owner CHUNGHWA TELECOM CO LTD