CN106710592A - Speech recognition error correction method and speech recognition error correction device used for intelligent hardware equipment - Google Patents
Speech recognition error correction method and speech recognition error correction device used for intelligent hardware equipment Download PDFInfo
- Publication number
- CN106710592A CN106710592A CN201611246867.0A CN201611246867A CN106710592A CN 106710592 A CN106710592 A CN 106710592A CN 201611246867 A CN201611246867 A CN 201611246867A CN 106710592 A CN106710592 A CN 106710592A
- Authority
- CN
- China
- Prior art keywords
- keyword
- word
- intelligent hardware
- candidate
- extracted
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
- 238000000034 method Methods 0.000 title claims abstract description 51
- 238000012937 correction Methods 0.000 title claims abstract description 47
- 238000005516 engineering process Methods 0.000 claims abstract description 20
- 150000001875 compounds Chemical class 0.000 claims description 12
- 230000008569 process Effects 0.000 claims description 9
- 230000002452 interceptive effect Effects 0.000 claims description 5
- 238000005265 energy consumption Methods 0.000 abstract description 5
- 230000006870 function Effects 0.000 description 10
- 241000282326 Felis catus Species 0.000 description 6
- 235000015278 beef Nutrition 0.000 description 5
- XLYOFNOQVPJJNP-UHFFFAOYSA-N water Substances O XLYOFNOQVPJJNP-UHFFFAOYSA-N 0.000 description 4
- 230000008901 benefit Effects 0.000 description 3
- 235000013372 meat Nutrition 0.000 description 3
- 238000004590 computer program Methods 0.000 description 2
- 239000000284 extract Substances 0.000 description 2
- 230000003993 interaction Effects 0.000 description 2
- 241000238558 Eucarida Species 0.000 description 1
- 230000015572 biosynthetic process Effects 0.000 description 1
- 230000008859 change Effects 0.000 description 1
- 238000001035 drying Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003786 synthesis reaction Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F40/00—Handling natural language data
- G06F40/30—Semantic analysis
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/10—Speech classification or search using distance or distortion measures between unknown speech and reference templates
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
- G10L15/1815—Semantic context, e.g. disambiguation of the recognition hypotheses based on word meaning
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L15/18—Speech classification or search using natural language modelling
- G10L15/1822—Parsing for meaning understanding
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/26—Speech to text systems
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L2015/086—Recognition of spelled words
-
- G—PHYSICS
- G10—MUSICAL INSTRUMENTS; ACOUSTICS
- G10L—SPEECH ANALYSIS TECHNIQUES OR SPEECH SYNTHESIS; SPEECH RECOGNITION; SPEECH OR VOICE PROCESSING TECHNIQUES; SPEECH OR AUDIO CODING OR DECODING
- G10L15/00—Speech recognition
- G10L15/08—Speech classification or search
- G10L2015/088—Word spotting
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Acoustics & Sound (AREA)
- Human Computer Interaction (AREA)
- Multimedia (AREA)
- Artificial Intelligence (AREA)
- Theoretical Computer Science (AREA)
- General Health & Medical Sciences (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Machine Translation (AREA)
Abstract
The invention discloses a speech recognition error correction method and a speech recognition error correction device used for intelligent hardware equipment. The speech recognition error correction method comprises steps that a speech signal received by the intelligent hardware equipment is converted into text information by adopting a speech recognition technology; a key word is extracted from the text information; the extracted key word is matched with a vocabulary the business-related of the intelligent hardware, and one or a plurality of candidate word/s matched with the key word is/are selected from the vocabulary; the error correction processing of the key word in the text information is carried out according to the selected one or a plurality of candidate word/s. According to the technical scheme, the function characteristic of the intelligent hardware is fully considered, and by using the preset business-related vocabulary, the intelligent error correction of the key word parsed from the speed recognition result is carried out, and the speech recognition accuracy is obviously improved, and occupied resources are less, and therefore the requirement of the intelligent hardware equipment on low energy consumption is satisfied.
Description
Technical field
The present invention relates to technical field of voice recognition, and in particular to the speech recognition error correction side in a kind of intelligent hardware devices
Method and device.
Background technology
The development of speech recognition technology causes user and intelligent hardware devices (such as intelligent watch, mobile phone, drive recorder)
Interaction become more convenient.Shown below is in the prior art, user is entered using speech recognition technology with intelligent hardware devices
Several examples of row interaction:
1) by the instruction morphing of user it is word by speech recognition technology;
2) user view is understood by semantic analysis technology;
3) literal resource that will be found by speech synthesis technique changes into voice, feeds back to user.
Wherein, the process of semantic analysis relies on the accuracy rate of speech recognition very much, and speech recognition accuracy is difficult to reach
100%, such as speech recognition can be made various mistakes are occurred when user is children, the characteristics of it pronounces indistinctly clear.
The content of the invention
In view of the above problems, it is proposed that the present invention so as to provide one kind overcome above mentioned problem or at least in part solve on
State speech recognition error correction method and the device in the intelligent hardware devices of problem.
According to one aspect of the present invention, there is provided the speech recognition error correction method in a kind of intelligent hardware devices, including:
The voice signal that intelligent hardware devices are received is converted into text information by speech recognition technology;
Keyword is extracted from the text information;
The keyword that will the be extracted vocabulary related to Intelligent hardware business is matched, and is selected from vocabulary and the pass
One or more candidate's words of keyword matching;
Correction process is carried out to the keyword in the text information according to one or more selected candidate's words.
Alternatively, the method is further included:Pre-set one or more business languages with the intelligent hardware devices
One or more of sound cross-correlation fix clause;In the position of each fixation clause acceptance of the bid note keyword;
The keyword that extracted from the text information includes:By the text information and one or more of fixations
Clause is matched;The position of the keyword according to the fixed clause acceptance of the bid note for matching, from the corresponding positions of the text information
Put extraction keyword.
Alternatively, the method is further included:It is the keyword tag type information in each fixation clause;It is determined that intelligence
The type information of the related vocabulary of hardware business;
The keyword that will the be extracted vocabulary related to Intelligent hardware business carries out matching to be included:According to what is matched
The type information of the keyword in fixed clause determines the type information of extracted keyword, according to the keyword for being extracted
The keyword that type information will be extracted is matched with the vocabulary of type matching.
Alternatively, one or more the candidate's words selected by the basis are carried out to the keyword in the text information
Correction process includes:
To selected with the Keywords matching each candidate's words and phrases, according to extracted keyword and candidate's word
Matching degree be candidate's word marking;
If the fraction of the best result candidate word of the keyword is greater than or equal to the first confidence value, the best result is used
Candidate word corrects the keyword;
If the fraction of the best result candidate word of the keyword is higher than the second confidence value but less than the first confidence value,
Further voice dialogue then is carried out with user, to be confirmed whether to need to correct the keyword with the best result candidate word;
If the fraction of the best result candidate word of the keyword is less than or equal to the second confidence value, do not corrected.
Alternatively, described marking for candidate's word with the matching degree of candidate's word according to extracted keyword includes:
Fraction is divided into high, medium and low three gear bands from high to low;
If keyword is identical with the phonetic of the candidate word, simply tone is different, then given a mark in the range of high tap position;
If keyword is identical with the initial consonant in the phonetic of the candidate word or simple or compound vowel of a Chinese syllable part, beaten in the range of middle gear
Point;
If the initial consonant, simple or compound vowel of a Chinese syllable in keyword and the phonetic of the candidate word are different from, given a mark in the range of low-grade location.
Alternatively, the method is further included:
If being extracted multiple keywords from the text information, by the fraction of the best result candidate word of each keyword
It is multiplied, obtains the fraction of the plurality of keyword;
If the fraction of multiple keywords is greater than or equal to the 3rd confidence value, with the best result candidate word of each keyword
Correct each keyword;
If the fraction of multiple keywords enters to advance higher than the 4th confidence value but less than the 3rd confidence value with user
The voice dialogue of one step, each keyword is corrected with the best result candidate word for being confirmed whether to need each keyword;
If the fraction of multiple keywords is less than or equal to the 4th confidence value, do not corrected.
Alternatively, the method is further included:
According to the corresponding service service for correcting the result output intelligent hardware devices.
According to another aspect of the present invention, there is provided the speech recognition error correction device in a kind of intelligent hardware devices, including:
Voice recognition unit, the voice signal for being suitable to receive intelligent hardware devices is changed written by speech recognition technology
Word information;
Keyword extracting unit, is suitable to extract keyword from the text information;
Matching unit, is suitable to the keyword that will the be extracted vocabulary related to Intelligent hardware business and is matched, from vocabulary
In select one or more candidate's words with the Keywords matching;
Error correction unit, is suitable to one or more the candidate's words according to selected by and the keyword in the text information is entered
Row correction process.
Alternatively, the device is further included:Dispensing unit, is suitable to pre-set one or more with the Intelligent hardware
One or more of the business interactive voice association of equipment fix clause;In the position of each fixation clause acceptance of the bid note keyword;
The keyword extracting unit, is suitable to carry out the text information and one or more of fixed clause
Match somebody with somebody;The position of the keyword according to the fixed clause acceptance of the bid note for matching, extracts crucial from the relevant position of the text information
Word.
Alternatively, the dispensing unit, is further adapted for each keyword tag type information fixed in clause;Really
Determine the type information of the related vocabulary of Intelligent hardware business;
The matching unit, the type information for being suitable to the keyword in the fixed clause for matching determines what is extracted
The type information of keyword, keyword and the vocabulary of type matching that the type information according to the keyword for being extracted will be extracted
Matched.
Alternatively, the error correction unit, is suitable to selected with the Keywords matching each candidate's words and phrases, according to
Extracted keyword is candidate's word marking with the matching degree of candidate's word;If the best result candidate word of the keyword
Fraction be greater than or equal to the first confidence value, then with the best result candidate word correction keyword;If the keyword
Best result candidate word fraction higher than the second confidence value but less than the first confidence value, then carry out further language with user
Sound is talked with, to be confirmed whether to need to correct the keyword with the best result candidate word;If the best result of the keyword is waited
Select the fraction of word less than or equal to the second confidence value, do not corrected.
Alternatively, the error correction unit, is suitable to fraction is divided into high, medium and low three gear bands from high to low;If
Keyword is identical with the phonetic of the candidate word, and simply tone is different, then given a mark in the range of high tap position;If keyword and the time
Select initial consonant in the phonetic of word or simple or compound vowel of a Chinese syllable part identical, then given a mark in the range of middle gear;If keyword and the candidate word
Initial consonant, simple or compound vowel of a Chinese syllable in phonetic are different from, then given a mark in the range of low-grade location.
Alternatively, the error correction unit, is further adapted for when multiple keywords are extracted from the text information, will
The fractional multiplication of the best result candidate word of each keyword, obtains the fraction of the plurality of keyword;If the fraction of multiple keywords
Greater than or equal to the 3rd confidence value, then each keyword is corrected with the best result candidate word of each keyword;If multiple is closed
The fraction of keyword then carries out further voice dialogue higher than the 4th confidence value but less than the 3rd confidence value with user, with
It is confirmed whether to need the best result candidate word of each keyword to correct each keyword;If the fraction of multiple keywords be less than or
Equal to the 4th confidence value, do not corrected.
Alternatively, the device is further included:
Business service unit, is suitable to according to the corresponding service service for correcting the result output intelligent hardware devices.
From the foregoing, intelligent hardware devices are received by technical scheme first with speech recognition technology
Voice signal carries out speech recognition, is converted into text information, further parse text information obtain it is therein several
Keyword, these keywords are matched by the vocabulary related to Intelligent hardware business, determine one or more candidate words
Language, finally carries out error correction using the candidate's word for obtaining to keyword.The technical scheme has taken into full account the function of Intelligent hardware
Characteristic, intelligent correction is carried out using default business correlation vocabulary to the keyword parsed in voice identification result, is significantly carried
The degree of accuracy of speech recognition has been risen, and the resource for taking is less, meets the demand of intelligent hardware devices low energy consumption.
Described above is only the general introduction of technical solution of the present invention, in order to better understand technological means of the invention,
And can be practiced according to the content of specification, and in order to allow the above and other objects of the present invention, feature and advantage can
Become apparent, below especially exemplified by specific embodiment of the invention.
Brief description of the drawings
By reading the detailed description of hereafter preferred embodiment, various other advantages and benefit is common for this area
Technical staff will be clear understanding.Accompanying drawing is only used for showing the purpose of preferred embodiment, and is not considered as to the present invention
Limitation.And in whole accompanying drawing, identical part is denoted by the same reference numerals.In the accompanying drawings:
Fig. 1 shows the speech recognition error correction method in a kind of intelligent hardware devices according to an embodiment of the invention
Schematic flow sheet;
Fig. 2 shows the speech recognition error correction device in a kind of intelligent hardware devices according to an embodiment of the invention
Structural representation.
Specific embodiment
The exemplary embodiment of the disclosure is more fully described below with reference to accompanying drawings.Although showing the disclosure in accompanying drawing
Exemplary embodiment, it being understood, however, that may be realized in various forms the disclosure without should be by embodiments set forth here
Limited.Conversely, there is provided these embodiments are able to be best understood from the disclosure, and can be by the scope of the present disclosure
Complete conveys to those skilled in the art.
Fig. 1 shows the speech recognition error correction method in a kind of intelligent hardware devices according to an embodiment of the invention
Schematic flow sheet, as shown in figure 1, the method includes:
Step S110, text information is converted into by the voice signal that intelligent hardware devices are received by speech recognition technology.
Wherein, intelligent hardware devices can be smart mobile phone, intelligent watch, intelligent robot etc..Speech recognition technology is
It is a kind of just in gradual perfection and the technology that has come into operation, such as the Siri functions in i Phone, it is possible to realize to user
Sound speech recognition is carried out to perform the functions such as opening camera, digital map navigation in i Phone.
Step S120, extracts keyword from text information.
For example, user is desirable for the function of telling a story in children's wrist-watch, have issued " I wants to listen story --- small red cap "
Voice signal, then " story " is keyword corresponding with the function, and " small red cap " is then and the story classification under the function
Corresponding keyword.That is, keyword is related to Intelligent hardware business.Therefore following step is continued executing with:
Step S130, the keyword that will the be extracted vocabulary related to Intelligent hardware business is matched, and is selected from vocabulary
Go out one or more the candidate's words with Keywords matching.
When actually used, the pronunciation of user is possible and nonstandard, so results in sound and does not express one's ideas, for example, child's handle
" small red cap " has been said into " Xiao Hong caps " (sound), then the word message that the voice signal is changed may be by speech recognition technology
" Xiao Hong caps ".It is evident that the word can not be matched completely with " small red cap ", but the similarity of the two is very high, by people
Work judge can draw child it is actual think expression be exactly " small red cap ".But in fact, the text that voice signal is converted into
Word message energy fuzzy matching to keyword be probably very many, for example, in story storehouse may also have " small red cat " story.
So for a keyword, the candidate's word selected from vocabulary is probably one, it is also possible to multiple.
Step S140, is carried out at error correction according to one or more selected candidate's words to the keyword in text information
Reason.
It can be seen that, the method shown in Fig. 1 is entered first with speech recognition technology to the voice signal that intelligent hardware devices are received
Row speech recognition, is converted into text information, further parses text information and obtains several keywords therein, by this
A little keywords are matched by the vocabulary related to Intelligent hardware business, determine one or more candidate's words, are finally utilized
The candidate's word for obtaining carries out error correction to keyword.The technical scheme has taken into full account the functional characteristic of Intelligent hardware, using pre-
If business correlation vocabulary intelligent correction is carried out to the keyword parsed in voice identification result, be obviously improved speech recognition
The degree of accuracy, and take resource it is less, meet the demand of intelligent hardware devices low energy consumption.
In one embodiment of the invention, the method shown in Fig. 1 is further included:Pre-set one or more and intelligence
One or more of the business interactive voice association of energy hardware device fix clause;In each fixation clause acceptance of the bid note keyword
Position;Keyword is extracted from text information to be included:Text information is matched with one or more fixation clause;According to phase
The position of the keyword of the fixed clause acceptance of the bid note of matching, keyword is extracted from the relevant position of text information.
In previous embodiment, " I will listen story --- Xiao Hong caps " it is exactly a kind of fixed clause, and " I can be summarized as
Want to listen story XXX ", " XXX " therein corresponds to a keyword.Again for example, the semanteme of the sentence can also be expressed as:" I
Want to listen the story of Xiao Hong caps ", then for fixed clause be " I wants to listen the story of XXX ".Support that song is broadcast in Intelligent hardware
When putting, user can use clause as " I wants to listen the XXX of XXX ", for example, " I wants to listen the lustily water of Liu De China ", etc..
It is evident that, by the text information of " I wants to listen the story of Xiao Hong caps ", this is fixed with " I wants to listen the story of XXX "
Clause is matched, it is possible to it is determined that " Xiao Hong caps " is keyword.Further, can also determine " Xiao Hong caps " it is corresponding be therefore
Thing title.Therefore in one embodiment of the invention, the above method is further included:It is the keyword in each fixation clause
Type information;Determine the type information of the related vocabulary of Intelligent hardware business;The keyword and Intelligent hardware that will be extracted
The related vocabulary of business carries out matching to be included:The type information of the keyword in the fixed clause for matching determines to be extracted
Keyword type information, the word of keyword that the type information according to the keyword for being extracted will be extracted and type matching
Table is matched.
For " telling a story " this business function, the vocabulary related to business can be story name;For " song is broadcast
Put " for this business function, the vocabulary related to business can be song title, song style, singer's name etc..Implement upper one
In example, it may be determined that position of each keyword in fixed clause, then believed by the corresponding type of the keyword of each position
Breath, it is possible to judge carry out the matching of candidate's word using the related vocabulary of which business.For in a upper embodiment
" Xiao Hong caps ", it is possible to matched using story noun list;For " Liu Dehua ", it is possible to use singer noun list enters
Row matching;For " lustily water ", it is possible to use song noun list is matched.
After keyword and corresponding candidate's word is obtained, the real language of keyword can be determined according to candidate's word
Justice.For example, what child wanted to listen is on earth the story of " small red cat " or the story of " small red cap ".In one of the invention implementation
In example, in the method shown in Fig. 1, the keyword in text information is entangled according to one or more selected candidate's words
Fault reason includes:To selected with Keywords matching each candidate's words and phrases, according to extracted keyword and candidate's word
Matching degree be candidate's word marking;If the fraction of the best result candidate word of keyword is greater than or equal to the first confidence level
Value, then correct keyword with the best result candidate word;If the fraction of the best result candidate word of keyword is higher than the second confidence level
It is worth but less than the first confidence value, then carries out further voice dialogue with user, is confirmed whether to need to be waited with the best result
Word is selected to correct keyword;If the fraction of the best result candidate word of keyword is less than or equal to the second confidence value, do not entangled
Just.
For example, " Xiao Hong caps " corresponding candidate word has " small red cat " and " small red cap " two words." Xiao Hong caps " with it is " small red
Cap " has two words identical, and only one of which word voice is different;" Xiao Hong caps " has a word identical with " small red cat ", two other word
Voice is different.So fractional result may be:The fraction of " small red cap " is 0.6, and the fraction of " small red cat " is 0.5.So with
It is all higher than 0.45 due to two fractions of candidate's word, then the time of selection wherein best result as a example by 0.45 that one confidence value is
Word, that is, " small red cap " is selected to correct " Xiao Hong caps " this keyword.
Shown below is several strategy examples given a mark to candidate's word language:In one embodiment of the invention,
In the above method, being given a mark for candidate's word with the matching degree of candidate's word according to extracted keyword includes:By fraction from
It is high to Low to be divided into high, medium and low three gear bands;If keyword is identical with the phonetic of the candidate word, simply tone is different,
Then given a mark in the range of high tap position;If keyword is identical with the initial consonant in the phonetic of the candidate word or simple or compound vowel of a Chinese syllable part, in
Marking in gear bands;If the initial consonant, simple or compound vowel of a Chinese syllable in keyword and the phonetic of the candidate word are different from, in low-grade location scope
Interior marking.
For example, the fraction range of high tap position is [0.45,1], and the fraction range of middle gear is (0.4,0.45), low-grade
The fraction range of position is for [0,0.4).Because " small red cat " and " small red cap " belongs to " Xiao Hong caps ", " phonetic is identical, simply tone
It is different ", then using the scoring criterion of high tap position.Try another example:Certain customers " n ", the initial consonant pronunciation of " l " are regardless of causing
" beef " has been said into " stream meat ", then due to " beef " this candidate's word and " stream meat " this keyword part initial consonant not
Together, simple or compound vowel of a Chinese syllable part is identical, then using the scoring criterion of middle gear.Another example is seen again:User due to memory deviation, by " Liu De
China " has been remembered into " Zhang Dehua ", then the initial consonant of " Liu " and " opening ", simple or compound vowel of a Chinese syllable are different from, but other two word is identical, in this feelings
Can be using the scoring criterion of low-grade location under condition.
So after the fraction for getting candidate's word, it is possible to according to the method in a upper embodiment, judge best result
Candidate's word confidence level.For example, the fraction of " beef " is 0.44, less than the first confidence value 0.45 but higher than the second confidence
Angle value 0.4, then user can be inquired:" what you thought is ' beef '" being confirmed whether to need " to be flowed with " beef " corrections
Meat ".If the fraction of the best result candidate word of keyword is less than or equal to the second confidence value, do not corrected, because at this moment
Even if being corrected, it is also possible to deviate from the original meaning of user.
In one embodiment of the invention, the above method is further included:If being extracted multiple from text information
Keyword, then by the fractional multiplication of the best result candidate word of each keyword, obtain the fraction of the plurality of keyword;If multiple is closed
The fraction of keyword is greater than or equal to the 3rd confidence value, then correct each keyword with the best result candidate word of each keyword;If
The fraction of multiple keywords then carries out further voice pair higher than the 4th confidence value but less than the 3rd confidence value with user
Words, each keyword is corrected with the best result candidate word for being confirmed whether to need each keyword;If the fraction of multiple keywords is less than
Or equal to the 4th confidence value, do not corrected.
Give in the present embodiment when keyword is for multiple, how to calculate the example of candidate word fraction.For example, "
The fraction of moral China " is 0.3, " lustily water " user pronunciation very standard, and the fraction of corresponding " lustily water " candidate word is 1, then this
Two fractions of keyword are 0.3 × 1=0.3.
In one embodiment of the invention, the above method is further included:According to correcting, result output intelligence is hard
The corresponding service service of part equipment.
For example, in the case where technical solution of the present invention is not performed, child says to intelligent robot of accompanying and attending to:" you can say
The story of Xiao Hong caps", because " Xiao Hong caps " and " small red cap " is mismatched, the reply that child obtains is for " I will not also for this
." at this moment mother robot of being accompanied and attended to intelligence again be made that correct instruction " telling the story of small red cap ", machine of intelligently accompanying and attending to
People correctly obtains the resource of " small red cap " this story, carries out voice and tells about.
And in the present embodiment, due to being said in child, " you can tell the story of Xiao Hong caps" when, by speech recognition
Correct, correctly can be corrected as " small red cap " " Xiao Hong caps ", it is possible to correctly obtain the money of " small red cap " this story
Source, carries out voice and tells about.
Fig. 2 shows the speech recognition error correction device in a kind of intelligent hardware devices according to an embodiment of the invention
Structural representation, as shown in Fig. 2 the speech recognition error correction device 200 in intelligent hardware devices includes:
Voice recognition unit 210, the voice signal for being suitable to receive intelligent hardware devices is changed by speech recognition technology
Into text information.
Keyword extracting unit 220, is suitable to extract keyword from text information.
Matching unit 230, is suitable to the keyword that will the be extracted vocabulary related to Intelligent hardware business and is matched, from word
One or more the candidate's words with Keywords matching are selected in table.
Error correction unit 240, is suitable to one or more the candidate's words according to selected by and the keyword in text information is entered
Row correction process.
It can be seen that, the device shown in Fig. 2 is hard to intelligence first with speech recognition technology by the mutual cooperation of each unit
The voice signal that part equipment is received carries out speech recognition, is converted into text information, further parses text information and obtains
Several keywords therein, these keywords are matched by the vocabulary related to Intelligent hardware business, determine one
Or multiple candidate's words, error correction finally is carried out to keyword using the candidate's word for obtaining.The technical scheme has taken into full account intelligence
The functional characteristic of energy hardware, intelligence is carried out using default business correlation vocabulary to the keyword parsed in voice identification result
Error correction, has been obviously improved the degree of accuracy of speech recognition, and the resource for taking is less, meets the need of intelligent hardware devices low energy consumption
Ask.
In one embodiment of the invention, said apparatus also include:Dispensing unit, is suitable to pre-set one or more
One or more associated with the business interactive voice of intelligent hardware devices fix clause;It is crucial in each fixation clause acceptance of the bid note
The position of word;Keyword extracting unit 220, is suitable to be matched text information with one or more fixation clause;According to phase
The position of the keyword of the fixed clause acceptance of the bid note of matching, keyword is extracted from the relevant position of text information.
In one embodiment of the invention, in said apparatus, dispensing unit is further adapted for each and fixes in clause
Keyword tag type information;Determine the type information of the related vocabulary of Intelligent hardware business;Matching unit 230, is suitable to root
Type information according to the keyword in the fixed clause for matching determines the type information of extracted keyword, according to being extracted
The keyword that will be extracted of type information of keyword matched with the vocabulary of type matching.
In one embodiment of the invention, in said apparatus, error correction unit 240 is suitable to selected and keyword
Each candidate's words and phrases of matching, the matching degree according to extracted keyword and candidate's word is candidate's word marking;If
The fraction of the best result candidate word of keyword is greater than or equal to the first confidence value, then correct crucial with the best result candidate word
Word;If the fraction of the best result candidate word of keyword is higher than the second confidence value but less than the first confidence value, with user
Further voice dialogue is carried out, to be confirmed whether to need to correct keyword with the best result candidate word;If keyword is most
The fraction of high score candidate word is less than or equal to the second confidence value, is not corrected.
In one embodiment of the invention, in said apparatus, error correction unit 240 is suitable to from high to low divide fraction
It is high, medium and low three gear bands;
If keyword is identical with the phonetic of the candidate word, simply tone is different, then given a mark in the range of high tap position;If
Keyword is identical with the initial consonant in the phonetic of the candidate word or simple or compound vowel of a Chinese syllable part, then given a mark in the range of middle gear;If keyword
It is different from the initial consonant in the phonetic of the candidate word, simple or compound vowel of a Chinese syllable, then is given a mark in the range of low-grade location.
In one embodiment of the invention, in said apparatus, error correction unit 240 is further adapted for working as from text information
In when being extracted multiple keywords, by the fractional multiplication of the best result candidate word of each keyword, obtain the plurality of keyword point
Number;If the fraction of multiple keywords is greater than or equal to the 3rd confidence value, corrected with the best result candidate word of each keyword
Each keyword;If the fraction of multiple keywords is carried out higher than the 4th confidence value but less than the 3rd confidence value with user
Further voice dialogue, each keyword is corrected with the best result candidate word for being confirmed whether to need each keyword;If multiple is closed
The fraction of keyword is less than or equal to the 4th confidence value, is not corrected.
In one embodiment of the invention, said apparatus are further included:Business service unit, is suitable to according at correction
Reason result exports the corresponding service service of intelligent hardware devices.
It should be noted that the specific embodiment of above-mentioned each device embodiment is specific with foregoing corresponding method embodiment
Implementation method is identical, will not be repeated here.
In sum, technical scheme, first with the language that speech recognition technology is received to intelligent hardware devices
Message number carries out speech recognition, is converted into text information, further parses text information and obtains several passes therein
Keyword, these keywords are matched by the vocabulary related to Intelligent hardware business, determine one or more candidate's words,
Error correction finally is carried out to keyword using the candidate's word for obtaining.The technical scheme has taken into full account that the function of Intelligent hardware is special
Property, intelligent correction is carried out to the keyword parsed in voice identification result using default business correlation vocabulary, it is obviously improved
The degree of accuracy of speech recognition, and the resource for taking is less, meets the demand of intelligent hardware devices low energy consumption.
It should be noted that:
Algorithm and display be not inherently related to any certain computer, virtual bench or miscellaneous equipment provided herein.
Various fexible units can also be used together with based on teaching in this.As described above, construct required by this kind of device
Structure be obvious.Additionally, the present invention is not also directed to any certain programmed language.It is understood that, it is possible to use it is various
Programming language realizes the content of invention described herein, and the description done to language-specific above is to disclose this hair
Bright preferred forms.
In specification mentioned herein, numerous specific details are set forth.It is to be appreciated, however, that implementation of the invention
Example can be put into practice in the case of without these details.In some instances, known method, structure is not been shown in detail
And technology, so as not to obscure the understanding of this description.
Similarly, it will be appreciated that in order to simplify one or more that the disclosure and helping understands in each inventive aspect, exist
Above to the description of exemplary embodiment of the invention in, each feature of the invention is grouped together into single implementation sometimes
In example, figure or descriptions thereof.However, the method for the disclosure should be construed to reflect following intention:I.e. required guarantor
The application claims of shield features more more than the feature being expressly recited in each claim.More precisely, such as following
Claims reflect as, inventive aspect is all features less than single embodiment disclosed above.Therefore,
Thus the claims for following specific embodiment are expressly incorporated in the specific embodiment, and wherein each claim is in itself
All as separate embodiments of the invention.
Those skilled in the art are appreciated that can be carried out adaptively to the module in the equipment in embodiment
Change and they are arranged in one or more equipment different from the embodiment.Can be the module or list in embodiment
Unit or component be combined into a module or unit or component, and can be divided into addition multiple submodule or subelement or
Sub-component.In addition at least some in such feature and/or process or unit exclude each other, can use any
Combine to all features disclosed in this specification (including adjoint claim, summary and accompanying drawing) and so disclosed appoint
Where all processes or unit of method or equipment are combined.Unless expressly stated otherwise, this specification (including adjoint power
Profit is required, summary and accompanying drawing) disclosed in each feature can the alternative features of or similar purpose identical, equivalent by offer carry out generation
Replace.
Although additionally, it will be appreciated by those of skill in the art that some embodiments described herein include other embodiments
In included some features rather than further feature, but the combination of the feature of different embodiments means in of the invention
Within the scope of and form different embodiments.For example, in the following claims, embodiment required for protection is appointed
One of meaning mode can be used in any combination.
All parts embodiment of the invention can be realized with hardware, or be run with one or more processor
Software module realize, or with combinations thereof realize.It will be understood by those of skill in the art that can use in practice
Microprocessor or digital signal processor (DSP) are known realizing the voice in intelligent hardware devices according to embodiments of the present invention
The some or all functions of some or all parts in other error correction device.The present invention is also implemented as performing this
In described method some or all equipment or program of device (for example, computer program and computer program
Product).It is such to realize that program of the invention be stored on a computer-readable medium, or can be with one or many
The form of individual signal.Such signal can be downloaded from internet website and obtained, or be provided on carrier signal, or with
Any other form is provided.
It should be noted that above-described embodiment the present invention will be described rather than limiting the invention, and ability
Field technique personnel can design alternative embodiment without departing from the scope of the appended claims.In the claims,
Any reference symbol being located between bracket should not be configured to limitations on claims.Word "comprising" is not excluded the presence of not
Element listed in the claims or step.Word "a" or "an" before element is not excluded the presence of as multiple
Element.The present invention can come real by means of the hardware for including some different elements and by means of properly programmed computer
It is existing.If in the unit claim for listing equipment for drying, several in these devices can be by same hardware branch
To embody.The use of word first, second, and third does not indicate that any order.These words can be explained and run after fame
Claim.
Claims (10)
1. the speech recognition error correction method in a kind of intelligent hardware devices, wherein, the method includes:
The voice signal that intelligent hardware devices are received is converted into text information by speech recognition technology;
Keyword is extracted from the text information;
The keyword that will the be extracted vocabulary related to Intelligent hardware business is matched, and is selected from vocabulary and the keyword
One or more candidate's words of matching;
Correction process is carried out to the keyword in the text information according to one or more selected candidate's words.
2. the method for claim 1, wherein
The method is further included:Pre-set what one or more were associated with the business interactive voice of the intelligent hardware devices
One or more fix clause;In the position of each fixation clause acceptance of the bid note keyword;
The keyword that extracted from the text information includes:By the text information and one or more of fixed clause
Matched;The position of the keyword according to the fixed clause acceptance of the bid note for matching, carries from the relevant position of the text information
Take keyword.
3. method as claimed in claim 1 or 2, wherein,
The method is further included:It is the keyword tag type information in each fixation clause;Determine Intelligent hardware business phase
The type information of the vocabulary of pass;
The keyword that will the be extracted vocabulary related to Intelligent hardware business carries out matching to be included:According to the fixation for matching
The type information of the keyword in clause determines the type information of extracted keyword, according to the type of the keyword for being extracted
The keyword that information will be extracted is matched with the vocabulary of type matching.
4. the method as any one of claim 1-3, wherein, one or more the candidate's words selected by the basis
Carrying out correction process to the keyword in the text information includes:
To selected with the Keywords matching each candidate's words and phrases, according to extracted keyword and candidate's word
With degree for candidate's word is given a mark;
If the fraction of the best result candidate word of the keyword is greater than or equal to the first confidence value, with best result candidate
Word corrects the keyword;
If the fraction of the best result candidate word of the keyword is higher than the second confidence value but less than the first confidence value, with
User carries out further voice dialogue, to be confirmed whether to need to correct the keyword with the best result candidate word;
If the fraction of the best result candidate word of the keyword is less than or equal to the second confidence value, do not corrected.
5. the method as any one of claim 1-4, wherein, it is described according to extracted keyword and candidate's word
Matching degree is that candidate's word marking includes:
Fraction is divided into high, medium and low three gear bands from high to low;
If keyword is identical with the phonetic of the candidate word, simply tone is different, then given a mark in the range of high tap position;
If keyword is identical with the initial consonant in the phonetic of the candidate word or simple or compound vowel of a Chinese syllable part, given a mark in the range of middle gear;
If the initial consonant, simple or compound vowel of a Chinese syllable in keyword and the phonetic of the candidate word are different from, given a mark in the range of low-grade location.
6. the method as any one of claim 1-5, wherein, the method is further included:
If being extracted multiple keywords from the text information, by the fraction phase of the best result candidate word of each keyword
Multiply, obtain the fraction of the plurality of keyword;
If the fraction of multiple keywords is greater than or equal to the 3rd confidence value, corrected with the best result candidate word of each keyword
Each keyword;
If the fraction of multiple keywords is carried out further higher than the 4th confidence value but less than the 3rd confidence value with user
Voice dialogue, each keyword is corrected with the best result candidate word for being confirmed whether to need each keyword;
If the fraction of multiple keywords is less than or equal to the 4th confidence value, do not corrected.
7. the method as any one of claim 1-6, wherein, the method is further included:
According to the corresponding service service for correcting the result output intelligent hardware devices.
8. the speech recognition error correction device in a kind of intelligent hardware devices, wherein, the device includes:
Voice recognition unit, the voice signal for being suitable to receive intelligent hardware devices is converted into word letter by speech recognition technology
Breath;
Keyword extracting unit, is suitable to extract keyword from the text information;
Matching unit, is suitable to the keyword that will the be extracted vocabulary related to Intelligent hardware business and is matched, and is selected from vocabulary
Go out one or more the candidate's words with the Keywords matching;
Error correction unit, is suitable to one or more the candidate's words according to selected by and the keyword in the text information is entangled
Fault is managed.
9. device as claimed in claim 8, wherein, the device is further included:Dispensing unit, be suitable to pre-set one or
Multiple one or more associated with the business interactive voice of the intelligent hardware devices fix clause;In each fixation clause
Mark the position of keyword;
The keyword extracting unit, is suitable to be matched the text information with one or more of fixed clause;Root
According to the position of the keyword of the fixed clause acceptance of the bid note for matching, keyword is extracted from the relevant position of the text information.
10. device as claimed in claim 8 or 9, wherein,
The dispensing unit, is further adapted for each keyword tag type information fixed in clause;Determine Intelligent hardware
The type information of the related vocabulary of business;
The matching unit, the type information for being suitable to the keyword in the fixed clause for matching determines extracted key
The type information of word, the keyword that the type information according to the keyword for being extracted will be extracted is carried out with the vocabulary of type matching
Matching.
Priority Applications (2)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611246867.0A CN106710592B (en) | 2016-12-29 | 2016-12-29 | Voice recognition error correction method and device in intelligent hardware equipment |
PCT/CN2017/116165 WO2018121275A1 (en) | 2016-12-29 | 2017-12-14 | Method and apparatus for error connection of voice recognition in smart hardware device |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201611246867.0A CN106710592B (en) | 2016-12-29 | 2016-12-29 | Voice recognition error correction method and device in intelligent hardware equipment |
Publications (2)
Publication Number | Publication Date |
---|---|
CN106710592A true CN106710592A (en) | 2017-05-24 |
CN106710592B CN106710592B (en) | 2021-05-18 |
Family
ID=58906036
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201611246867.0A Active CN106710592B (en) | 2016-12-29 | 2016-12-29 | Voice recognition error correction method and device in intelligent hardware equipment |
Country Status (2)
Country | Link |
---|---|
CN (1) | CN106710592B (en) |
WO (1) | WO2018121275A1 (en) |
Cited By (21)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107239547A (en) * | 2017-06-05 | 2017-10-10 | 北京智能管家科技有限公司 | Voice error correction method, terminal and storage medium for ordering song by voice |
CN107977356A (en) * | 2017-11-21 | 2018-05-01 | 新疆科大讯飞信息科技有限责任公司 | Method and device for correcting recognized text |
CN108021554A (en) * | 2017-11-14 | 2018-05-11 | 无锡小天鹅股份有限公司 | Audio recognition method, device and washing machine |
WO2018121275A1 (en) * | 2016-12-29 | 2018-07-05 | 北京奇虎科技有限公司 | Method and apparatus for error connection of voice recognition in smart hardware device |
CN108647346A (en) * | 2018-05-15 | 2018-10-12 | 苏州东巍网络科技有限公司 | A kind of the elderly's voice interactive method and system for wearable electronic |
CN108877792A (en) * | 2018-05-30 | 2018-11-23 | 北京百度网讯科技有限公司 | For handling method, apparatus, electronic equipment and the computer readable storage medium of voice dialogue |
CN109522550A (en) * | 2018-11-08 | 2019-03-26 | 和美(深圳)信息技术股份有限公司 | Text information error correction method, device, computer equipment and storage medium |
CN109754791A (en) * | 2017-11-03 | 2019-05-14 | 财团法人资讯工业策进会 | Acoustic-controlled method and system |
CN109951354A (en) * | 2019-03-12 | 2019-06-28 | 北京奇虎科技有限公司 | A kind of terminal device recognition methods, system and storage medium |
CN109947264A (en) * | 2017-12-21 | 2019-06-28 | 北京搜狗科技发展有限公司 | A kind of information display method, device and electronic equipment |
CN110120986A (en) * | 2018-02-05 | 2019-08-13 | 腾讯科技(深圳)有限公司 | Acquisition methods, device and the equipment of a kind of electronic equipment information |
CN110211592A (en) * | 2019-05-17 | 2019-09-06 | 北京华控创为南京信息技术有限公司 | Intelligent sound data processing equipment and method |
CN110232921A (en) * | 2019-06-21 | 2019-09-13 | 深圳市酷开网络科技有限公司 | Voice operating method, apparatus, smart television and system based on service for life |
CN110689891A (en) * | 2019-11-20 | 2020-01-14 | 广东奥园奥买家电子商务有限公司 | Voice interaction method and device based on public display device |
CN110909127A (en) * | 2019-11-07 | 2020-03-24 | 中铁大桥科学研究院有限公司 | Method and system for inputting and inquiring bridge inspection information |
CN111160013A (en) * | 2019-12-30 | 2020-05-15 | 北京百度网讯科技有限公司 | Text error correction method and device |
CN112580324A (en) * | 2020-12-24 | 2021-03-30 | 北京百度网讯科技有限公司 | Text error correction method and device, electronic equipment and storage medium |
CN113420547A (en) * | 2021-08-25 | 2021-09-21 | 深圳市豪华科技有限公司 | Wrongly written word error correction method of instant messaging software and related equipment |
CN113763944A (en) * | 2020-09-29 | 2021-12-07 | 浙江思考者科技有限公司 | AI video cloud interactive system based on simulation person logic knowledge base |
WO2022135414A1 (en) * | 2020-12-24 | 2022-06-30 | 深圳Tcl新技术有限公司 | Speech recognition result error correction method and apparatus, and terminal device and storage medium |
CN115438995A (en) * | 2022-09-21 | 2022-12-06 | 青岛酷特智能股份有限公司 | Service processing method and equipment for garment customization enterprise based on knowledge graph |
Families Citing this family (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN109977412B (en) * | 2019-03-29 | 2022-12-27 | 北京林业大学 | Method and device for correcting field value of voice recognition text and storage controller |
CN110232129B (en) * | 2019-06-11 | 2020-09-29 | 北京百度网讯科技有限公司 | Scene error correction method, device, equipment and storage medium |
CN110516237B (en) * | 2019-08-15 | 2022-12-09 | 重庆长安汽车股份有限公司 | Short text phrase extraction method, system and storage medium |
CN113808577A (en) * | 2021-09-18 | 2021-12-17 | 平安银行股份有限公司 | Intelligent extraction method and device of voice abstract, electronic equipment and storage medium |
Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050049868A1 (en) * | 2003-08-25 | 2005-03-03 | Bellsouth Intellectual Property Corporation | Speech recognition error identification method and system |
CN1979638A (en) * | 2005-12-02 | 2007-06-13 | 中国科学院自动化研究所 | Method for correcting error of voice identification result |
CN101154228A (en) * | 2006-09-27 | 2008-04-02 | 西门子公司 | Partitioned pattern matching method and device thereof |
CN101188109A (en) * | 2006-11-20 | 2008-05-28 | 富士通株式会社 | Speech recognition apparatus, speech recognition method and recording medium having speech recognition software program recorded thereon |
CN101655837A (en) * | 2009-09-08 | 2010-02-24 | 北京邮电大学 | Method for detecting and correcting error on text after voice recognition |
CN102682763A (en) * | 2011-03-10 | 2012-09-19 | 北京三星通信技术研究有限公司 | Method, device and terminal for correcting named entity vocabularies in voice input text |
CN103514882A (en) * | 2012-06-30 | 2014-01-15 | 北京百度网讯科技有限公司 | Voice identification method and system |
CN104269168A (en) * | 2014-09-24 | 2015-01-07 | 上海伯释信息科技有限公司 | Voice recognition method with higher work efficiency |
CN104871239A (en) * | 2013-01-16 | 2015-08-26 | 夏普株式会社 | Electronic apparatus and vacuum cleaner |
CN105374356A (en) * | 2014-08-29 | 2016-03-02 | 株式会社理光 | Speech recognition method, speech assessment method, speech recognition system, and speech assessment system |
CN105488027A (en) * | 2015-11-30 | 2016-04-13 | 百度在线网络技术(北京)有限公司 | Keyword pushing method and apparatus |
CN105632499A (en) * | 2014-10-31 | 2016-06-01 | 株式会社东芝 | Method and device for optimizing voice recognition result |
CN106098060A (en) * | 2016-05-19 | 2016-11-09 | 北京搜狗科技发展有限公司 | The correction processing method of voice and device, the device of correction process for voice |
Family Cites Families (3)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US7058575B2 (en) * | 2001-06-27 | 2006-06-06 | Intel Corporation | Integrating keyword spotting with graph decoder to improve the robustness of speech recognition |
CN103186523B (en) * | 2011-12-30 | 2017-05-10 | 富泰华工业(深圳)有限公司 | Electronic device and natural language analyzing method thereof |
CN106710592B (en) * | 2016-12-29 | 2021-05-18 | 北京奇虎科技有限公司 | Voice recognition error correction method and device in intelligent hardware equipment |
-
2016
- 2016-12-29 CN CN201611246867.0A patent/CN106710592B/en active Active
-
2017
- 2017-12-14 WO PCT/CN2017/116165 patent/WO2018121275A1/en active Application Filing
Patent Citations (13)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20050049868A1 (en) * | 2003-08-25 | 2005-03-03 | Bellsouth Intellectual Property Corporation | Speech recognition error identification method and system |
CN1979638A (en) * | 2005-12-02 | 2007-06-13 | 中国科学院自动化研究所 | Method for correcting error of voice identification result |
CN101154228A (en) * | 2006-09-27 | 2008-04-02 | 西门子公司 | Partitioned pattern matching method and device thereof |
CN101188109A (en) * | 2006-11-20 | 2008-05-28 | 富士通株式会社 | Speech recognition apparatus, speech recognition method and recording medium having speech recognition software program recorded thereon |
CN101655837A (en) * | 2009-09-08 | 2010-02-24 | 北京邮电大学 | Method for detecting and correcting error on text after voice recognition |
CN102682763A (en) * | 2011-03-10 | 2012-09-19 | 北京三星通信技术研究有限公司 | Method, device and terminal for correcting named entity vocabularies in voice input text |
CN103514882A (en) * | 2012-06-30 | 2014-01-15 | 北京百度网讯科技有限公司 | Voice identification method and system |
CN104871239A (en) * | 2013-01-16 | 2015-08-26 | 夏普株式会社 | Electronic apparatus and vacuum cleaner |
CN105374356A (en) * | 2014-08-29 | 2016-03-02 | 株式会社理光 | Speech recognition method, speech assessment method, speech recognition system, and speech assessment system |
CN104269168A (en) * | 2014-09-24 | 2015-01-07 | 上海伯释信息科技有限公司 | Voice recognition method with higher work efficiency |
CN105632499A (en) * | 2014-10-31 | 2016-06-01 | 株式会社东芝 | Method and device for optimizing voice recognition result |
CN105488027A (en) * | 2015-11-30 | 2016-04-13 | 百度在线网络技术(北京)有限公司 | Keyword pushing method and apparatus |
CN106098060A (en) * | 2016-05-19 | 2016-11-09 | 北京搜狗科技发展有限公司 | The correction processing method of voice and device, the device of correction process for voice |
Cited By (29)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2018121275A1 (en) * | 2016-12-29 | 2018-07-05 | 北京奇虎科技有限公司 | Method and apparatus for error connection of voice recognition in smart hardware device |
CN107239547A (en) * | 2017-06-05 | 2017-10-10 | 北京智能管家科技有限公司 | Voice error correction method, terminal and storage medium for ordering song by voice |
CN107239547B (en) * | 2017-06-05 | 2019-05-28 | 北京儒博科技有限公司 | Voice error correction method, terminal and storage medium for ordering song by voice |
CN109754791A (en) * | 2017-11-03 | 2019-05-14 | 财团法人资讯工业策进会 | Acoustic-controlled method and system |
CN108021554A (en) * | 2017-11-14 | 2018-05-11 | 无锡小天鹅股份有限公司 | Audio recognition method, device and washing machine |
CN107977356A (en) * | 2017-11-21 | 2018-05-01 | 新疆科大讯飞信息科技有限责任公司 | Method and device for correcting recognized text |
CN109947264A (en) * | 2017-12-21 | 2019-06-28 | 北京搜狗科技发展有限公司 | A kind of information display method, device and electronic equipment |
CN109947264B (en) * | 2017-12-21 | 2023-03-14 | 北京搜狗科技发展有限公司 | Information display method and device and electronic equipment |
CN110120986B (en) * | 2018-02-05 | 2021-10-19 | 腾讯科技(深圳)有限公司 | Method, device and equipment for acquiring electronic equipment information |
CN110120986A (en) * | 2018-02-05 | 2019-08-13 | 腾讯科技(深圳)有限公司 | Acquisition methods, device and the equipment of a kind of electronic equipment information |
CN108647346A (en) * | 2018-05-15 | 2018-10-12 | 苏州东巍网络科技有限公司 | A kind of the elderly's voice interactive method and system for wearable electronic |
CN108877792B (en) * | 2018-05-30 | 2023-10-24 | 北京百度网讯科技有限公司 | Method, apparatus, electronic device and computer readable storage medium for processing voice conversations |
CN108877792A (en) * | 2018-05-30 | 2018-11-23 | 北京百度网讯科技有限公司 | For handling method, apparatus, electronic equipment and the computer readable storage medium of voice dialogue |
CN109522550A (en) * | 2018-11-08 | 2019-03-26 | 和美(深圳)信息技术股份有限公司 | Text information error correction method, device, computer equipment and storage medium |
CN109951354A (en) * | 2019-03-12 | 2019-06-28 | 北京奇虎科技有限公司 | A kind of terminal device recognition methods, system and storage medium |
CN109951354B (en) * | 2019-03-12 | 2021-08-10 | 北京奇虎科技有限公司 | Terminal equipment identification method, system and storage medium |
CN110211592A (en) * | 2019-05-17 | 2019-09-06 | 北京华控创为南京信息技术有限公司 | Intelligent sound data processing equipment and method |
CN110232921A (en) * | 2019-06-21 | 2019-09-13 | 深圳市酷开网络科技有限公司 | Voice operating method, apparatus, smart television and system based on service for life |
CN110909127A (en) * | 2019-11-07 | 2020-03-24 | 中铁大桥科学研究院有限公司 | Method and system for inputting and inquiring bridge inspection information |
CN110689891A (en) * | 2019-11-20 | 2020-01-14 | 广东奥园奥买家电子商务有限公司 | Voice interaction method and device based on public display device |
CN111160013A (en) * | 2019-12-30 | 2020-05-15 | 北京百度网讯科技有限公司 | Text error correction method and device |
CN111160013B (en) * | 2019-12-30 | 2023-11-24 | 北京百度网讯科技有限公司 | Text error correction method and device |
CN113763944A (en) * | 2020-09-29 | 2021-12-07 | 浙江思考者科技有限公司 | AI video cloud interactive system based on simulation person logic knowledge base |
WO2022135414A1 (en) * | 2020-12-24 | 2022-06-30 | 深圳Tcl新技术有限公司 | Speech recognition result error correction method and apparatus, and terminal device and storage medium |
CN112580324A (en) * | 2020-12-24 | 2021-03-30 | 北京百度网讯科技有限公司 | Text error correction method and device, electronic equipment and storage medium |
CN112580324B (en) * | 2020-12-24 | 2023-07-25 | 北京百度网讯科技有限公司 | Text error correction method, device, electronic equipment and storage medium |
CN113420547A (en) * | 2021-08-25 | 2021-09-21 | 深圳市豪华科技有限公司 | Wrongly written word error correction method of instant messaging software and related equipment |
CN115438995A (en) * | 2022-09-21 | 2022-12-06 | 青岛酷特智能股份有限公司 | Service processing method and equipment for garment customization enterprise based on knowledge graph |
CN115438995B (en) * | 2022-09-21 | 2023-10-10 | 青岛酷特智能股份有限公司 | Business processing method and equipment for clothing customization enterprise based on knowledge graph |
Also Published As
Publication number | Publication date |
---|---|
WO2018121275A1 (en) | 2018-07-05 |
CN106710592B (en) | 2021-05-18 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN106710592A (en) | Speech recognition error correction method and speech recognition error correction device used for intelligent hardware equipment | |
CN104166462B (en) | The input method and system of a kind of word | |
CN110675854B (en) | Chinese and English mixed speech recognition method and device | |
CN107729321A (en) | A kind of method for correcting error of voice identification result | |
WO2016067418A1 (en) | Conversation control device and conversation control method | |
CN109686361A (en) | A kind of method, apparatus of speech synthesis calculates equipment and computer storage medium | |
CN111415656B (en) | Speech semantic recognition method, device and vehicle | |
US20060229864A1 (en) | Method, device, and computer program product for multi-lingual speech recognition | |
CN103309926A (en) | Chinese and English-named entity identification method and system based on conditional random field (CRF) | |
US11093110B1 (en) | Messaging feedback mechanism | |
CN107451119A (en) | Method for recognizing semantics and device, storage medium, computer equipment based on interactive voice | |
CN109614623B (en) | Composition processing method and system based on syntactic analysis | |
CN109524008A (en) | A kind of audio recognition method, device and equipment | |
Álvarez et al. | Towards customized automatic segmentation of subtitles | |
Thomas et al. | Code-switching in Indic Speech Synthesisers. | |
JP2016001242A (en) | Question sentence creation method, device, and program | |
CN105895076B (en) | A kind of phoneme synthesizing method and system | |
CN110175337A (en) | A kind of textual presentation method and device | |
Goldman et al. | The SIWIS database: a multilingual speech database with acted emphasis | |
Tsiakoulis et al. | Dialogue context sensitive HMM-based speech synthesis | |
CN113177114A (en) | Natural language semantic understanding method based on deep learning | |
Saunders et al. | Towards using prosody to scaffold lexical meaning in robots | |
CN109871528A (en) | The method for recognizing semantics and device of voice data, storage medium, computer equipment | |
CN112530406A (en) | Voice synthesis method, voice synthesis device and intelligent equipment | |
CN106294306B (en) | A kind of information processing method and electronic equipment |
Legal Events
Date | Code | Title | Description |
---|---|---|---|
PB01 | Publication | ||
PB01 | Publication | ||
SE01 | Entry into force of request for substantive examination | ||
GR01 | Patent grant | ||
GR01 | Patent grant |