CN107316639A - A kind of data inputting method and device based on speech recognition, electronic equipment - Google Patents
A kind of data inputting method and device based on speech recognition, electronic equipment Download PDFInfo
- Publication number
- CN107316639A CN107316639A CN201710358744.4A CN201710358744A CN107316639A CN 107316639 A CN107316639 A CN 107316639A CN 201710358744 A CN201710358744 A CN 201710358744A CN 107316639 A CN107316639 A CN 107316639A
- Authority
- CN
- China
- Prior art keywords
- recognition result
- voice
- word
- words
- words recognition
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 50
- 238000004590 computer program Methods 0.000 claims description 8
- 235000013399 edible fruits Nutrition 0.000 claims description 5
- 230000006399 behavior Effects 0.000 claims 1
- 238000010586 diagram Methods 0.000 description 6
- 238000005516 engineering process Methods 0.000 description 2
- 230000004927 fusion Effects 0.000 description 1
- 238000007689 inspection Methods 0.000 description 1
- 230000000750 progressive effect Effects 0.000 description 1
- 238000012163 sequencing technique 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
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
-
- 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
- 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
- G10L2015/088—Word spotting
-
- 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
- G10L15/22—Procedures used during a speech recognition process, e.g. man-machine dialogue
- G10L2015/225—Feedback of the input speech
Landscapes
- Engineering & Computer Science (AREA)
- Computational Linguistics (AREA)
- Health & Medical Sciences (AREA)
- Audiology, Speech & Language Pathology (AREA)
- Human Computer Interaction (AREA)
- Physics & Mathematics (AREA)
- Acoustics & Sound (AREA)
- Multimedia (AREA)
- User Interface Of Digital Computer (AREA)
Abstract
The invention provides a kind of data inputting method based on speech recognition, the problem of phonetic entry efficiency is low in the prior art is solved.Methods described includes：Input voice is identified, at least two voice identification results are obtained；Every institute's speech recognition result is the words recognition result sequence that each the corresponding words recognition result of word voice is constituted in the input voice；Determine whole words recognition results in corresponding at least two voice identification results of each word voice；According to the order in the words recognition result sequence, the corresponding at least part words recognition result of each word is shown；According to selection operation of the user to the words recognition result of display, information to be entered is generated.Method disclosed by the invention, information to be entered is constituted by the way that the corresponding multiple words recognition results of each word voice are shown, and according to the selection of user, is accurately generated information to be entered and is completed input, further increases the efficiency of phonetic entry.
Description
Technical field
Field, more particularly to a kind of data inputting method and dress based on speech recognition are inputted the present invention relates to information
Put, electronic equipment.
Background technology
With the development of speech recognition technology, increasing electronic equipment or defeated using row information is entered by speech recognition
Enter.Pronunciation inputting method of the prior art is usually：First, the speech data of user's input is obtained；Then, to the language of acquisition
Sound data are identified；Finally, voice identification result is directly inputted.Or, when voice identification result has it is multiple when, with list shape
Formula shows multiple recognition results, then, the recognition result selected according to the selection input user of user.But, when user's input
Voice is nonstandard or during input homonym, the recognition result of some or some words may be inaccurate or there are multiple candidates
Recognition result, therefore, the voice identification result or the recognition result of list display directly inputted may not be that user really will
The information of input, in this case, user can only delete the voice identification result of identification mistake, then re-start voice defeated
Enter.In the prior art, also a kind of way is to show a plurality of tabular form of voice identification result one, selects defeated for user
Enter.This method equally exists the problem of a plurality of voice identification result is not user's information really to be inputted.
It can be seen that, the data inputting method of the prior art based on speech recognition is at least present：Due to single word voice
Recognition result is inaccurate or there are multiple candidates, causes input or shows the recognition result of user inaccurate, it is necessary to user's weight
New input, the problem of input efficiency is low caused by.
The content of the invention
Technical problems to be solved in this application are：Data inputting method of the prior art based on speech recognition is present
Input efficiency it is low the problem of.
In order to solve the above problems, the embodiments of the invention provide a kind of data inputting method based on speech recognition, bag
Include：
Input voice is identified, at least two voice identification results are obtained；Wherein, every institute's speech recognition result
For the words recognition result sequence that the corresponding words recognition result of each word voice is constituted in the input voice；
Determine whole words recognition knots in corresponding at least two voice identification results of each word voice
Really；
According to the input sequence of the word voice, the corresponding at least part words recognition of each word voice is shown
As a result；
According to selection operation of the user at least part words recognition result of display, information to be entered is generated.
On the other hand, the embodiments of the invention provide a kind of message input device based on speech recognition, including：
Sound identification module, for input voice to be identified, obtains at least two voice identification results；Wherein, often
Bar institute speech recognition result is the word that each the corresponding words recognition result of word voice is constituted in the input voice
Recognition result sequence；
Words recognition result determining module, for determining that corresponding at least two voices of each word voice are known
Whole words recognition results in other result；
Words recognition result display module, for the input sequence according to the word voice, shows the words recognition
As a result the corresponding at least part words recognition result of each word voice that determining module is determined；
Input information generating module, for described in being shown according to user to the words recognition result display module at least
The selection operation of partial words recognition result, generates information to be entered.
The third aspect, the embodiments of the invention provide a kind of electronic equipment, including memory, processor and is stored in described
On memory and the computer program that can run on a processor, this hair is realized during computer program described in the computing device
The data inputting method based on speech recognition described in bright embodiment.
The third aspect, the embodiments of the invention provide a kind of computer-readable recording medium, is stored thereon with computer journey
Sequence, the program realizes the step of the data inputting method based on speech recognition described in the embodiment of the present invention when being executed by processor
Suddenly.
Data inputting method based on speech recognition disclosed in embodiments of the invention, by knowing to input voice
Not, at least two voice identification results are obtained；Wherein, every institute's speech recognition result is each word in the input voice
The words recognition result sequence of the corresponding words recognition result composition of language voice；Determine that each word voice is corresponding described
Whole words recognition results at least two voice identification results；And according to the order in the words recognition result sequence,
Show the corresponding at least part words recognition result of each word；Finally, according to words recognition result of the user to display
Selection operation, generate information to be entered, solve that data inputting method based on speech recognition in the prior art is present is defeated
The problem of entering inefficiency.By the way that the corresponding multiple words recognition results of each word voice are shown, and according to user
Selection composition information to be entered, accurately generate information to be entered and complete input, further increase the efficiency of phonetic entry.
Brief description of the drawings
In order to illustrate the technical solution of the embodiments of the present invention more clearly, below will be in embodiment or description of the prior art
The required accompanying drawing used is briefly described, it should be apparent that, drawings in the following description are only some realities of the present invention
Example is applied, for those of ordinary skill in the art, without having to pay creative labor, can also be attached according to these
Figure obtains other accompanying drawings.
Fig. 1 is the data inputting method flow chart based on speech recognition of the embodiment of the present invention one；
Fig. 2 is the data inputting method flow chart based on speech recognition of the embodiment of the present invention two；
Fig. 3 is one of words recognition result display schematic diagram in embodiment two；
Fig. 4 is one of words recognition result display schematic diagram in embodiment two；
Fig. 5 is the data inputting method flow chart based on speech recognition of the embodiment of the present invention three；
Fig. 6 is one of words recognition result display schematic diagram in embodiment three；
Fig. 7 is two of words recognition result display schematic diagram in embodiment three；
Fig. 8 is one of the message input device structural representation of the embodiment of the present invention four based on speech recognition；
Fig. 9 is the two of message input device structural representation of the embodiment of the present invention four based on speech recognition；
Figure 10 is the three of message input device structural representation of the embodiment of the present invention four based on speech recognition.
Embodiment
Below in conjunction with the accompanying drawing in the embodiment of the present invention, the technical scheme in the embodiment of the present invention is carried out clear, complete
Site preparation is described, it is clear that described embodiment is a part of embodiment of the invention, rather than whole embodiments.Based on this hair
Embodiment in bright, the every other implementation that those of ordinary skill in the art are obtained under the premise of creative work is not made
Example, belongs to the scope of protection of the invention.
Embodiment one
A kind of data inputting method based on speech recognition disclosed by the invention, as shown in figure 1, methods described includes:Step
Rapid 100 to step 130.
Step 100, input voice is identified, obtains at least two voice identification results.
Wherein, every institute's speech recognition result is the corresponding words recognition of each word voice in the input voice
As a result the words recognition result sequence constituted.
When user is by application, the phonetic entry control of such as interface of input method is inputted after voice, carried by calling system
The interface of confession, can obtain the input voice of user.Then, using will further call input of the speech recognition engine to acquisition
Voice is identified, to obtain voice identification result.
Generally, speech recognition engine carries out segment processing according to the acoustic characteristic of speech data to the voice of input first,
Then, each section of voice is identified respectively.User, would generally be by of short duration pronunciation between different terms when inputting voice
Interval, therefore, after segment processing, every section of voice will correspond to the voice of a word.Speech recognition engine is to every section of language
After sound is identified, it will obtain the words recognition result of each word voice.Because different user stings the difference of word and accent,
When it is implemented, same word voice may recognize arrive multiple words recognition results.Or, input unisonance for user
The situation of word or homonym, speech recognition engine, which may be recognized, obtains multiple words recognition results.Therefore, generally, each word
Language voice will obtain at least one words recognition result.
Then, speech recognition engine by language model to the corresponding words recognition knot of each word voice in input voice
Fruit is modified, and is further combined into sequence of terms.When some word voice has multiple words recognition results, speech recognition
Engine will obtain inputting at least two voice identification results of voice.
Step 110, whole words in corresponding at least two voice identification results of each word voice are determined
Language recognition result.
Speech recognition engine, with reference to language model, obtains at least two voices according to the words recognition result of every section of voice
After recognition result, at least two voice identification results exported according to speech recognition engine can determine that every section of voice is right respectively
The words recognition result answered, i.e., each word voice distinguishes corresponding words recognition result.The language exported with speech recognition engine
Sound recognition result is：" Can ' t you help me " and " exemplified by Can you help me ", it may be determined that input voice in：The
The corresponding words recognition result of one word voice is：" the corresponding words recognition of Can ' t " and " Can ", second word voice
As a result it is：" you ", the corresponding words recognition result of the 3rd word voice are：" help ", the 4th corresponding word of word voice
Language recognition result is：“me”.When it is implemented, the data format of the voice identification result exported according to speech recognition engine can be with
It is determined that each word voice distinguishes corresponding words recognition result.
Step 120, according to the input sequence of the word voice, the corresponding at least part of each word voice is shown
Words recognition result.
It may be one to input the corresponding words recognition result of each word voice in voice, it is also possible to be multiple, in order to
It is easy to by selecting the corresponding words recognition result of each word voice to constitute the corresponding input information of the input voice, this
Invention is shown the corresponding words recognition result of each word voice using corresponding mode.For example：According to words recognition
As a result sequencing of the corresponding word voice in input voice, the word of each word voice is shown in corresponding list of locations
Language recognition result.Such as the corresponding words recognition result of some word voice is more than 1, then the word is shown in the form of list
The corresponding whole words recognition results of voice or the partial words recognition result for recognizing highest scoring.Or, such as some word language
The corresponding words recognition result of sound is more than 1, then the corresponding whole words recognitions of the word voice are shown in the form of driver plate
As a result or identification highest scoring partial words recognition result.Or, such as corresponding words recognition result of some word voice is big
In 1, then use and highlight the words recognition result that form shows identification highest scoring in the word voice, to indicate
User can select other words recognition results of the word voice by triggering the words recognition result.
Step 130, according to selection operation of the user to the words recognition result of display, information to be entered is generated.
User can select the words recognition result of display by operations such as click, slips, then the word knowledge by selecting
Other result composition phrase or sentence, are used as information to be entered.
Data inputting method based on speech recognition disclosed in the embodiment of the present invention, by the way that input voice is identified,
Obtain at least two voice identification results；Wherein, every institute's speech recognition result is each word in the input voice
The words recognition result sequence of the corresponding words recognition result composition of voice；Determine each word voice it is corresponding it is described extremely
Whole words recognition results in few two voice identification results；And according to the order in the words recognition result sequence, show
Show the corresponding at least part words recognition result of each word；Finally, according to user to the words recognition result of display
Selection operation, generates information to be entered, solves the input that the data inputting method based on speech recognition is present in the prior art
The problem of inefficiency.By the way that the corresponding multiple words recognition results of each word voice are shown, and according to user's
Selection composition information to be entered, accurately generates information to be entered and completes input, further increase the efficiency of phonetic entry.
Embodiment two
A kind of data inputting method based on speech recognition disclosed in the embodiment of the present invention, as shown in Fig. 2 methods described bag
Include:Step 200 is to step 230.
Step 200, input voice is identified, obtains at least two voice identification results.
Wherein, every institute's speech recognition result is the corresponding words recognition of each word voice in the input voice
As a result the words recognition result sequence constituted.
Input voice is identified, the embodiment of at least two voice identification results is obtained referring to embodiment
One, here is omitted.
Step 210, whole words in corresponding at least two voice identification results of each word voice are determined
Language recognition result.
Determine whole words recognition knots in corresponding at least two voice identification results of each word voice
The embodiment of fruit is referring to embodiment one, and here is omitted.
In the present embodiment, the voice identification result exported using speech recognition engine as：" Can ' tyouhitit " and " Can
Exemplified by you heat it ", it may be determined that in input voice：The corresponding words recognition result of first word voice is：“Can’
The corresponding words recognition result of t " and " Can ", second word voice is：The corresponding word of " you ", the 3rd word voice is known
Other result is：" hit " and " heat ", the corresponding words recognition result of the 4th word voice are：“it”.
Step 220, according to the input sequence of the word voice, the same word voice of correspondence is shown side by side in relevant position
Multiple words recognition results.
Know when in the voice identification result that speech recognition engine is exported, being present in the corresponding multiple words of same word voice
During other result, the step of showing each word voice corresponding at least part words recognition result further comprises：In phase
Position is answered to show multiple words recognition results of the same word voice of correspondence side by side.
Still by taking the words recognition result determined in previous step as an example, first word voice and the 3rd word voice point
Dui Ying not multiple words recognition results " Can ' t " and " Can ", therefore, in first position display arranged side by side, first word voice
Corresponding two words recognition results, the corresponding two words recognition knots of the 3rd word voice are shown the 3rd position side by side
Really " hit " and " heat ".Second word voice and the 4th word voice are corresponded with words recognition result respectively, in phase
Position is answered only to show a words recognition result.Concrete term recognition result display schematic diagram is as shown in Figure 3.
When it is implemented, the input sequence in order to intuitively embody word voice, and constitute defeated by words recognition result
Enter the possibility of information, can be connected words recognition result by arrow, formed word chain, as shown in figure 4, i.e. with
The form of word chain shows the corresponding words recognition result of each word voice；Wherein, by the same word voice of correspondence
Multiple words recognition results as different terms chain node arranged side by side；Will be with the one-to-one words recognition knot of word voice
Really as the common node of the different terms chain.The word is known it can be seen from the words recognition result shown in Fig. 4
The information for the input that other result can be constituted has four kinds, is respectively：“Can’t you hit it”、“Can you heat it”、
" Can ' t you heat it " and " Can you hit it ".By showing the same word voice of correspondence side by side in relevant position
Multiple words recognition results, add user input information washability.
Step 230, according to user to the selection operation of the multiple words recognition result shown side by side, generation includes institute
State the information to be entered for the words recognition result that selection operation is directed to.
When it is implemented, being corresponded to when in the voice identification result that speech recognition engine is exported, being present in same word voice
Multiple words recognition results when, according to selection operation of the user at least part words recognition result of display, generation
Information to be entered, further comprises：According to user to the selection operation of the multiple words recognition result shown side by side, generation
The information to be entered for the words recognition result being directed to including the selection operation.
When it is implemented, the selection operation includes：The words recognition result shown side by side is clicked on, performs and slides hand
The words recognition result of the gesture Jing Guo showing side by side and with the one-to-one words recognition result of word voice.With Fig. 3 and Fig. 4
Exemplified by the words recognition result of middle display, user can by clicking on successively " Can ' t " and " hit ", select words recognition result,
Then, according to user to the selection operation of the multiple words recognition result shown side by side, generation includes the selection operation
For words recognition result " Can ' t " and " hit " information to be entered, such as：“Can’t you hit it”.User can be with
Words recognition result is selected by performing slip gesture.For example, the word chain that user is shown by finger along along Fig. 4 is slided
Dynamic, finger sequentially passes through node where first to fourth node on word chain, i.e. " Can ", " you ", " heat ", " it ",
The words recognition result of the node passed through by the slide, constitutes information to be entered：“Can you heat it”.
User can select the words recognition result of display by operations such as click, slips, then the word knowledge by selecting
Other result composition phrase or sentence, are used as information to be entered.
Data inputting method based on speech recognition disclosed in the embodiment of the present invention, by the way that input voice is identified,
Obtain at least two voice identification results；Wherein, every institute's speech recognition result is each word in the input voice
The words recognition result sequence of the corresponding words recognition result composition of voice；Determine each word voice it is corresponding it is described extremely
Whole words recognition results in few two voice identification results；According to the input sequence of the word voice, in relevant position
Multiple words recognition results of the same word voice of correspondence are shown side by side；Finally, it is the multiple to what is shown side by side according to user
The selection operation of words recognition result, generation includes the information to be entered for the words recognition result that the selection operation is directed to, and solves
The problem of data inputting method based on speech recognition is present in the prior art the input efficiency of having determined is low.By by each word
The corresponding multiple words recognition results of language voice are shown, and constitute information to be entered according to the selection of user, are accurately generated
Information to be entered simultaneously completes input, further increases the efficiency of phonetic entry.Also, by showing multiple words recognitions side by side
As a result, it is effectively improved the washability of information to be entered.
Embodiment three
Correspondingly, the embodiment of the invention also discloses a kind of data inputting method based on speech recognition, as shown in figure 5,
Methods described includes:Step 500 is to step 540.
Step 500, input voice is identified, obtains at least two voice identification results.
Wherein, every institute's speech recognition result is the corresponding words recognition of each word voice in the input voice
As a result the words recognition result sequence constituted.
Input voice is identified, the embodiment of at least two voice identification results is obtained referring to embodiment
One, here is omitted.
Step 510, whole words in corresponding at least two voice identification results of each word voice are determined
Language recognition result.
Determine whole words recognition knots in corresponding at least two voice identification results of each word voice
The embodiment of fruit is referring to embodiment one, and here is omitted.
In the present embodiment, the voice identification result exported using speech recognition engine as：" Can ' t you he lp me " and
" exemplified by Can you help me ", it may be determined that in input voice：The corresponding words recognition result of first word voice is：
" the corresponding words recognition result of Can ' t " and " Can ", second word voice is：" you ", the 3rd word voice are corresponding
Words recognition result is：" help ", the corresponding words recognition result of the 4th word voice are：“me”.
Step 520, according to the input sequence of the word voice, show corresponding with same word voice with the first pattern
Multiple words recognition results, and shown and the one-to-one words recognition result of word voice with the second pattern.
Wherein, first pattern is different from second pattern, and first pattern is used to indicate the words recognition knot
Fruit editable.
Know when in the voice identification result that speech recognition engine is exported, being present in the corresponding multiple words of same word voice
During other result, the step of showing each word voice corresponding at least part words recognition result further comprises：With
One pattern shows multiple words recognition results corresponding with same word voice, and is shown with the second pattern with word voice one by one
Corresponding words recognition result.
Still by taking the words recognition result determined in previous step as an example, first word voice and the 3rd word voice point
Not Dui Ying multiple words recognition results " Can ' t " and " Can ", therefore, first word are shown first position with the first pattern
The corresponding two words recognition results of language voice.Second word voice, the 3rd word voice and the 4th word voice point
Do not corresponded with words recognition result, one words recognition result is shown with the second pattern in relevant position.When it is implemented,
First pattern is the Show Styles for being different from the second pattern, and such as the first pattern is：Be highlighted, underscore form is shown,
Bubble is shown.Fig. 6 is that the first pattern is that underscore form is particularly shown schematic diagram.
When it is implemented, being corresponded to when in the voice identification result that speech recognition engine is exported, being present in same word voice
Multiple words recognition results when, according to selection operation of the user at least part words recognition result of display, generation
Information to be entered, further comprises：According to the trigger action of user, the words recognition result pair that the trigger action is directed to is shown
The candidate's word recognition result answered, candidate's word recognition result is selected from the words recognition result that the trigger action is directed to
Multiple words recognition results of corresponding word voice；The selection of candidate's word recognition result of display is grasped according to user
Make, generate information to be entered.
Step 530, according to the trigger action of user, the corresponding time of words recognition result that the trigger action is directed to is shown
Select words recognition result.
When it is implemented, when the words recognition result that user's triggering is shown with the first pattern, as clicked in Fig. 6
" Can ' t " then eject 700 in drop-down menu or subwindow, such as Fig. 7, user are shown in the drop-down menu or subwindow
Triggering words recognition result " the corresponding candidate's word recognition results of Can ' t ", such as：“Can”.
Step 540, according to selection operation of the user to candidate's word recognition result of display, generation includes the choosing
Select the information to be entered for candidate's word recognition result that operation is directed to.
User is detected to the selection operation of the candidate's word recognition result shown in drop-down menu or pop-up window, such as：Inspection
Survey user and click on the operation of some candidate's word recognition result, then, candidate's word recognition result that user is clicked on as with
Candidate's word recognition result of family selection, and the candidate's word recognition result selected according to user and other words recognition results are common
With composition information to be entered.Wherein, candidate's word recognition result is selected from the words recognition knot that the trigger action is directed to
Really " multiple words recognition results of the corresponding word voices of Can ' t ", such as：“Can’t”、“Can”、“Count”.To be used in Fig. 7
Words recognition result is clicked at family, and " after Can ' t ", ejection drop-down menu shows exemplified by " Can " and " Count " that user can be further
Candidate's word recognition result " Can " is clicked on, to select " Can " as the words recognition result of first word voice.Finally, root
The candidate's word recognition result " Can " selected according to user and other words recognition results：" you ", " help " and " me " common group
Into information to be entered：“Can you help me”.
When it is implemented, according to selection operation of the user at least part words recognition result of display, generation
During information to be entered, when the trigger action according to user, the corresponding time of words recognition result that the trigger action is directed to is shown
After the step of selecting words recognition result, methods described also includes：Phonetic entry control is shown, for re-entering the triggering
Operate the phonetic entry icon 710 in the corresponding word voice of words recognition result being directed to, such as Fig. 7.By showing that voice is defeated
Voice can be re-entered with instruction user by entering control, and current term recognition result is modified.By detecting user couple in real time
The trigger action of the phonetic entry control, and obtain the output of phonetic entry control, that is, the word voice re-entered.
It is described by the words recognition result of the word voice re-entered, update the word that the trigger action is directed to and know
The corresponding candidate's word recognition result of other result, including：The word voice re-entered is entered with reference to the input voice
Row speech recognition, the words recognition result updated；By the words recognition result of the renewal, update trigger action and be directed to
The corresponding candidate's word recognition result of words recognition result.After the word voice re-entered is got, by again
The word voice of input is replaced and is originally inputted corresponding segment in voice, by the voice re-entered and original to input voice progress
Voice is merged, and calls speech recognition engine that the voice after fusion is identified again, the word language re-entered
The renewal words recognition result of sound.Finally, by the words recognition result of the renewal, the word knowledge that trigger action is directed to is updated
The corresponding candidate's word recognition result of other result.
Data inputting method based on speech recognition disclosed in the embodiment of the present invention, by the way that input voice is identified,
Obtain at least two voice identification results；Wherein, every institute's speech recognition result is each word in the input voice
The words recognition result sequence of the corresponding words recognition result composition of voice；Determine each word voice it is corresponding it is described extremely
Whole words recognition results in few two voice identification results；According to the input sequence of the word voice, with the first pattern
Display multiple words recognition results corresponding with same word voice, and show one-to-one with word voice with the second pattern
Words recognition result；Then, according to the trigger action of user, show that the words recognition result that the trigger action is directed to is corresponding
Candidate's word recognition result；Finally, according to selection operation of the user to candidate's word recognition result of display, generation treats defeated
Enter information, solve the problem of data inputting method based on speech recognition is present in the prior art input efficiency is low.
By showing the words recognition result with multiple candidates with different patterns, using Show Styles as instruction user and
Whether words recognition result may be selected, and when needing selection, just triggering shows candidate's word recognition result, can effectively save
Show the display space that words recognition result takes.By showing phonetic entry control simultaneously in display candidate's word recognition result
Part, is easy to user when the voice identification result of word voice is not that user really thinks the information of input, quick to update input language
Sound, it is the efficiency that is transfused to information further to improve by voice.
Example IV
Correspondingly, the embodiment of the invention also discloses a kind of message input device based on speech recognition, as shown in figure 8,
Described device includes:
Sound identification module 810, for input voice to be identified, obtains at least two voice identification results；Wherein,
Every institute's speech recognition result is the word that each the corresponding words recognition result of word voice is constituted in the input voice
Language recognition result sequence；
Words recognition result determining module 820, for determining corresponding at least two languages of each word voice
Whole words recognition results in sound recognition result；
Words recognition result display module 830, for the input sequence according to the word voice, shows that the word is known
The corresponding at least part words recognition result of each word voice that other result determining module 820 is determined；
Information generating module 840 is inputted, for the institute shown according to user to the words recognition result display module 830
The selection operation of at least part words recognition result is stated, information to be entered is generated.
Optionally, as shown in figure 9, the words recognition result display module 830 includes：
First display unit 8301, multiple words recognitions for showing the same word voice of correspondence side by side in relevant position
As a result.
Optionally, the input information generating module 840 further comprises：
First input information generating unit 8401, for according to user to the multiple words recognition result that shows side by side
Selection operation, generation include the information to be entered for the words recognition result that the selection operation is directed to.
Optionally, the selection operation includes：Click on the words recognition result shown side by side, perform slip gesture warp
Cross the words recognition result that shows side by side and with the one-to-one words recognition result of word voice.
Optionally, as shown in Figure 10, the words recognition result display module 830 includes：
Second display unit 8302, for showing multiple words recognition knots corresponding with same word voice with the first pattern
Really, and with the second pattern show and the one-to-one words recognition result of word voice；Wherein, first pattern is used to indicate
The words recognition result editable.
Optionally, the words recognition result display module 830 further comprises：
3rd display unit 8303, for the trigger action according to user, shows that the word that the trigger action is directed to is known
The corresponding candidate's word recognition result of other result, candidate's word recognition result is selected from the word that the trigger action is directed to
Multiple words recognition results of the corresponding word voice of recognition result；
Optionally, the input information generating module 840 further comprises：
Second input information generating unit 8402, for the choosing according to user to candidate's word recognition result of display
Operation is selected, generation includes the information to be entered generation information to be entered for candidate's word recognition result that the selection operation is directed to.
Optionally, as shown in Figure 10, the words recognition result display module 830 also includes：
4th display unit 8304, for showing phonetic entry control, is directed to for re-entering the trigger action
The corresponding word voice of words recognition result；
Words recognition result updating block 8305, for the words recognition result of the word voice by re-entering, more
The corresponding candidate's word recognition result of words recognition result that the new trigger action is directed to.
Optionally, the words recognition result updating block 8305 specifically for：
Speech recognition is carried out to the word voice re-entered with reference to the input voice, the word updated is known
Other result；By the words recognition result of the renewal, the corresponding candidate word of words recognition result that trigger action is directed to is updated
Language recognition result.
Message input device based on speech recognition disclosed in the embodiment of the present invention, by the way that input voice is identified,
Obtain at least two voice identification results；Wherein, every institute's speech recognition result is each word in the input voice
The words recognition result sequence of the corresponding words recognition result composition of voice；Determine each word voice it is corresponding it is described extremely
Whole words recognition results in few two voice identification results；According to the input sequence of the word voice, with the first pattern
Display multiple words recognition results corresponding with same word voice, and show one-to-one with word voice with the second pattern
Words recognition result；Then, according to the trigger action of user, show that the words recognition result that the trigger action is directed to is corresponding
Candidate's word recognition result；Finally, according to selection operation of the user to candidate's word recognition result of display, generation treats defeated
Enter information, solve the problem of data inputting method based on speech recognition is present in the prior art input efficiency is low.
By showing the words recognition result with multiple candidates with different patterns, using Show Styles as instruction user and
Whether words recognition result may be selected, and when needing selection, just triggering shows candidate's word recognition result, can effectively save
Show the display space that words recognition result takes.By showing phonetic entry control simultaneously in display candidate's word recognition result
Part, is easy to user when the voice identification result of word voice is not that user really thinks the information of input, quick to update input language
Sound, it is the efficiency that is transfused to information further to improve by voice.
Correspondingly, invention additionally discloses a kind of electronic equipment, including memory, processor and be stored in the memory
Computer program that is upper and can running on a processor, realizes that the present invention is implemented described in the computing device during computer program
The data inputting method based on speech recognition described in example one to embodiment three.The electronic equipment can be：It is smart mobile phone, flat
Plate computer, game machine, electronic reader etc..
Correspondingly, invention additionally discloses a kind of computer-readable recording medium, be stored thereon with computer program, the journey
The data inputting method based on speech recognition described in the embodiment of the present invention one to embodiment three is realized when sequence is executed by processor
The step of.
Each embodiment in this specification is described by the way of progressive, what each embodiment was stressed be with
Between the difference of other embodiment, each embodiment identical similar part mutually referring to.For device and electronics
For apparatus embodiments, because it is substantially similar to embodiment of the method, so description is fairly simple, related part is referring to method
The part explanation of embodiment.
The application is carried out to a kind of data inputting method based on speech recognition, device and electronic equipment of offer above
It is discussed in detail, specific case used herein is set forth to the principle and embodiment of the application, above example
Explanation be only intended to help and understand the present processes and its core concept；Simultaneously for those of ordinary skill in the art,
According to the thought of the application, it will change in specific embodiments and applications, in summary, in this specification
Hold the limitation that should not be construed as to the application.
Through the above description of the embodiments, those skilled in the art can be understood that each embodiment can
Realized by the mode of software plus required general hardware platform, naturally it is also possible to realized by hardware.Based on such reason
Solution, the part that above-mentioned technical proposal substantially contributes to prior art in other words can be embodied in the form of software product
Come, the computer software product can be stored in a computer-readable storage medium, such as ROM/RAM, magnetic disc, CD, including
Some instructions are to cause a computer equipment (can be personal computer, server, or network equipment etc.) is performed each
Method described in some parts of individual embodiment or embodiment.
Claims (11)
1. a kind of data inputting method based on speech recognition, it is characterised in that including：
Input voice is identified, at least two voice identification results are obtained；Wherein, every institute's speech recognition result is served as reasons
The words recognition result sequence of the corresponding words recognition result composition of each word voice in the input voice；
Determine whole words recognition results in corresponding at least two voice identification results of each word voice；
According to the input sequence of the word voice, the corresponding at least part words recognition knot of each word voice is shown
Really；
According to selection operation of the user at least part words recognition result of display, information to be entered is generated.
2. the method as described in claim 1, it is characterised in that the corresponding at least part of the display each word voice
The step of words recognition result, including：
Show multiple words recognition results of the same word voice of correspondence side by side in relevant position.
3. method as claimed in claim 2, it is characterised in that described to be known according to user at least part word of display
The step of selection operation of other result, generation information to be entered, including：
According to user to the selection operation of the multiple words recognition result shown side by side, generation includes the selection operation pin
To words recognition result information to be entered.
4. method as claimed in claim 3, it is characterised in that the selection operation includes：Click on the institute's predicate shown side by side
Language recognition result, perform slip gesture by the words recognition result that shows side by side and with the one-to-one word of word voice
Language recognition result.
5. the method as described in claim 1, it is characterised in that the corresponding at least part of the display each word voice
The step of words recognition result, including：
Multiple words recognition results corresponding with same word voice are shown with the first pattern, and shown and word with the second pattern
The one-to-one words recognition result of voice；Wherein, first pattern is used to indicate the words recognition result editable.
6. method as claimed in claim 5, it is characterised in that described to be known according to user at least part word of display
The step of selection operation of other result, generation information to be entered, including：
According to the trigger action of user, the corresponding candidate's words recognition knot of words recognition result that the trigger action is directed to is shown
Really, candidate's word recognition result is selected from many of the corresponding word voice of words recognition result that the trigger action is directed to
Individual words recognition result；
According to selection operation of the user to candidate's word recognition result of display, generation includes what the selection operation was directed to
The information to be entered of candidate's word recognition result.
7. method as claimed in claim 6, it is characterised in that the trigger action according to user, shows the triggering behaviour
After the step of making the words recognition result that is directed to corresponding candidate's word recognition result, in addition to：
Phonetic entry control is shown, for re-entering the corresponding word language of words recognition result that the trigger action is directed to
Sound；
By the words recognition result of the word voice re-entered, the words recognition result pair that the trigger action is directed to is updated
The candidate's word recognition result answered.
8. method as claimed in claim 7, it is characterised in that described by the words recognition knot of the word voice re-entered
Really, the step of updating words recognition result that the trigger action is directed to corresponding candidate's word recognition result, including：
Speech recognition, the words recognition knot updated are carried out to the word voice re-entered with reference to the input voice
Really；
By the words recognition result of the renewal, the corresponding candidate's word knowledge of words recognition result that trigger action is directed to is updated
Other result.
9. a kind of message input device based on speech recognition, it is characterised in that including：
Sound identification module, for input voice to be identified, obtains at least two voice identification results；Wherein, every institute
Speech recognition result is the words recognition that each the corresponding words recognition result of word voice is constituted in the input voice
As a result sequence；
Words recognition result determining module, for determining the corresponding at least two speech recognition knots of each word voice
Whole words recognition results in fruit；
Words recognition result display module, for the input sequence according to the word voice, shows the words recognition result
The corresponding at least part words recognition result of each word voice that determining module is determined；
Information generating module is inputted, for the described at least part shown according to user to the words recognition result display module
The selection operation of words recognition result, generates information to be entered.
10. a kind of electronic equipment, including memory, processor and it is stored on the memory and can runs on a processor
Computer program, it is characterised in that realize claim 1 to 8 any one described in the computing device during computer program
The data inputting method based on speech recognition described in claim.
11. a kind of computer-readable recording medium, is stored thereon with computer program, it is characterised in that the program is by processor
The step of data inputting method based on speech recognition described in claim 1 to 8 any one claim is realized during execution.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710358744.4A CN107316639A (en) | 2017-05-19 | 2017-05-19 | A kind of data inputting method and device based on speech recognition, electronic equipment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201710358744.4A CN107316639A (en) | 2017-05-19 | 2017-05-19 | A kind of data inputting method and device based on speech recognition, electronic equipment |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN107316639A true CN107316639A (en) | 2017-11-03 |
Family
ID=60181507
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201710358744.4A Pending CN107316639A (en) | 2017-05-19 | 2017-05-19 | A kind of data inputting method and device based on speech recognition, electronic equipment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN107316639A (en) |
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109215661A (en) * | 2018-08-30 | 2019-01-15 | 上海与德通讯技术有限公司 | Speech-to-text method, apparatus equipment and storage medium |
| CN110473543A (en) * | 2019-09-25 | 2019-11-19 | 北京蓦然认知科技有限公司 | A kind of audio recognition method, device |
| CN111475129A (en) * | 2019-01-24 | 2020-07-31 | 北京京东尚科信息技术有限公司 | Method and equipment for displaying candidate homophones through voice recognition |
| CN116825107A (en) * | 2023-08-21 | 2023-09-29 | 北京集度科技有限公司 | Voice interaction method and device, electronic equipment and storage medium |
Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923854A (en) * | 2010-08-31 | 2010-12-22 | 中国科学院计算技术研究所 | Interactive speech recognition system and method |
| CN102903361A (en) * | 2012-10-15 | 2013-01-30 | Itp创新科技有限公司 | Instant call translation system and instant call translation method |
| CN103366741A (en) * | 2012-03-31 | 2013-10-23 | 盛乐信息技术（上海）有限公司 | Voice input error correction method and system |
| CN104538032A (en) * | 2014-12-19 | 2015-04-22 | 中国科学院计算技术研究所 | Chinese voice recognition method and system fusing user feedback |
| CN105161094A (en) * | 2015-06-26 | 2015-12-16 | 徐信 | System and method for manually adjusting cutting point in audio cutting of voice |
| CN105244026A (en) * | 2015-08-24 | 2016-01-13 | 陈娟 | Voice processing method and device |
| CN105513589A (en) * | 2015-12-18 | 2016-04-20 | 百度在线网络技术（北京）有限公司 | Speech recognition method and speech recognition device |
| CN106683662A (en) * | 2015-11-10 | 2017-05-17 | 中国电信股份有限公司 | Speech recognition method and device |
-
2017
- 2017-05-19 CN CN201710358744.4A patent/CN107316639A/en active Pending
Patent Citations (8)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101923854A (en) * | 2010-08-31 | 2010-12-22 | 中国科学院计算技术研究所 | Interactive speech recognition system and method |
| CN103366741A (en) * | 2012-03-31 | 2013-10-23 | 盛乐信息技术（上海）有限公司 | Voice input error correction method and system |
| CN102903361A (en) * | 2012-10-15 | 2013-01-30 | Itp创新科技有限公司 | Instant call translation system and instant call translation method |
| CN104538032A (en) * | 2014-12-19 | 2015-04-22 | 中国科学院计算技术研究所 | Chinese voice recognition method and system fusing user feedback |
| CN105161094A (en) * | 2015-06-26 | 2015-12-16 | 徐信 | System and method for manually adjusting cutting point in audio cutting of voice |
| CN105244026A (en) * | 2015-08-24 | 2016-01-13 | 陈娟 | Voice processing method and device |
| CN106683662A (en) * | 2015-11-10 | 2017-05-17 | 中国电信股份有限公司 | Speech recognition method and device |
| CN105513589A (en) * | 2015-12-18 | 2016-04-20 | 百度在线网络技术（北京）有限公司 | Speech recognition method and speech recognition device |
Non-Patent Citations (2)
| Title |
|---|
| JUN OGATA等: ""Speech Repair: Quick Error Correction Just by Using Selection Operation for Speech Input Interfaces"", 《INTERSPEECH-2005:9TH EUROPEAN CONFERENCE ON SPEECH COMMNUNICATION AND TECHNOLOGY》 * |
| KEITH VERTANEN等: ""Intelligently Aiding Human-Guided Correction of Speech Recognition"", 《TWENTY-FOURTH AAAI CONFERENCE ON ARTIFICIAL INTELLIGENCE》 * |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109215661A (en) * | 2018-08-30 | 2019-01-15 | 上海与德通讯技术有限公司 | Speech-to-text method, apparatus equipment and storage medium |
| CN111475129A (en) * | 2019-01-24 | 2020-07-31 | 北京京东尚科信息技术有限公司 | Method and equipment for displaying candidate homophones through voice recognition |
| CN110473543A (en) * | 2019-09-25 | 2019-11-19 | 北京蓦然认知科技有限公司 | A kind of audio recognition method, device |
| CN116825107A (en) * | 2023-08-21 | 2023-09-29 | 北京集度科技有限公司 | Voice interaction method and device, electronic equipment and storage medium |
| CN116825107B (en) * | 2023-08-21 | 2023-11-21 | 北京集度科技有限公司 | Voice interaction method and device, electronic equipment and storage medium |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| TWI437449B (en) | Multi-mode input method and input method editor system | |
| TWI510965B (en) | Input method editor integration | |
| JP2021089739A (en) | Question answering method and language model training method, apparatus, device, and storage medium | |
| US9569231B2 (en) | Device, system, and method for providing interactive guidance with execution of operations | |
| TWI443551B (en) | Method and system for an input method editor and computer program product | |
| CN107316639A (en) | A kind of data inputting method and device based on speech recognition, electronic equipment | |
| US8370808B2 (en) | Apparatus and a method for generating a test case | |
| US20140236596A1 (en) | Emotion detection in voicemail | |
| WO2018075191A1 (en) | Techniques for scheduling language models and character recognition models for handwriting inputs | |
| CN105161095B (en) | Method and device for picture composition of speech recognition syntax tree | |
| CN112286485B (en) | Method and device for controlling application through voice, electronic equipment and storage medium | |
| CN101989128B (en) | Method and device for showing candidate words in fixed position | |
| CN110060674A (en) | Form management method, apparatus, terminal and storage medium | |
| JP2010044637A (en) | Data processing apparatus, method, and program | |
| CN106484131A (en) | A kind of input error correction method and input subtraction unit | |
| CN114168718A (en) | Information processing apparatus, method and information recording medium | |
| CN112541052B (en) | Method, device, equipment and storage medium for determining answers to questions | |
| CN115879469B (en) | Text data processing method, model training method, device and medium | |
| US11928444B2 (en) | Editing files using a pattern-completion engine implemented using a machine-trained model | |
| KR20210127613A (en) | Method and apparatus for generating conversation, electronic device and storage medium | |
| CN111880668A (en) | Input display method and device and electronic equipment | |
| CN107247518A (en) | Method, device, equipment and the readable storage medium storing program for executing of polynary association's input | |
| CN108351710A (en) | Continuous keyboard identification | |
| CN104317416B (en) | A kind of method and apparatus of input | |
| CN104077105B (en) | A kind of information processing method and a kind of electronic equipment |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| RJ01 | Rejection of invention patent application after publication | ||
| RJ01 | Rejection of invention patent application after publication |
Application publication date: 20171103 |