CN109840062A - Auxiliary input device and recording medium - Google Patents

Auxiliary input device and recording medium Download PDF

Info

Publication number
CN109840062A
CN109840062A CN201711211098.5A CN201711211098A CN109840062A CN 109840062 A CN109840062 A CN 109840062A CN 201711211098 A CN201711211098 A CN 201711211098A CN 109840062 A CN109840062 A CN 109840062A
Authority
CN
China
Prior art keywords
variable slot
item value
mentioned
option
pronunciation
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
Application number
CN201711211098.5A
Other languages
Chinese (zh)
Other versions
CN109840062B (en
Inventor
酒主佳祐
宫部泰成
矶部庄三
后藤和之
杉浦千加志
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Toshiba Corp
Toshiba Digital Solutions Corp
Original Assignee
Toshiba Corp
Toshiba Digital Solutions Corp
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Toshiba Corp, Toshiba Digital Solutions Corp filed Critical Toshiba Corp
Priority to CN201711211098.5A priority Critical patent/CN109840062B/en
Publication of CN109840062A publication Critical patent/CN109840062A/en
Application granted granted Critical
Publication of CN109840062B publication Critical patent/CN109840062B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Abstract

A kind of auxiliary input device and recording medium, even if the item value in the recognition result based on voice recognition comprising can also suitably determine to input towards the variable slot of formatted data in the case where mistake.The auxiliary input device of embodiment has template storage unit, acquisition unit and determination section.Template storage unit storage format template, the format module is that the option of the variable slot of option type of item value is selected with the template with the entry name accordingly formatted data of 1 or more variable slot of cuit value, the entry name of each variable slot and from multiple options to be described together with respective pronunciation.Acquisition unit obtains recognition result data, which is recognition result data obtained from the sound as issuing to user carries out voice recognition, includes record and pronunciation.The pronunciation of pronunciation and the entry name and option that in above-mentioned format module describe of the determination section based on above-mentioned recognition result data determines the item value that the variable slot towards above-mentioned formatted data inputs.

Description

Auxiliary input device and recording medium
Technical field
Embodiments of the present invention are related to auxiliary input device and recording medium.
Background technique
In the past, there is known the technologies assisted using voice recognition technology come the input operation to user.Carry out sound The Audio Recognition Engine of identification is roughly divided into grammer type and dictation (dictation) type.Grammer type is the rule based on pre-production Then (grammer) carries out the mode of voice recognition, is easy to improve accuracy of identification, but presence can not identify the speech not followed the principles Feature.Also, the production needs professional knowledge relevant to voice recognition of rule, there is also suitably make, management rule Operation becomes many and diverse disadvantage.On the other hand, listening rewritable type is gradually to identify the method arbitrarily made a speech not based on specific rule, Versatility is high, but there are the features that accuracy of identification is low compared with grammer type.
Herein, consider through voice recognition technology come to variable slot (slot) input item towards formatted data (formdata) The operation of mesh value is assisted.In the present specification, formatted data refers to the data that format determines, is with corresponding with entry name The data of 1 or more variable slot of ground cuit value.For example, to the operator by having carried out live (field) operation In the case that the input of the formatted data towards workpaper book carried out is assisted, all expect to abide by for all operators It follows this case that the speech of rule and unrealistic, considers to wish in the voice recognition of speech using listening the voice recognition of rewritable type to draw It holds up.But listen the Audio Recognition Engine of rewritable type accuracy of identification as described above low, therefore can not determine to tie identification in many cases Which of fruit is partially used as the item value of which variable slot to input.Even if accordingly, it is desired to provide one kind is in recognition result Also the new technology of the item value inputted towards the variable slot of formatted data can be suitably determined in the case where comprising mistake.
Summary of the invention
Project to be solved by this invention is to provide a kind of auxiliary input device and recording medium, even if being based on sound Sound identification recognition result in comprising mistake in the case where can also suitably determine towards formatted data variable slot input item Mesh value.
The auxiliary input device of embodiment has template storage unit, acquisition unit and determination section.The storage of template storage unit Format module, the format module are with the formatted data with entry name accordingly 1 or more variable slot of cuit value Template, the entry name of each variable slot and from multiple options select item value option type variable slot option with it is respective Pronunciation is described together.Acquisition unit obtains recognition result data, the recognition result data be sound by being issued to user into Recognition result data obtained from row voice recognition include record and pronunciation.Reading of the determination section based on above-mentioned recognition result data The pronunciation of method and the entry name described in above-mentioned format module and option determines that the variable slot towards above-mentioned formatted data inputs Item value.
According to the auxiliary input device of above structure, even if the feelings comprising mistake in the recognition result based on voice recognition Under condition, the item value inputted towards the variable slot of formatted data can be also suitably determined.
Detailed description of the invention
Fig. 1 is the figure for showing an example of formatted data.
Fig. 2 is the figure for showing an example of format module.
Fig. 3 is the figure for showing an example of recognition result data.
Fig. 4 is the figure for showing an example of the formatted data after having input item value.
Fig. 5 is the block diagram for showing the functional structure example of auxiliary input device.
Fig. 6 is the flow chart for showing an example of the processing sequence carried out by extraction unit.
Fig. 7 is the figure for showing an example of candidate list.
Fig. 8 is the flow chart for showing an example of the processing sequence carried out by option type item value determination section.
Fig. 9 is the flow chart for showing an example of the processing sequence carried out by free description type item value determination section.
Figure 10 is the flow chart for showing an example of the processing sequence carried out by Template Manager portion.
Figure 11 is the figure for showing an example of formatted data that the item value input of a part of variable slot finishes.
Figure 12 is the figure for showing an example of format module.
Figure 13 is the figure for showing an example of recognition result data.
Figure 14 is the figure for showing an example of the formatted data in the variable slot not inputted after cuit value.
Figure 15 is the block diagram for showing the hardware configuration example of auxiliary input device.
Specific embodiment
Hereinafter, being described in detail referring to auxiliary input device and program of the attached drawing to embodiment.
The summary > of < embodiment
The auxiliary input device of present embodiment is used to be identified as obtained from the speech progress voice recognition to user Result data come determine towards formatted data variable slot input item value.Imagination is known in the sound that the speech for user carries out Using in not can be not based on the Audio Recognition Engine for listening rewritable type for specifically regularly carrying out voice recognition, in recognition result number Situation comprising mistake in.To enable being suitably determined according to such recognition result data comprising mistake towards lattice The item value of the variable slot input of formula data uses the format module made in advance by template construct person in the present embodiment. In addition, hereinafter, formatted data as process object, the workpaper of the illustrated job result for reporting field operation The formatted data of book, but the formatted data that can be applicable in is not limited to this.
Formatted data is the data that format determines, accordingly has 1 or more the change for item value input with entry name Measuring tank.Fig. 1 shows an example of formatted data.Formatted data A shown in FIG. 1 is the lattice of the workpaper book of not yet cuit value The example of formula data, the change of variable slot (variable slot No.1), project entitled " eye-witness " with project entitled " Work places " The variable slot (variable slot No.3) and entitled " the special thing of project of measuring tank (variable slot No.2), project entitled " replacement component " This 4 variable slot B of the variable slot (variable slot No.4) of item ".
Variable slot B includes option type variable slot B1 and free this 2 types of description type variable slot B2.Option type variable slot B1 It is the variable slot B that item value is selected from multiple options.Free description type variable slot B2 is the variable slot for being freely written item value B.In the formatted data A illustrated by Fig. 1,3 variable slot B of variable slot No.1 to variable slot No.3 are option type variable slots The variable slot B of B1, variable slot No.4 are free description type variable slot B2.As the item value of above-mentioned each variable slot B, input has base In the character string of the speech of operator.
Format module is the template of formatted data A, by the entry name and option of each variable slot B possessed by formatted data A The option of the item value of type variable slot B1 and respective pronunciation (it is shown in the corresponding Japanese specification of the present invention with hiragana, But shown in the present invention with the Roman character of small letter) it describes together.Fig. 2 shows an examples of format module.Format shown in Fig. 2 Template C is the example of format module corresponding with formatted data A illustrated by Fig. 1, for variable slot No.1 to variable slot No.3 3 option type variable slot B1, describing together with the option of entry name and item value has respective pronunciation, for variable slot No.4 Free description type variable slot B2, description have entry name and its pronunciation.Multiple readings can also be described for 1 entry name or option Method.Also, in format module C, describing to have indicates that each variable slot B is the kind of the variable slot of option type or free description type Class.
Format module C is as described above by template construct person's pre-production.Format module C includes mistake as basis is used for Recognition result data come determine towards formatted data A each variable slot B input item value hint information utilize, but do not need into Rule in the Audio Recognition Engine of row grammer type makes such detailed parameter setting relevant to voice recognition.That is, mould Even if plate producer does not have professional knowledge relevant to voice recognition, only by by each variable slot B of corresponding formatted data A Entry name and the option and pronunciation of option type variable slot B1 set can easily make format module C together, and Also it is easy to carry out its management.
An example of recognition result data obtained from carrying out voice recognition as the speech to user is shown in Fig. 3.Fig. 3's (a) an example of the speech D of the workpaper book for making the operator by carrying out field operation is shown, (b) of Fig. 3 is shown An example of the recognition result data E as obtained from carrying out voice recognition using speech D of the Audio Recognition Engine to (a) of Fig. 3. As shown in (b) of Fig. 3, record and pronunciation are included in recognition result data E.In the example shown in Fig. 3, it makes a speech " Kawasaki " of D It is misidentified as " rock is rugged ", " air filter " for the D that makes a speech is misidentified as " so filter ", is wrapped in recognition result data E Containing mistake.
In addition, the pronunciation of recognition result data E can also and not used Audio Recognition Engine, but is separately arranged The pronunciation information generation unit that pronunciation is generated according to the record of the recognition result data E of Audio Recognition Engine output, uses the reading The output of method information generation unit.Also, in the present embodiment, the speech D as shown in (a) of Fig. 3 comprising multiple sentences is by one Secondary input Audio Recognition Engine and obtained the recognition result data E as (b) of Fig. 3, but the timing of voice recognition be appoint Meaning.For example, it is also possible to be formed as following structure: each sentence for constituting speech D is gradually inputted Audio Recognition Engine, in real time Ground carries out voice recognition.
The auxiliary input device of present embodiment for example obtains the recognition result data as illustrated by (b) of Fig. 3 E, pronunciation based on recognition result data E and such format module C is described as illustrated in Figure 2 entry name and The pronunciation of option determines the item value of each variable slot B input towards such formatted data A as illustrated in Figure 1.Fig. 4 is shown Have input an example of the formatted data A of the item value determined by the auxiliary input device of present embodiment.Pass through present embodiment Auxiliary input device processing, as shown in figure 4, in option type variable slot B1 (the variable slot of project entitled " Work places " No.1 in), " Kawasaki " is inputted as item value, in the option type variable slot B1 (variable slot No.2) of project entitled " eye-witness " In, " middle village " is inputted as item value, in the option type variable slot B1 (variable slot No.3) of project entitled " replacement component ", is made " air filter " is inputted for item value.Also, in the free description type variable slot B2 of project entitled " special item ", as Item value input " special item: the closedown school for being required film on the move stumbles and almost falls ".
For being input to the item value " river of the option type variable slot B1 (variable slot No.1) of project entitled " Work places " It is rugged ", due to including the pronunciation " ka-wa- with the option " Kawasaki " described in format module C in the pronunciation of recognition result data E Sa-ki " similar " wa-sa-ki ", therefore selected as the item value of option type variable slot B1.But " Kawasaki " exists It also include that therefore, it is necessary to determine will be in the option of the option type variable slot B1 (variable slot No.2) of project entitled " eye-witness " The option type variable slot B1 of " wa-sa-ki " corresponding " Kawasaki " as which of variable slot No.1 and variable slot No.2 Item value.Also, the option type variable slot B1 (variable slot No.2) for being input to project entitled " eye-witness " item value " in Village ", due to including the pronunciation " na-ka- with the option " middle village " described in format module C in the pronunciation of recognition result data E Mu-ra " similar " ka-mu-ra ", therefore selected as the item value of option type variable slot B1.But in entry name In option for the option type variable slot B1 (variable slot No.2) of " eye-witness ", also comprising the pronunciation institute with recognition result data E " wa-sa-ki " for including corresponding " Kawasaki ", it is therefore desirable to determine to regard which of " middle village " and " Kawasaki " as variable slot The item value of the option type variable slot B1 of No.2.For the auxiliary input device of present embodiment, even if being input to like this There are in the case where multiple candidates of competing conjunction, can also suitably determine towards each variable slot B input for the item value of variable slot B Item value.
The concrete example > of < auxiliary input device
Secondly, the concrete example of the auxiliary input device of present embodiment is described in detail.Fig. 5 is to show this embodiment party The figure of the functional structure example of the auxiliary input device 100 of formula.The auxiliary input device 100 of present embodiment is for example as shown in Figure 5 Have formatted data storage unit 10, template storage unit 20, acquisition unit 30, determination section 40 and Template Manager portion 50.
10 storage format data A of formatted data storage unit.For example, making format module C (referring to figure according to template construct person 2) this case generates formatted data A (referring to Fig.1) corresponding with format module C, not yet cuit value, and stores In formatted data storage unit 10.In turn, each variable slot B of the formatted data A by being stored to format data store 10 is defeated Enter the item value determined by determination section 40, generate the formatted data A for having input item value (referring to Fig. 4).Have input item value Formatted data A can be read from formatted data storage unit 10 as needed, and with for example carried out by display display, towards note The arbitrary way of output output such as the data transmission of the output for printing of record paper, outwardly part device.
20 storage format template C of template storage unit.Template construct person is able to use Template Manager portion 50 and makes format module C, the format module C that should be made by template construct person are incorporated in template storage unit 20.The format that template storage unit 20 is stored Template C is read by determination section 40.In the case that template storage unit 20 is stored with multiple format module C, determination section 40 for example may be used According to the operation for the user for selecting format module C to be used or according to the recognition result data E obtained with acquisition unit 30 Between consistent degree, select the format module C that reads from template storage unit 20.In addition, template construct person is also able to use mould Board management portion 50 carries out editor or deletion of the format module C that template storage unit 20 stores etc..
Acquisition unit 30 is obtained through Audio Recognition Engine to user's (work to carry out field operation in the present embodiment Dealer) speech carry out recognition result data E (referring to Fig. 3) obtained from voice recognition.The recognition result that acquisition unit 30 obtains Data E includes record and pronunciation as described above.Acquisition unit 30 can be by receiving the input for the sound that user is made a speech and using Internal Audio Recognition Engine carries out voice recognition to the sound inputted to obtain recognition result data E, can also obtain logical It crosses external Audio Recognition Engine and recognition result data E obtained from voice recognition is carried out to the sound that user issues.That is, being directed to The Audio Recognition Engine that the sound that user is issued carries out voice recognition can be 100 institute of auxiliary input device of present embodiment The structure having is also possible to such as the terminal installation that user uses or inputs providing the external server of voice recognition service The structure that device outside auxiliary device 100 has.The recognition result data E that acquisition unit 30 obtains is transferred to determination section 40。
The pronunciation for the recognition result data E that determination section 40 is obtained based on acquisition unit 30 and the lattice read from template storage unit 20 The pronunciation of entry name and option that formula template C is described determine towards formatted data A each variable slot B input item value. Determination section 40 includes extraction unit 41, option type item value determination section 42 and free description type item value determination section as shown in Figure 5 43, and handled according to the sequence implementation.
Extraction unit 41 is from the character string extraction and format module C of the pronunciation for the recognition result data E that acquisition unit 30 obtains The entry name of description or the pronunciation of option it is similar segment word column.In turn, extraction unit 41 will be from the reading of recognition result data E The entry name of segment word that the character string of method extracts column and the format module C used in the extraction of segment word column or Candidate of the combination of person's option as the corresponding part between recognition result data E and format module C, is registered in candidate list. The details of candidate list is aftermentioned.
Fig. 6 is the flow chart for showing an example of the processing sequence carried out by extraction unit 41.If processing shown in fig. 6 starts, Then extraction unit 41 is first by candidate list initialization (step S101).Secondly, the project that extraction unit 41 will describe in format module C 1 in name and option is set in variable TEXT (step S102).At this point, being set in the entry name or option of variable TEXT Pronunciation is also kept.
Secondly, extraction unit 41 from the character string of the pronunciation of recognition result data E extract with the pronunciation of TEXT it is completely the same or The consistent segment word column (step S103) in person part.At this point, extraction unit 41 can for example be removed to recognition result data E After removing filling word or parsing pre-treatment as the part removed other than the part for being considered as noun by morpheme, into The processing of row step S103.In turn, extraction unit 41 be directed to extract in step s 103 segment word column, calculate relative to The similar degree (step S104) of the pronunciation of TEXT.In the case where having extracted multiple portions character string in step s 103, In step S104, similar degree is calculated for various pieces character string.
As similar degree used herein, such as alpha-numeric, the segment word of the pronunciation relative to TEXT can be enumerated The alpha-numeric ratio etc. of column.Such as when the option for the variable slot No.3 for setting format module C shown in Fig. 2 in TEXT is " empty In the case where air filter ", in the character string of the pronunciation of recognition result data E shown in (b) from Fig. 3, extract " empty with option Part pronunciation " e-a-fi-ru-ta " consistent " fi-ru-ta " (the 1st row) of air filter " and " fi-ru " (the 3rd row).At this In the case of, the text number indicated with hiragana of " e-a-fi-ru-ta " is 6, the text indicated with hiragana of " fi-ru-ta " Number is 4, and the text number indicated with hiragana of " fi-ru " is 3, therefore can be calculated, and the similar degree of " fi-ru-ta " is 4/6= 0.67, the similar degree of " fi-ru " is 3/6=0.5.In this way, arranging for extracted segment word, the pronunciation with TEXT is calculated More similar is then the similar degree of bigger value.
In addition, the calculation method of similar degree is not limited to the above method, it is able to use well known technology.For example, it is also possible to It is using editor (levenshtein) distance (to the operation ability of the insertion of 2 character string application how many times texts, deletion, replacement As identical character string), (adjacent N number of character string or single contamination continuously wants N-gram in a certain character string Element) concordance rate calculate the structure of similar degree.
Secondly, extraction unit 41 is directed to the segment word column extracted in step s 103, judgement calculates in step S104 Similar degree whether be predetermined threshold value (such as 0.5) or more (step S105).In turn, when the similar degree of segment word column In the case where more than threshold value (step S105: yes), extraction unit 41 by the segment word arrange be set in TEXT entry name or The combination of person's option is registered in candidate list (step S106) as candidate.
On the other hand, in the case where the similar degree deficiency threshold value of segment word column (step S105: no), step is skipped The processing of S106, without the registration towards candidate list.In this way, being formed as only arranging the segment word that similar degree is threshold value or more It is registered in the structure of candidate list, can be for example as a result, 1 text or 2 texts by the text number indicated with hiragana Degree short segment word arrange from candidate remove, can effectively avoid handling the bad feelings for becoming many and diverse because candidate number is huge Condition.
Then, extraction unit 41 judges whether that being directed to all entry names described in format module C and option carries out Processing (step S107), untreated entry name or option (step S107: no) if it exists, then return step S102 lays equal stress on Later processing is carried out again.In turn, it has carried out when for all entry names and option described in format module C from step After processing until rapid S102 to step S106 (step S107: yes), terminate a series of processing shown in fig. 6.
Fig. 7 is the figure for showing an example of the candidate list F generated by the above-mentioned processing carried out by extraction unit 41.Fig. 7 Shown in candidate list F be based on recognition result data E shown in format module C and Fig. 3 (b) shown in Fig. 2 generate time Mend the example of list F.In candidate list F, it is registered with as candidate from the character string of the pronunciation of recognition result data E and extracts Segment word column and the segment word column extraction used in format module C entry name or option combination.It waits The registration data for mending list F for example includes candidate classification, variable slot No., entry name, item value, segment word as shown in Figure 7 Column, offset, similar degree, type of variable slot etc..
Candidate classification indicates that candidate is using entry name as object or using option (item value) as object.Variable slot No. indicate candidate using which variable slot B as object.Entry name indicates the entry name of corresponding variable slot B.Item value indicates to wait It mends using option (item value) as the option (item value) in the case where object.Offset indicates the part in recognition result data E The appearance position of character string indicates the beginning of the beginning texts of segment word column from the character string of the pronunciation of recognition result data E Which text rises is.Similar degree indicates the similar degree calculated in the step S104 of Fig. 6.The type of variable slot is indicated in candidate Corresponding variable slot B is option type variable slot B1 or free description type variable slot B2 in the case where using entry name as object. The registration data of each candidate comprising above- mentioned information is accordingly stored with the candidate ID foundation being uniquely assigned for each candidate. Hereinafter, by the candidate for being registered in candidate list F, using option (item value) as the candidate of object, (candidate classification is project The candidate of value) it is known as " item value candidate ", the candidate (candidate classification is candidate of entry name) using entry name as object is claimed For " entry name candidate ".
In addition, the processing of extraction unit 41 shown in fig. 6 and the structure of candidate list F shown in Fig. 7 are an example, and unlimited In this.As long as extraction unit 41 is extracted and the entry name of format module C or choosing from the character string of the pronunciation of recognition result data E The segment word that the pronunciation of item is similar arranges and is registered in the structure of candidate list F.As long as also, candidate list F can Know the structure of the segment word extracted by extraction unit 41 column and the corresponding relationship between entry name or option.In template In the case that storage unit 20 is stored with multiple format module C, candidate list F is also possible to also make with identification other than candidate ID The template ID of format module C establishes the structure for accordingly storing each registration data.
Option type item value determination section 42 is based on the segment word similar with the pronunciation of option described in format module C In positional relationship between the similar degree of column and segment word column and other parts character strings at least any one determine court The item value of the option type variable slot B1 input of formatted data A.
Fig. 8 is the flow chart for showing an example of the processing sequence carried out by option type item value determination section 42.If shown in Fig. 8 Processing start, then one of candidate list F item value candidate for being included is set in by option type item value determination section 42 first Variable V ALUE (step S201).
Secondly, option type item value determination section 42 determines whether that there are projects for the item value candidate for being set in VALUE Name or the identical competing conjunction candidate (step S202) of item value.For example, work as the candidate ID " 001 " of candidate list F shown in Fig. 7 In the case that item value candidate is set in VALUE, the item value candidate of the identical candidate ID " 003 " of item value " Kawasaki " becomes Competing conjunction candidate.Also, in the case that the item value candidate of candidate ID " 003 " is set in VALUE, item value " Kawasaki " is identical Candidate ID " 001 " item value candidate and the identical candidate ID " 004 " of entry name " eye-witness " item value candidate become it is competing Close candidate.Also, in the case that the item value candidate of candidate ID " 006 " is set in VALUE, item value " air filter " The item value candidate of identical candidate ID " 007 " becomes competing conjunction candidate.Also, when the item value candidate of candidate ID " 007 " is set In the case where VALUE, the item value candidate of the identical candidate ID " 006 " of item value " air filter " is waited as competing conjunction It mends.
If the judgement of step S202 the result is that competing conjunction candidate (step S202: no) is not present, option type item value determines Determine the item value (step S203) that the item value of VALUE is determined as corresponding variable slot B by portion 42.On the other hand, when there are competing In the case where conjunction candidate (step S202: yes), whether option type item value determination section 42 subsequently determines that the similar degree of VALUE than competing The similar degree for closing candidate is big (step S204).In turn, if the similar degree of VALUE it is bigger than the competing similar degree for closing candidate (step S204: It is), then step S203 is advanced to, the item value of VALUE is determined as to the item value of corresponding variable slot B.For example, when Fig. 7's In the case that the item value candidate of candidate ID " 006 " is set in VALUE, the similar degree 0.67 of VALUE is than being used as competing conjunction candidate Candidate ID " 007 " item value candidate similar degree 0.5 it is big, therefore, the item value " air filter " of VALUE is decided to be The item value of the option type variable slot B1 of variable slot No.3.
On the other hand, when the similar degree of VALUE be not larger than it is competing close candidate similar degree in the case where (step S204: It is no), option type item value determination section 42 subsequently determines that the similar degree of VALUE (step whether smaller than the competing similar degree for closing candidate S205).In turn, if the similar degree of VALUE is smaller (step S205: yes) than the competing similar degree for closing candidate, terminate to be directed to be set to The processing of the item value candidate of VALUE, advances to subsequent steps S211.
On the other hand, in the case that the similar degree of VALUE is small unlike the competing similar degree for closing candidate, i.e. the class of VALUE Like spend it is equal with the competing conjunction similar degree of candidate in the case where (step S205: no), option type item value determination section 42 subsequently determines that With the presence or absence of entry name candidate (step S206) corresponding with the entry name of VALUE.In turn, if it exists with the entry name of VALUE Corresponding entry name candidate (step S206: yes), then option type item value determination section 42 subsequently determines that VALUE and its entry name are waited Whether the difference of the offset between benefit is smaller (step S207) than the difference of the offset between competing conjunction candidate and its entry name candidate, if offset Difference it is smaller than competing conjunction candidate (step S207: yes), then advance to step S203, the item value of VALUE be determined as corresponding change The item value of measuring tank B.
For example, existing and VALUE in the case that the item value candidate of the candidate ID " 004 " of Fig. 7 is set in VALUE The corresponding candidate ID " 002 " of item value " eye-witness " entry name candidate, the offset of the VALUE relative to the entry name candidate Difference be 25-17=8, in contrast, competing conjunction candidate (the item value candidate of candidate ID " 001 " relative to the entry name candidate And the entry name candidate of candidate ID " 003 ") offset difference be 17-2=15, offset difference it is smaller than competing conjunction candidate.Therefore, The item value " middle village " of VALUE is decided to be the item value of the option type variable slot B1 of variable slot No.2.
On the other hand, when the difference of the offset between VALUE and its entry name candidate is competing conjunction candidate and its entry name candidate Between offset difference more than in the case where (step S207: no), terminate for be set in VALUE item value candidate place Reason, advances to subsequent steps S211.
Also, when there is no (step S206: no) in the case where entry name candidate corresponding with the entry name of VALUE, choosings Item type item value determination section 42 then determines whether there is entry name candidate (step corresponding with the competing conjunction entry name of candidate S208).In turn, entry name candidate (step S208: yes) corresponding with the competing conjunction entry name of candidate if it exists, then option type project Value determination section 42 then determines whether that there are entry name other item value candidates (step S209) identical with competing conjunction candidate.Into And such item value candidate (step S209: yes) if it exists, then option type item value determination section 42 subsequently determines that competing conjunction candidate Whether the difference of the offset between entry name candidate is bigger than the difference of the offset between other item value candidates and entry name candidate (step S210), if the competing difference for closing the offset between candidate and entry name candidate is than other item value candidates and entry name candidate Between offset difference it is big (step S210: yes), then advance to step S203, the item value of VALUE be determined as corresponding change The item value of measuring tank B.
For example, when Fig. 7 candidate ID " 001 " item value candidate be set in VALUE in the case where, as with as competing The corresponding entry name candidate of entry name of the item value candidate of the candidate ID " 003 " of candidate is closed, there are the projects of " 002 " candidate ID Name candidate.Also, as entry name with as the identical other projects of the competing conjunction item value candidate of candidate ID " 003 " of candidate It is worth candidate, there are the item value candidates of candidate ID " 004 ".In turn, the item value candidate as the competing candidate ID " 003 " for closing candidate The difference of offset between the entry name candidate of candidate ID " 002 " is 17-2=15, in contrast, is waited as other item values The difference of offset between the item value candidate of the candidate ID " 004 " of benefit and the entry name candidate of candidate ID " 002 " is 25-17=8, The competing difference for closing the offset between candidate and entry name candidate than the offset between other item value candidates and entry name candidate it It is poor big.Therefore, the item value " Kawasaki " of VALUE is decided to be the item value of the option type variable slot B1 of variable slot No.1.
In addition, when there is no (step S208: no) in the case where item value candidate corresponding with the competing conjunction entry name of candidate, Or there is no entry name others item value candidates identical with competing conjunction candidate in the case where it is (step S209: no) or competing The difference for closing the offset between candidate and entry name candidate is the difference of the offset between other item value candidates and entry name candidate In situation below (step S210: no), terminate to advance to step for the processing for the item value candidate for being set in VALUE S211。
Then, option type item value determination section 42 determines whether all item value candidates for including for candidate list F Processing (step S211) is had been carried out, if it exists untreated item value candidate (step S211: no), then return step S201 repeats later processing.In turn, when all item value candidates for including for candidate list F have been carried out After processing until from step S201 to step S210 (step S211: yes), the item value determined is input to formatted data A Corresponding option type variable slot B1 (step S212), terminate a series of processing shown in Fig. 8.
By option type item value determination section 42 carry out more than processing, for example, as shown in figure 4, in the change of formatted data A In the option type variable slot B1 of measuring tank No.1, " Kawasaki " is inputted as item value.Also, in the option type variable of variable slot No.2 In slot B1, " middle village " is inputted as item value.Also, it is defeated as item value in the option type variable slot B1 of variable slot No.3 Enter " air filter ".
In addition, the processing of option type item value determination section 42 shown in Fig. 8 is an example, however it is not limited to this.Option type item As long as (segment word that item value candidate includes is arranged relative to project similar degree of the mesh value determination section 42 based on item value candidate The similar degree of value) and the item value candidate and other item value candidate or entry name candidate between positional relationship it is (each to wait The difference of the offset of segment word that benefit includes column) at least any one structure to determine the item value of option type variable slot B1 ?.For example, in the above example, the size of similar degree is paid the utmost attention in compared between competing conjunction candidate, but can also The score for each candidate is calculated with comprehensive descision similar degree and positional relationship, option type is determined based on the score of each candidate The item value of variable slot B1.Also, the difference of the offset for the segment word column in the above example, including using each candidate carrys out table Show the positional relationship between candidate, but finger other the quantity of existing other candidates between 2 candidates also can be used Mark is to indicate the positional relationship between candidate.
Free description type item value determination section 43 is based on the item with the free description type variable slot B2 described in format module C The similar degree of the similar segment word column of the pronunciation of mesh name and the position of segment word column, according to the note of recognition result data E It carries, determines the item value that the free description type variable slot B2 towards formatted data A is inputted.
Fig. 9 is the flow chart for showing an example of the processing sequence carried out by free description type item value determination section 43.If Fig. 9 Shown in handle and start, then the entry name candidate that free description type item value determination section 43 includes with reference first to candidate list F The type of variable slot determines the entry name candidate (step S301) that whether there is free description type variable slot B2 in candidate list F. Herein, it if the entry name candidate (step S301: no) of free description type variable slot B2 is not present in candidate list F, directly ties Beam processing.
On the other hand, if in candidate list F in the presence of free description type variable slot B2 entry name candidate (step S301: It is), then free description type item value determination section 43 is then by the project for the free description type variable slot B2 for including in candidate list F One of name candidate is set in variable V ALUE (step S302).In turn, free description type item value determination section 43 is then for setting Determine whether that there are the identical competing conjunction candidate (steps of entry name in the entry name candidate of the free description type variable slot B2 of VALUE S303)。
Herein, the identical competing conjunction candidate (step S303: no) of entry name if it does not exist, then free description type item value determines Portion 43, which will correspond to from the segment word of VALUE, to be arranged to part corresponding with the entry name of other variable slot B or item value Character string it is tight before until the record of recognition result data E of part be determined as the item of corresponding free description type variable slot B2 Mesh value (step S304).In addition, in the pronunciation of recognition result data E, when after the segment word of VALUE column there is no with In the case where the corresponding segment word column of the entry name or item value of other variable slot B, arranged with from the segment word of VALUE The record of the corresponding recognition result data E in part until the end of pronunciation is decided to be the free description type variable slot B2's Item value.
For example, the entry name candidate for the free description type variable slot B2 for including in candidate list F shown in Fig. 7 is only candidate Entry name competing conjunction candidate identical with the entry name candidate is not present in the entry name candidate of ID " 008 ".Also, in recognition result In the pronunciation of data E, after the segment word column " to-kki-ji-ko-u " of the entry name candidate of candidate ID " 008 ", do not deposit It is arranged in segment word corresponding with the entry name of other variable slot B or item value.Thus, and from recognition result data E's The record of the corresponding recognition result data E in part of " to-kki-ji-ko-u " of pronunciation until end be " special item: The closedown school that film is required in movement stumbles and almost falls." it is decided to be the free description type variable of variable slot No.4 The item value of slot B2.
On the other hand, in the case where competing conjunction candidate identical there are entry name (step S303: yes), free description type item Mesh value determination section 43 subsequently determines that whether the similar degree of VALUE is bigger (step S305) than the competing similar degree for closing candidate.In turn, if The similar degree of VALUE is bigger (step S305: yes) than the competing similar degree for closing candidate, then advances to step S304, will correspond to from VALUE segment word column to segment word corresponding with the entry name of other variable slot B or item value arrange it is tight before be The record of the recognition result data E of part only is determined as the item value of corresponding free description type variable slot B2.
On the other hand, in the similar degree of VALUE is the competing similar degree situation below for closing candidate (step S305: no), The processing for terminating the entry name candidate for the free description type variable slot B2 for being set in VALUE, advances to step S306.
Then, free description type item value determination section 43 determines whether for all freedom for including in candidate list F The entry name candidate of description type variable slot B2 has been carried out processing (step S306), and untreated free description type becomes if it exists The entry name candidate (step S306: no) of measuring tank B2, then return step S302 and repeatedly later processing.In turn, when for time The entry name candidate for mending all free description type variable slot B2 for including in list F has been carried out from step S302 to step After processing until S305 (step S306: yes), the item value determined is input to the corresponding of formatted data A and is freely described Type variable slot B2 (step S307) terminates a series of processing shown in Fig. 9.
By free description type item value determination section 43 carry out more than processing, for example, as shown in figure 4, in formatted data A Variable slot No.4 free description type variable slot B2 in, as item value input " special item: be required film on the move Closedown school stumble and almost fall.".
In addition, the processing of free description type item value determination section 43 shown in Fig. 9 is an example, however it is not limited to this.Freely As long as the similar degree of entry name candidate of the description type item value determination section 43 based on free description type variable slot B2 (freely describes The segment word for including in the entry name candidate of type variable slot B2 arranges the similar degree relative to entry name) and the entry name candidate Position (offsets for the segment word column for including in the entry name candidate) is simultaneously input to according to the record of recognition result data E decision The structure of the item value of free description type variable slot B2.For example, in the above example, when there are competing conjunction candidate the case where If lower similar degree is bigger than the competing conjunction candidate, judge that the item value candidate for the free description type variable slot B2 for being set in VALUE is Segment word corresponding to VALUE is arranged the recognition result number of later part by the entry name of the free description type variable slot B2 The item value of the free description type variable slot B2 is determined as according to the record of E.But it is also possible to be formed as following structure: even if In the case where the similar degree of VALUE is bigger than the competing similar degree for closing candidate, in the pronunciation of recognition result data E, when in VALUE Segment word column after and position arrange close to the segment word exist the entry name for corresponding to others variable slot B or In the case where the segment word column of item value, it is judged as that competing conjunction candidate is the entry name of the free description type variable slot B2, it will be right The record for the recognition result data E that the segment word that Ying Yujing closes candidate arranges later part is determined as the free description type variable The item value of slot B2.
Template Manager portion 50 carries out the change in the format module C that template storage unit 20 stores according to the operation of template construct person Production, editor, deletion of measuring tank B etc., the format module C stored to template storage unit 20 are managed.Template Manager portion 50 is such as It include variable slot type configuration part 51, entry name configuration part 52 and set of options portion 53 shown in Fig. 5.
Each variable slot B that variable slot type configuration part 51 has for formatted data A as variable slot B type and set Any of option type and free description type.Entry name configuration part 52 sets the project for each variable slot B that formatted data A has Name and its pronunciation.Set of options portion 53 set the item value for the variable slot B1 of option type that formatted data A has option and its Pronunciation.
Figure 10 is the flow chart for showing an example of the processing sequence carried out by Template Manager portion 50.If processing shown in Fig. 10 Start, then Template Manager portion 50 determines the action type (operation of template construct person for the operation that template construct person is carried out first It is which of the production of variable slot B, editor, deletion) (step S401).In turn, if action type is the system of variable slot B Make, then entry name configuration part 52 (is walked according to the entry name and its pronunciation of the operation setting variable slot B to be made of template construct person Rapid S402), variable slot type configuration part 51 sets the type (step S403) of variable slot B.In turn, determine the variable to be made The type of slot B is free description type or option type (step S404), and the variable slot B to production is option type variable slot B1, Then set of options portion 53 (is walked according to the option and pronunciation of the item value of the operation setting option type variable slot B1 of template construct person Rapid S405).In addition, the setting for pronunciation, can separately be arranged the choosing of the entry name, item value according to template construct person input Item generates the pronunciation information generation unit of pronunciation, sets the output of the pronunciation information generation unit.Also, 1 project can also be directed to Name, the multiple pronunciations of set of options of item value.
Also, if the action type of template construct person is the editor of variable slot B, Template Manager portion 50 is selected to edit Object variable slot B (step S407), the entry name of selected variable slot B is carried out according to the operation of template construct person, is become The type of measuring tank B, item value option etc. editor (step S408).If also, the action type of template construct person is variable The deletion of slot B, then Template Manager portion 50 is selected to the variable slot B (step S409) for the object deleted, according to template construct person Operation by selected variable slot B delete (step S410).In addition, said to the example for deleting variable slot B herein It is bright, but in the case where deleting format module C, as long as being formed as the format for being selected to the object deleted in step S409 The structure of template C.
Then, Template Manager portion 50 determines whether the variable slot B (step S406) to be operated there is also template construct person.Into And the variable slot B (step S406: yes) to be operated if it exists, then return step S401 and repeat later processing, if not In the presence of the variable slot B (step S406: no) to be operated, then terminate a series of processing shown in Fig. 10.In addition, shown in Fig. 10 The processing in Template Manager portion 50 is an example, however it is not limited to this.As long as Template Manager portion 50 can be according to template construct person's Operation carries out production, editor, the deletion etc. of the variable slot B in the format module C that template storage unit 20 stores, and stores to template The structure that the format module C that portion 20 stores is managed.
The effect > of < embodiment
As enumerated specific example above and being described in detail by, the auxiliary input device 100 of present embodiment Each change towards formatted data A is being determined using recognition result data E obtained from carrying out voice recognition as the speech to user When the item value of measuring tank B input, format module C is used as hint information.In turn, pronunciation based on recognition result data E and The pronunciation of the entry name and option that describe in format module C determines the item value that each variable slot B towards formatted data A is inputted. Thus, auxiliary input device 100 according to the present embodiment, i.e., in convenient recognition result data E comprising mistake in the case where, It can suitably determine the item value inputted towards each variable slot B of formatted data A.Also, user can be according to not following rule Speech then but relatively flexibly speech and input suitable item value towards each variable slot of formatted data A.
Also, for the format module C used in the auxiliary input device 100 of present embodiment as hint information, Only by the option of the entry name of each variable slot B of formatted data A and the item value of option type variable slot B1 and respective pronunciation It sets, does not need relevant to voice recognition detailed as the rule production in the Audio Recognition Engine of grammer type together Thin parameter setting.Thus, even if template construct person does not have professional knowledge relevant to voice recognition and can easily make yet Make format module C, and its management is also easy.
Also, auxiliary input device 100 according to the present embodiment, even if there is option type variable slot in formatted data A In the case where the variable slot B of this 2 types of B1 He free description type variable slot B2,1 time for capable of also being carried out according to user is sent out It says and suitably determines above-mentioned option type variable slot B1 and the respective item value of free description type variable slot B2.
1 > of < variation
In addition, the embodiment being explained above is an example, it is able to carry out various deformations.For example, above-mentioned Embodiment in, the similar degree (portion that in item value candidate includes of the option type item value determination section 42 based on item value candidate Point similar degree of the character string relative to item value) and the item value candidate and others item value candidates or entry name candidate it Between positional relationship (difference of the offset for the segment word column for including in each candidate) at least any one determine towards option type The item value of variable slot B1 input.Also, in the above-described embodiment, free description type item value determination section 43 is based on freely Similar degree (the part for including in the entry name candidate of free description type variable slot B2 of the entry name candidate of description type variable slot B2 Similar degree of the character string relative to entry name) and the entry name candidate the position (segment word for including in the entry name candidate The offset of column) to determine the item value inputted towards free description type variable slot B2 according to the record of recognition result data E.But In the case where having had input item value in a part of variable slot B of formatted data A, option type item value determination section 42 or free The input status of formatted data A can be used also to determine to input towards option type variable slot B1 in description type item value determination section 43 Item value or towards free description type variable slot B2 input item value.
Figure 11 be show a part of variable slot B item value inputted the formatted data A finished an example figure, Tu12Shi The figure of an example of format module C corresponding with formatted data A shown in Figure 11 is shown.Also, Figure 13 is to show determining towards figure The figure of an example of the recognition result data E used when the item value of the variable slot B input of formatted data A shown in 11.
In the example shown in Figure 11 to Figure 13, by the processing of extraction unit 41, from the text of the pronunciation of recognition result data E Word column, as similar with item value (option) " 1 hour (i-chi-ji-ka-n) " of option type variable slot B1 of variable slot No.1 Segment word column extract " i-chi-ji-ka-n ", the item value (option) and segment word column combination by as project Value candidate is registered in candidate list F.Also, by the processing of extraction unit 41, from the character string of the pronunciation of recognition result data E, Item value (option) " side line (i-chi-ji-ga-wa-ha- as the option type variable slot B1 with variable slot No.2 I-ka-n) " similar segment word column extract " i-chi-ji-ka-n ", the group of the item value (option) and segment word column Conjunction is registered in candidate list F as item value candidate.
In this variation, when the multiple item value candidates for as described above including identical segment word column are registered in time Mend list F in the case where, option type item value determination section 42 not merely with the comparison of the similar degree of above-mentioned item value candidate or with The comparison of positional relationship between other candidates also utilizes the input status of formatted data A, determines the option of formatted data A The item value of type variable slot B1.For example, in the case where item value pays the utmost attention to the variable slot B not inputted, although with variable slot The corresponding item value candidate of item value (option) " 1 hour " of the option type variable slot B1 of No.1 is than the option with variable slot No.2 Item value (side line) corresponding item value candidate similar degree of type variable slot B1 is big, but as shown in figure 11, variable slot The item value of the option type variable slot B1 of No.1, which has inputted, to be finished, in contrast, the option type variable slot B1's of variable slot No.2 Item value not yet inputs.Thus, option type item value determination section 42 uses the item with the option type variable slot B1 of variable slot No.2 The corresponding item value candidate of mesh value (side line), is determined as " side line " for the item value of variable slot No.2.
Also, in the example shown in Figure 11 to Figure 13, by the processing of extraction unit 41, from the reading of recognition result data E In the character string of method, as the part similar with entry name " points for attention " of free description type variable slot B2 of variable slot No.3 Character string extracts " chu-u-i ", and the combination of the entry name and segment word column is registered in candidate column as entry name candidate Table F.
In this variation, when the entry name candidate of free description type variable slot B2 as described above is registered in candidate column In the case where table F, free description type item value determination section 43 not merely with the similar degree of entry name candidate or position, also utilize lattice The input status of formula data A determines the item value of the free description type variable slot B2 of formatted data A.For example, preferential in item value In the case where considering the variable slot B not inputted, even if including the free description type variable slot with variable slot No.3 in candidate list F The corresponding entry name candidate of B2 is finished since the item value of the free description type variable slot B2 of variable slot No.3 has inputted, Without using the decision of the item value of the entry name candidate.
Figure 14 is an example for showing the formatted data A for having input item value towards the variable slot B not inputted according to this modification Figure.With Figure 11 shown in compared with formatted data A, for the option type variable slot for the variable slot No.2 that item value not yet inputs B1 inputs " side line " determined by option type item value determination section 42 as item value.In addition, excellent to item value herein First consider that the example for the variable slot B not inputted is illustrated, but it is also assumed that based on the speech of rear input come to item value It has inputted the variable slot B finished and has been updated particularly significant, priority discipline value has inputted the variable slot B finished.
As above, in this variation, when had input item value in a part of variable slot B in formatted data A the case where Under, option type item value determination section 42 or free description type item value determination section 43 are also come using the input status of formatted data A Determine the item value of option type variable slot B1 or the item value of free description type variable slot B2.Thus, it according to this modification, can More suitably determine the item value that each variable slot B towards formatted data A is inputted.
More than in addition, imagination also uses formatted data A's other than the similar degree for using item value candidate or position The example of input status still, such as can also be formed as following structure: when the item value candidate of a certain variable slot B only has In the case that one is included in candidate list F, if the item value of the variable slot B in formatted data A not yet inputs, use The item value candidate determines the item value of the variable slot B not inputted, like this, and without using the similar degree of item value candidate or Position, but the input status based on the item value candidate and formatted data A for including in candidate list F are determined towards formatted data A Variable slot B input item value.
2 > of < variation
Also, in the format module C in addition to describe have the above-mentioned entry name of each variable slot B or the option of item value and its Except the information such as pronunciation, the type of variable slot B, also describe have supplemental information in the case where, option type item value determination section 42 or The supplemental information described in format module C also can be used in free description type item value determination section 43, determines to become towards option type The item value of measuring tank B1 input or the item value inputted towards free description type variable slot B2.
As the supplemental information described in format module C, such as can to enumerate the variable slot B high to priority additional Preferential input marking.The preferential input marking expression should be preferentially to the variable slot B cuit value.When for option type variable In the case that slot B1 attached preferential input marking, option type item value determination section 42 preferentially determines to be attached with the preferential input mark The item value of the option type variable slot B1 of note.Also, when the feelings for being attached with preferential input marking to free description type variable slot B2 Under condition, free description type item value determination section 43 preferentially determines the free description type variable slot B2 for being attached with the preferential input marking Item value.
Also, it as the other examples of the supplemental information described in format module C, can enumerate in formatted data A The information of each variable slot B to put in order.In this case, option type item value determination section 42 or free description type item value are determined Determine portion 43 and be directed to the item value candidate or entry name candidate for being registered in candidate list F by extraction unit 41, considers each format number According in A while putting in order, determining the item value of option type variable slot B1 or the item value of free description type variable slot B2.
As above, in this variation, in the case that describing in format module C has supplemental information, option type item value Determination section 42 or free description type item value determination section 43 also determine option type using the supplemental information described in format module C The item value of the item value of variable slot B1 or free description type variable slot B2.Thus, according to this modification, can more suitably it determine The item value that fixed each variable slot B towards formatted data A is inputted.
More than in addition, imagination also uses other than the similar degree for using item value candidate or position to be remembered in format module C The example for the supplemental information stated, but it is also possible to be formed as following structure: and without using the similar degree of item value candidate or position It sets, but based on the supplemental information described in the item value candidate and template C for including in candidate list F, it determines towards formatted data A Variable slot B input item value.
3 > of < variation
Also, in the above-described embodiment, imaginary formatted data A is to include option type variable slot B1 and free description type The form of variable slot B2 both sides has the choosing determined towards the option type variable slot B1 item value inputted to auxiliary input device 100 The free description type item value of item type item value determination section 42 and decision towards the free description type variable slot B2 item value inputted is determined The structure for determining 43 both sides of portion is illustrated.But in the formatted data A in the form of not including free description type variable slot B2 In the case where process object, auxiliary input device 100 can also be formed as not having free description type item value determination section 43 structure.
< supplementary explanation >
The auxiliary input device 100 of present embodiment for example can be by constituting the hardware of common computer and by calculating The program (software) that machine executes, which cooperates, to be realized.For example, scheduled program can be executed by computer to realize above-mentioned acquirement Portion 30, determination section 40 (extraction unit 41, option type item value determination section 42 and free description type item value determination section 43) and Template Manager portion 50 (variable slot type configuration part 51, entry name configuration part 52 and set of options portion 53).Further, it is possible to use The mass storage device that computer has realizes above-mentioned formatted data storage unit 10 and template storage unit 20.
Figure 15 is the block diagram for showing the hardware configuration example of auxiliary input device 100 of present embodiment.Auxiliary input device The storage devices such as 100 hardware processor, RAM102 or ROM103 such as shown in figure 15 there is conduct to have CPU101, The mass storage devices such as HDD104, via network with the communication I/F105 that is communicated of outside and for connecting peripheral equipment Equipment I/F106 common hardware structure of computer.
At this point, above-mentioned program is for example recorded in disk, CD, semiconductor memory or similar record Medium and provided.As long as the recording medium that the recording medium computer system of logging program can be read, is deposited Storage form can be any mode.Also, it is also configured to above procedure being installed on computer in advance, also be configured to Above-mentioned program via Web Publishing is appropriately mounted at computer.
Be formed as by the program that above-mentioned computer executes comprising above-mentioned acquisition unit 30, determination section 40 (extraction unit 41, choosing Type item value determination section 42 and free description type item value determination section 43) and (variable slot type is set in Template Manager portion 50 Determine portion 51, entry name configuration part 52 and set of options portion 53) etc. each function part modular structure, suitably by processor It reads the program and executes, above-mentioned each portion is generated on the main memories such as RAM102.
In addition, the auxiliary input device 100 of present embodiment be also possible to a part of above-mentioned each function part or Person all uses ASIC (Application Specific Integrated Circuit, specific integrated circuit) or FPTA Dedicated hard-wired structures such as (Field-Programmable Gate Array, field programmable gate arrays).
Also, the auxiliary input device 100 of present embodiment is also configured to multiple stage computers to communicate The network system that mode connects, and above-mentioned each section is dispersed in multiple stage computers and the structure realized.Also, present embodiment Auxiliary input device 100 be also possible to the virtual machine acted on cloud system.
Embodiments of the present invention are illustrated above, but above embodiment is only mentioned as an example Show, is not intended to limit the range of invention.The new embodiment can be implemented in such a way that others are various, Neng Gou The range for not departing from the purport of invention carries out various omissions, displacement, change.Above embodiment and its deformation are also contained in invention Range or purport in, and be contained in recorded in technical solution invention and its equivalent range in.

Claims (11)

1. a kind of auxiliary input device, has:
Template storage unit, storage format template, the format module are with 1 or more with entry name accordingly cuit value Variable slot formatted data template, the entry name of each variable slot and from multiple options select item value option type change The option of measuring tank is described together with respective pronunciation;
Acquisition unit, obtains recognition result data, which obtained and the speech to user carries out voice recognition The recognition result data arrived include record and pronunciation;And
Determination section, pronunciation based on above-mentioned recognition result data and the entry name described in above-mentioned format module and option Pronunciation determines the item value that the variable slot towards above-mentioned formatted data inputs.
2. auxiliary input device according to claim 1, wherein
Above-mentioned determination section includes:
Extraction unit, the entry name for extracting and being described in above-mentioned format module from the character string of the pronunciation of above-mentioned recognition result data Or option pronunciation it is similar segment word column;And
Option type item value determination section is arranged based on the segment word similar with the pronunciation of option described in above-mentioned format module That is the similar degree of part 1 character string and the part 1 character string is the same as the reading with the entry name described in above-mentioned format module The similar segment word column of method are part 2 character string or other part 1 texts different from the part 1 character string In positional relationship between column at least any one, determine towards option type variable slot input item value.
3. auxiliary input device according to claim 1, wherein
Above-mentioned determination section includes:
Extraction unit, the entry name for extracting and being described in above-mentioned format module from the character string of the pronunciation of above-mentioned recognition result data Or option pronunciation it is similar segment word column;And
Option type item value determination section, based on the segment word similar with the pronunciation of option described in above-mentioned format module A part possessed by column and above-mentioned formatted data in multiple variable slots be entered item value in the case where above-mentioned format The input status of data determines the item value that the variable slot towards option type inputs.
4. auxiliary input device according to claim 1, wherein
Above-mentioned determination section includes:
Extraction unit, the entry name for extracting and being described in above-mentioned format module from the character string of the pronunciation of above-mentioned recognition result data Or option pronunciation it is similar segment word column;And
Option type item value determination section, based on the segment word similar with the pronunciation of option described in above-mentioned format module Column and the supplemental information described in above-mentioned template determine the item value that the variable slot towards option type inputs.
5. the auxiliary input device according to any one of claim 2~4, wherein
Above-mentioned determination section also includes free description type item value determination section, which is based on and project The pronunciation of the entry name of the variable slot for the free description type that value can be freely written it is similar segment word column similar degree and The position of segment word column determines the item that the variable slot towards free description type inputs according to the record of above-mentioned recognition result data Mesh value.
6. the auxiliary input device according to any one of claim 2~4, wherein
Above-mentioned determination section also includes free description type item value determination section, which is based on and project The pronunciation of the entry name of the variable slot for the free description type that value can be freely written similar segment word column and above-mentioned format A part in multiple variable slots possessed by data be entered item value in the case where above-mentioned formatted data input shape Condition determines the item value that the variable slot towards free description type inputs.
7. the auxiliary input device according to any one of claim 2~4, wherein
Above-mentioned determination section also includes free description type item value determination section, which is based on and project The similar segment word of the pronunciation of the entry name of the variable slot for the free description type that value can be freely written arranges and in above-mentioned mould The supplemental information described in plate determines the item value that the variable slot towards free description type inputs.
8. auxiliary input device according to claim 5, wherein
It is also equipped with the Template Manager portion for managing above-mentioned format module,
Above-mentioned Template Manager portion includes:
Variable slot type configuration part, for each variable slot possessed by above-mentioned formatted data, as variable slot type and set Any of option type and free description type;
Entry name configuration part sets the entry name and its pronunciation of each variable slot possessed by above-mentioned formatted data;And
Set of options portion sets the option and its pronunciation of the item value of the variable slot of option type possessed by above-mentioned formatted data.
9. auxiliary input device according to claim 6, wherein
It is also equipped with the Template Manager portion for managing above-mentioned format module,
Above-mentioned Template Manager portion includes:
Variable slot type configuration part, for each variable slot possessed by above-mentioned formatted data, as variable slot type and set Any of option type and free description type;
Entry name configuration part sets the entry name and its pronunciation of each variable slot possessed by above-mentioned formatted data;And
Set of options portion sets the option and its pronunciation of the item value of the variable slot of option type possessed by above-mentioned formatted data.
10. auxiliary input device according to claim 7, wherein
It is also equipped with the Template Manager portion for managing above-mentioned format module,
Above-mentioned Template Manager portion includes:
Variable slot type configuration part, for each variable slot possessed by above-mentioned formatted data, as variable slot type and set Any of option type and free description type;
Entry name configuration part sets the entry name and its pronunciation of each variable slot possessed by above-mentioned formatted data;And
Set of options portion sets the option and its pronunciation of the item value of the variable slot of option type possessed by above-mentioned formatted data.
11. a kind of recording medium, wherein
Record has the program for implementing function such as computer, which has the template storage of storage format template Portion, the format module be with the template with the entry name accordingly formatted data of 1 or more variable slot of cuit value, The entry name of each variable slot and the option of the variable slot of the option type of selection item value and respective pronunciation one from multiple options And described,
Above-mentioned function includes:
The function of recognition result data is obtained, which obtained and the speech to user carries out voice recognition Recognition result data, comprising record and pronunciation;And
The pronunciation of pronunciation and the entry name and option that are described in above-mentioned format module based on above-mentioned recognition result data, certainly The function for the item value that the fixed variable slot towards above-mentioned formatted data inputs.
CN201711211098.5A 2017-11-28 2017-11-28 Input support device and recording medium Active CN109840062B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201711211098.5A CN109840062B (en) 2017-11-28 2017-11-28 Input support device and recording medium

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201711211098.5A CN109840062B (en) 2017-11-28 2017-11-28 Input support device and recording medium

Publications (2)

Publication Number Publication Date
CN109840062A true CN109840062A (en) 2019-06-04
CN109840062B CN109840062B (en) 2022-10-28

Family

ID=66879363

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201711211098.5A Active CN109840062B (en) 2017-11-28 2017-11-28 Input support device and recording medium

Country Status (1)

Country Link
CN (1) CN109840062B (en)

Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH04243477A (en) * 1991-01-17 1992-08-31 Dainippon Printing Co Ltd Index word extraction method for natural language processing system
JPH11203325A (en) * 1998-01-16 1999-07-30 Tadamitsu Ryu Database generating method, record medium in which its program is stored and record medium recording group of data generated by the generating method
CN1470015A (en) * 2000-10-11 2004-01-21 ������������ʽ���� Information processing device, information processing method, and storage medium
CN1681565A (en) * 2002-09-13 2005-10-12 科乐美股份有限公司 Game device, game device control method, program, program distribution device, information storage medium
US6999932B1 (en) * 2000-10-10 2006-02-14 Intel Corporation Language independent voice-based search system
EP1631955A2 (en) * 2003-05-29 2006-03-08 Dictaphone Corporation Systems and methods utilizing natural language medical records
WO2011093025A1 (en) * 2010-01-29 2011-08-04 日本電気株式会社 Input support system, method, and program
CN102165437A (en) * 2008-07-25 2011-08-24 夏普株式会社 Information processing device and information processing method
CN102365639A (en) * 2009-04-06 2012-02-29 三菱电机株式会社 Retrieval device
CN104464731A (en) * 2013-09-20 2015-03-25 株式会社东芝 Data collection device, method, voice talking device and method
CN104488027A (en) * 2012-07-09 2015-04-01 独立行政法人情报通信研究机构 Speech processing system and terminal device
US20160260430A1 (en) * 2015-03-06 2016-09-08 Dell Products L.P. Voice-based input using natural language processing for interfacing with one or more devices
CN105981099A (en) * 2014-02-06 2016-09-28 三菱电机株式会社 Speech search device and speech search method
US20160343374A1 (en) * 2015-05-22 2016-11-24 Kabushiki Kaisha Toshiba Minutes taking system, minutes taking method, and image forming apparatus
WO2017010506A1 (en) * 2015-07-13 2017-01-19 帝人株式会社 Information processing apparatus, information processing method, and computer program
US20170323641A1 (en) * 2014-12-12 2017-11-09 Clarion Co., Ltd. Voice input assistance device, voice input assistance system, and voice input method

Patent Citations (16)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
JPH04243477A (en) * 1991-01-17 1992-08-31 Dainippon Printing Co Ltd Index word extraction method for natural language processing system
JPH11203325A (en) * 1998-01-16 1999-07-30 Tadamitsu Ryu Database generating method, record medium in which its program is stored and record medium recording group of data generated by the generating method
US6999932B1 (en) * 2000-10-10 2006-02-14 Intel Corporation Language independent voice-based search system
CN1470015A (en) * 2000-10-11 2004-01-21 ������������ʽ���� Information processing device, information processing method, and storage medium
CN1681565A (en) * 2002-09-13 2005-10-12 科乐美股份有限公司 Game device, game device control method, program, program distribution device, information storage medium
EP1631955A2 (en) * 2003-05-29 2006-03-08 Dictaphone Corporation Systems and methods utilizing natural language medical records
CN102165437A (en) * 2008-07-25 2011-08-24 夏普株式会社 Information processing device and information processing method
CN102365639A (en) * 2009-04-06 2012-02-29 三菱电机株式会社 Retrieval device
WO2011093025A1 (en) * 2010-01-29 2011-08-04 日本電気株式会社 Input support system, method, and program
CN104488027A (en) * 2012-07-09 2015-04-01 独立行政法人情报通信研究机构 Speech processing system and terminal device
CN104464731A (en) * 2013-09-20 2015-03-25 株式会社东芝 Data collection device, method, voice talking device and method
CN105981099A (en) * 2014-02-06 2016-09-28 三菱电机株式会社 Speech search device and speech search method
US20170323641A1 (en) * 2014-12-12 2017-11-09 Clarion Co., Ltd. Voice input assistance device, voice input assistance system, and voice input method
US20160260430A1 (en) * 2015-03-06 2016-09-08 Dell Products L.P. Voice-based input using natural language processing for interfacing with one or more devices
US20160343374A1 (en) * 2015-05-22 2016-11-24 Kabushiki Kaisha Toshiba Minutes taking system, minutes taking method, and image forming apparatus
WO2017010506A1 (en) * 2015-07-13 2017-01-19 帝人株式会社 Information processing apparatus, information processing method, and computer program

Also Published As

Publication number Publication date
CN109840062B (en) 2022-10-28

Similar Documents

Publication Publication Date Title
JP6762819B2 (en) Input support device and program
CN105264518B (en) Data processing equipment and story model building method
CN110619115B (en) Template creating method and device, electronic equipment and storage medium
CN105335360B (en) The method and apparatus for generating file structure
CN103632668B (en) A kind of method and apparatus for training English speech model based on Chinese voice information
JP2007058605A5 (en)
CN101882135B (en) Data processing method and device
JP2005174336A (en) Learning and use of generalized string pattern for information extraction
JP6645850B2 (en) Information management apparatus, information management method, and information management program
CN104077346A (en) Document creation support apparatus, method and program
CN107203574A (en) Data management and the polymerization of data analysis
JP4589910B2 (en) Conversation recording blogging device
CN108304377A (en) A kind of extracting method and relevant apparatus of long-tail word
CN102193920A (en) Name word stock generating method and device as well as text input system
CN103246642B (en) Information processor and information processing method
CN111815162A (en) Digital auditing tool and method
CN107169011A (en) The original recognition methods of webpage based on artificial intelligence, device and storage medium
JP2004233577A (en) Method for forming speech recognition dictionary
CN109840062A (en) Auxiliary input device and recording medium
CN104464731A (en) Data collection device, method, voice talking device and method
CN106528516A (en) Method for establishing analysis report
CN110956043A (en) Domain professional vocabulary word embedding vector training method, system and medium based on alias standardization
CN110956170A (en) Method, device, equipment and storage medium for generating passport machine-readable code sample
CN108614821B (en) Geological data interconnection and mutual-checking system
CN108052686A (en) A kind of abstract extraction method and relevant device

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
GR01 Patent grant
GR01 Patent grant