JP3762300B2 - Text input processing apparatus and method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide a device, method, and program for text input equipped with a user interface enabling a user to properly and easily edit an input text based upon a speech recognition result. <P>SOLUTION: The text input processor providing a dictation function of inputting a text by using speech recognition correlates an index with a pronounced character string and puts them side by side. Candidates having the same pronunciation are displayed in a candidate list where the index is specified is displayed. In a candidate list where the pronounced character string is specified, candidates having different pronunciations are displayed. <P>COPYRIGHT: (C)2003,JPO

Description

[0001]
BACKGROUND OF THE INVENTION
The present invention relates generally to natural language processing, and more particularly to a text input processing apparatus, method, and program using speech recognition technology that provides a dictation function for inputting text (sentence) by speech.
[0002]
[Prior art]
In the conventional example of the text input processing device using the speech recognition technology, the speech recognition result is displayed in a simple kana-kana mixed text format. When correcting the input text based on such speech recognition, if the correction target is selected and the candidate display operation is performed, in the conventional example, candidates with different notations and candidates with different pronunciations are all displayed. It was. In such a conventional example, when the input text displayed based on the speech recognition result is different from the one expected by the user, whether it is a difference in notation in the homophone or a difference in sound, Or, there is a problem that it is difficult to determine whether the unit of division is a word or the like. In particular, words that the user does not know or cannot read are displayed, which makes it difficult to correct the text. This is a problem specific to text input based on speech recognition that is different from text input using a keyboard. is there.
[0003]
In addition, when selecting correction candidates, recognition candidates due to multiple factors such as candidates with different notation and pronunciation are displayed, so it takes time to find the target candidate and the operation is complicated. There is also. Further, even when the display is made only after a new operation step after the transition to the candidate selection state, the operation until the user can obtain the target candidate becomes complicated.
[0004]
As described above, it is desired that the conventional example of the text input processing device using the speech recognition technology is provided with a user interface that can easily perform an operation of correcting (editing in a broad sense) the input speech text. .
[0005]
[Problems to be solved by the invention]
The present invention has been made in view of such circumstances, and provides a text input device, method, and program provided with a user interface that allows a user to appropriately and easily edit input text based on a speech recognition result. With the goal.
[0006]
[Means for Solving the Problems]
In order to solve the above problems and achieve the object, the present invention is configured as follows.
[0007]
A first text input processing device according to the present invention is a text input processing device that performs text input processing using speech recognition, and any one of a plurality of clauses constituting the text includes a plurality of abnormal sounds. Storage means for storing a speech recognition result having pronunciation character string candidates including words, display means for combining and displaying each of the pronunciation character strings of the plurality of phrases, and for any one of the phrases, the pronunciation characters Text input comprising: candidate display means for displaying a list of column candidates; and selection means for causing a user to select any one of the abnormal words from the displayed candidates. It is a processing device.
[0008]
A second text input processing device according to the present invention is a text input processing device that performs text input processing using speech recognition, and any one of a plurality of clauses constituting the text includes a plurality of texts. Storage means for storing speech recognition results having pronunciation character string candidates including allophone words and kanji kana mixed character strings including a plurality of homophones having different notations, and pronunciation characters of each of the plurality of phrases A first display means for displaying a combination of one of the columns, a second display means for displaying a combination of kanji characters in each of the plurality of phrases, and any one of the phrases, First candidate display means for displaying a list of pronunciation character string candidates; second candidate display means for displaying a list of candidate character strings mixed with the kanji characters for any one of the phrases; The first selection means for allowing the user to select any one of the phonetic character strings from the displayed pronunciation character string candidates and the kanji kana mixed character string candidates displayed as a list And a second selection means for allowing the user to select one homophone.
[0009]
A first text input processing method according to the present invention is a text input processing method for inputting text using speech recognition, wherein any one of a plurality of clauses constituting the text includes a plurality of abnormal sounds. A storage step of storing a speech recognition result having pronunciation character string candidates including a word; a display step of displaying a combination of the pronunciation character strings of the plurality of phrases; and the pronunciation character for any one of the phrases A text input process comprising: a candidate display step for displaying a list of column candidates; a selection step for allowing a user to select any one of the abnormal words from the candidates displayed in the list; Is the method.
[0010]
A second text input processing method according to the present invention is a text input processing method for inputting text using speech recognition, wherein any one of a plurality of clauses constituting the text includes a plurality of texts. Storing a speech recognition result having pronunciation character string candidates including allophone words and kanji kana mixed character string candidates including a plurality of homophones having different notations; and each pronunciation character of each of the plurality of phrases A first display step that displays a combination of one of the columns, a second display step that displays a combination of one of the kanji-kana mixed character strings of the plurality of phrases, and any one of the phrases, A first candidate display step for displaying a list of pronunciation character string candidates; and a second candidate for displaying a list of candidate character strings mixed with the kanji characters for any one of the phrases A first selection step for allowing the user to select any one of the abnormal words from the displayed pronunciation character string candidates displayed in a list; and the candidate kanji mixed character strings displayed as a list And a second selection step for allowing the user to select any one of the homophones. A text input processing method comprising:
[0011]
A first program according to the present invention is a program for processing input of text using speech recognition, and any one of a plurality of clauses constituting the text includes a plurality of allophones. Storage means for storing a speech recognition result having pronunciation character string candidates, display means for displaying a combination of the pronunciation character strings of the plurality of phrases, and a list of the pronunciation character string candidates for any one of the phrases This is a program for functioning as a candidate display means for displaying and a selection means for allowing the user to select any one of the abnormal words from among the candidates displayed as a list.
[0012]
A second program according to the present invention is a program for processing input of text using speech recognition, wherein any one of a plurality of clauses constituting the text is a plurality of allophones. A storage means for storing a speech recognition result having a candidate for a pronunciation character string including a character and a kanji kana mixed character string candidate including a plurality of homophones having different notations, and one of the pronunciation character strings of each of the plurality of phrases First display means for displaying in combination, second display means for combining and displaying one of the kanji-kana mixed character strings of each of the plurality of phrases, and for any one of the phrases, the candidate pronunciation string First candidate display means for displaying a list, Second candidate display means for displaying a list of candidates for the character string mixed with kanji for any one of the phrases, First selection means for allowing the user to select any one of the phonetic words from the pronunciation character string candidates, and any one homophone from the kanji kana mixed character string candidates displayed in a list. Is a program for causing the user to select the second selection means.
[0013]
DETAILED DESCRIPTION OF THE INVENTION
Hereinafter, embodiments of the present invention will be described with reference to the drawings.
[0014]
FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a text input device according to the present invention. The text input device according to the present embodiment can be configured based on, for example, a general-purpose computer provided with a device for speech recognition. The text input device is coupled to a speech input device such as a microphone and inputs speech uttered by a user. A voice input unit 11 that recognizes a voice input to the voice input unit 11, a candidate information holding unit 13 that holds a voice recognition result by the voice recognition unit 12, and an input device such as a keyboard or a mouse The candidate selection operation unit 14 for inputting information on the candidate selection operation performed by the user, and the main part of the present embodiment that controls the selective candidate display, and is input from the candidate selection operation unit 14 A candidate display control unit 15 that creates appropriate candidate display information from the information of the recognition result held by the candidate information holding unit 13 according to the operation information to be displayed, and a candidate table A display unit 16 for displaying the candidate display information created by the control unit 15 on the display, and a.
[0015]
The main components corresponding to the present invention are the candidate information holding unit 13, the candidate display control unit 15, and the candidate selection operation unit 14, and these components can be realized by, for example, computer software.
[0016]
FIG. 2 is a block diagram showing a schematic configuration of the speech recognition unit 12 in FIG. As shown in FIG. 2, the voice recognition unit 12 performs signal processing and analysis on the voice input unit 21 that inputs voice data, and the voice data input via the voice input unit 21, thereby An acoustic processing unit 22 that performs acoustic processing such as detection and clipping, or extraction and parameterization of feature amounts, and voice information parameterized by the acoustic processing unit 22 are registered in the word dictionary 26. A collation unit 23 that collates with a word string composed of words, is referred to in the collation process in the collation unit 23, and is also referred to in the collation process in the collation unit 23, similarly to the acoustic dictionary 24 that is composed of HMM (Hidden Markov Model) and the like And analyzing the word dictionary 25 composed of a statistical language model (n-gram) and the like and the word string obtained as a result of the matching process in the matching unit 23, A language processing unit 27 for reconfiguring in units of clauses and expanding homophones, a word dictionary 28 that stores various information about words that are referred to in the language processing in the language processing unit 27, and a language It comprises a candidate output unit 29 that manages the history of the language processing results in the processing unit 27 and performs candidate output.
[0017]
In the present embodiment configured as described above, it is assumed that the voice uttered by the user is input to the voice input unit 11 and the voice recognition unit 12 recognizes the one utterance immediately after the end of the utterance. Here, a case where candidate information as shown in FIG. 3 or FIG. 4 is output by the voice recognition unit 12 and stored in the candidate information holding unit 13 will be described as an example.
[0018]
When the candidate information of the recognition result is received from the candidate information holding unit 13, the candidate display control unit 15 immediately uses the combination of reading (pronunciation) and heading determined to be most appropriate, and displays the display unit 16 in FIG. 5. Candidate display with “reading (pronunciation)” is displayed. Such “reading” is the same information as so-called “ruby (reading pseudonym)”.
[0019]
Here, the candidate selection operation unit 14 is displayed on the display unit 16 such that the user selects and designates a part to be selected by the user with a keyboard, a mouse, or the like, or performs an operation for executing candidate display. An operation interface for selecting and specifying another candidate from a plurality of candidates is provided. Details thereof will be described later.
[0020]
Next, candidate information output by the speech recognition unit 12 will be described.
[0021]
FIG. 3 shows an example of candidate information output by the speech recognition unit 12. In the candidate information, a plurality of candidates having the same phrase boundary are stored in addition to the most appropriate (first rank) series of speech recognition results obtained by the speech recognition unit 12. Each candidate has information of reading (pronunciation character string) and heading (notation mixed with kanji characters). The candidate information also stores candidates of the same notation (pronunciation) (synonyms) and different candidates (pronunciations) (sound words, different words).
[0022]
As shown in FIG. 3, as phrase number 1, two pronunciation candidates “from the heart” recognized in the section from the position 0 to 40 in the speech are stored, including homophones. The notation determined to be the most appropriate among these is “from the heart”.
[0023]
In addition, as phrase number 2, a total of ten candidate pronunciations including “sound” and “atsuko” recognized in the section from position 40 to 60 following phrase number 1 are stored, including homophones. Among these, “hot” is determined to be the most appropriate. Examples of homophones include “thick”.
[0024]
Further, as phrase number 3, four pronunciation candidates “Orei”, “Orei”, “Ode”, and “Orehe” recognized in the section from position 60 to 88 following phrase number 2 include homophones. A total of 11 are stored. Of these, “thank you” is determined to be the most appropriate. Among the candidate pronunciations of allophones in the same section, “onrei”, “onde”, and “to me” are determined to be the most appropriate.
[0025]
FIG. 4 shows another example of candidate information output by the speech recognition unit 12. In this candidate information, clause numbers 1 to 3 are the same as those in FIG. Then, “here is” recognized in the section from the position 0 to 36 in the voice is stored as the phrase number 4, and “desu” is recognized as the phrase number 7 in the section from the position 36 to 52 following this. "Is stored, and" Konrei "recognized in the subsequent section from position 52 to 88 is stored as the phrase number 9, including the same word. Among these candidates, “9th”, “put out”, and “wedding” are determined to be most appropriate.
[0026]
Further, in the candidate information of FIG. 4, “here also” is recognized as the phrase number 5 in the section from the position 0 to 32 in the speech, and is recognized as the phrase number 6 in the section from the position 32 to 46 following this. “Sutton” recognized in the subsequent section from position 46 to 88 is stored as the phrase number “8”, including the same word. Among these candidates, “also here”, “just”, and “jump” are determined to be the most appropriate.
[0027]
That is, “heart”, “hot”, “thank you” with clause numbers 1 to 3 are the first series, and “9 days (here)” of clause numbers 4 and 7 and 9 are the other series with different clause boundaries. “Out”, “wedding”, and “here also”, “just” and “sweep” are stored.
[0028]
Here, operations of the candidate selection operation unit 14, the candidate display control unit 15, and the display unit 16 will be described.
[0029]
In the initial state, as shown in FIG. 5, it is assumed that “from the heart / from the heart”, “hot / hot”, “Ore / Thank you”, “I will give you / I will tell you” is displayed on the display unit 16.
[0030]
First, a case where a candidate gives a candidate display instruction for “headline”, for example, a case where a candidate display instruction is given by selecting “hot” will be described. When such an operation is performed, the operation information is given to the candidate display control unit 15 through the candidate selection operation unit 14. The candidate display control unit 15 displays the candidate that is a homonym of “hot” among the candidates corresponding to “hot” in the candidate information held in the candidate information holding unit 13 on the display unit 16 as shown in FIG. Display.
[0031]
Here, in the present embodiment, when a predetermined mode switching operation is given to the candidate display control unit 15, the candidate display control unit 15 reads the reading (as shown in FIG. 7 instead of the display information shown in FIG. Candidate display including even candidates (pronunciation, different words) with different pronunciations is configured. The candidate displays in FIGS. 6 and 7 are preferably configured to be switchable with each other in accordance with the mode switching operation.
[0032]
Furthermore, here, when an operation of selecting and specifying “thick” among the display candidates of FIG. 6 (or FIG. 7) is performed, the display unit 16 displays the selected “thick” as shown in FIG. In addition, when an operation of selecting and specifying “thickness” among the candidates displayed in FIG. 7 is performed, the display unit 16 displays the selected “thickness” and its reading (pronunciation) “atsukomo” in FIG. To display.
[0033]
Next, a case where a candidate display instruction for “reading” is given by the user, for example, a case where “Ore” is selected and a candidate display instruction is given will be described. When such an operation is performed, among the candidates corresponding to “thank you” in the candidate information, a reading (pronunciation) different from “Ore” is displayed as shown in FIG. At this time, as shown in FIG. 11, it is preferable to be able to switch to a mode in which notation (pronunciation) is displayed together with notation. The notation may be a notation determined to be most appropriate among candidates corresponding to the reading (pronunciation).
[0034]
Here, when an operation of selecting and specifying “onrei” among the display candidates in FIG. 10 is performed, the display unit 16 displays the selected “onrei” and the corresponding notation “thank you” as shown in FIG. To display. Note that the same applies to the case where an operation of selecting “Onrei / Thank you” is performed from the candidates displayed in FIG.
[0035]
According to the present embodiment as described above, the user can easily obtain an appropriate combination of reading (pronunciation) and notation based on the selective candidate display for “heading” and “reading”, and can obtain a desired text. Can be input (corrected, etc.).
[0036]
Here, another candidate display process based on the above-described configuration will be described. Another candidate display process is to display only “reading” without displaying the headline.
[0037]
The candidate display control unit 15 uses only the reading (pronunciation) determined to be the most appropriate, and causes the display unit 16 to perform candidate display of only the reading (pronunciation) as shown in FIG.
[0038]
In this case, when an instruction to select and display “Ore” reading (pronunciation) is displayed from the user via the candidate selection operation unit 14, the candidate corresponding to “Thank you” in the candidate information. Among them, the reading (pronunciation) different from “Ore” is displayed as shown in FIG. At this time, as shown in FIG. 15, it is preferable to enable switching to a mode in which the most appropriate notation is displayed together with reading (pronunciation).
[0039]
Furthermore, when the user performs an operation of selecting and specifying “onrei” from the candidates displayed in FIG. 14, the display unit 16 displays the selected “onrei” as shown in FIG. 16. . The same applies to the case where an operation of selecting “onre / thank you” from among the candidates displayed in FIG. 15 is performed.
[0040]
Next, the processing content in the candidate display control part 15 is demonstrated with reference to the flowchart of FIG.
[0041]
When candidate information is input from the speech recognition unit 12, the candidate display control unit 15 stores the candidate information in the candidate information holding unit 13 (step S31).
[0042]
Next, the candidate information of the first series determined to be most appropriate among the candidate information is acquired from the candidate information holding unit 13 (step S32), and setting information for determining whether or not to display the headline is determined (step S32). S33). You may comprise so that a user can set this setting information.
[0043]
In the case of setting to display a headline, display information using a combination of reading (pronunciation) and a headline is created (step S34). On the other hand, in the case of setting not to display the headline, display information using only reading (pronunciation) is created (step S35), and display on the display unit 16 is performed (step S36). The display in step S36 is display of candidate information of the first rank series.
[0044]
Thereafter, the candidate display control unit 15 shifts to a standby state for accepting a candidate display instruction from the user (step S37).
[0045]
Here, when an instruction for candidate display is given from the user, display information of candidate selection candidates is created (step S38) and displayed by the display unit 16. The processing content of step S38 will be described later. Subsequent to this candidate display operation, a transition is made to a standby state for accepting a candidate selection operation from the user (step S39).
[0046]
Here, when an operation for selecting a candidate is performed, the display of the display unit 16 is updated using a set of reading (pronunciation) and heading of the designated candidate (step S40), and the candidate display operation from the user is performed again. (Step S37).
[0047]
Next, details of candidate display processing for candidate selection (step S38) in the candidate display control unit 15 will be described with reference to the flowchart of FIG.
[0048]
First, when an instruction operation from the user to display a candidate (for example, mouse click) is detected, it is determined whether the designated location is a headline or a reading (pronunciation) (step S41). When the heading is designated, a predetermined setting content as to whether to display all candidates is referred to (step S42). In the case of setting not to display all candidates, for example, only homophone candidates are extracted from the candidate information holding unit 13 (step S43). In the case of setting to display all candidates, all candidates in the same section are extracted (step S44). Candidates extracted according to these settings are displayed on the display unit 16 in step S45.
[0049]
On the other hand, if it is determined in step S41 that the designated location is reading (pronunciation), the candidate information holding unit 13 is a candidate for different reading (pronunciation), and is most appropriate for each reading (pronunciation). In step S46, the notation candidates determined to be extracted are extracted. Furthermore, a predetermined setting content as to whether or not the headline display is used together is referred to (step S47). In accordance with such setting contents, a processing operation is selected as to whether only reading (pronunciation) is displayed as a candidate (step S48) or whether reading (pronunciation) and notation (heading) are combined and displayed as a candidate (step S49). . After such an operation, candidates are displayed on the display unit 16 in step S45.
[0050]
Here, the modification of the said embodiment is demonstrated.
[0051]
In the above-described embodiment, Hiragana “Yomi” is used as a phonetic character string displayed in combination with a heading, but Katakana or Romaji may be used. Also, a notation similar to actual pronunciation such as “Ore” may be used for “thank you” instead of “Ore”. Further, a display indicating an accent type may be added.
[0052]
Moreover, in the said embodiment, although Japanese is made into object, another language may be sufficient. For example, for Chinese, Pinyin or phonograms may be used as pronunciation character strings. A tone display may also be added.
[0053]
Moreover, in the said embodiment, although candidate display is performed per phrase, you may carry out per word and other units.
[0054]
The candidate selection operation method may be performed using a keyboard, a mouse, a pen, a voice operation, or the like. For any device that can specify a selection target and instruct execution, the present invention Is applicable.
[0055]
In the above-described embodiment, candidates are displayed by a so-called pop-up window, but a method of enumerating and displaying in another area such as the lower end of the screen may be used.
[0056]
In the above embodiment, different notation candidates for the same reading (pronunciation) are generated in advance in the speech recognition processing. However, in the speech recognition processing, different reading (pronunciation) are handled and are different. You may comprise so that the language processing expanded to the candidate of a description may be performed separately. For example, candidate expansion of different notations may be performed when a candidate display operation is performed.
[0057]
In the above embodiment, the speech recognition unit 12 includes the language processing unit 27, but instead of the processing unit 27, a language processing unit such as kana-kanji conversion that mainly processes keyboard input is used. It may be configured to be externally attached to the voice recognition unit 12.
[0058]
The present invention is not limited to the above-described embodiments and modifications, and can be implemented with various modifications. The present invention is effective as a method for configuring means for inputting text in various information processing apparatuses, and can be widely used in software for personal computers, word processor devices, portable information devices, and the like.
[0059]
【The invention's effect】
As described above, according to the present invention, it is possible to provide a text input device, method, and program provided with a user interface that allows a user to appropriately and easily edit input text based on a speech recognition result.
[Brief description of the drawings]
FIG. 1 is a block diagram showing a schematic configuration of an embodiment of a text input device according to the present invention.
2 is a block diagram showing a schematic configuration of a speech recognition unit 12 shown in FIG.
FIG. 3 is a diagram illustrating an example of candidate information output by the speech recognition unit 12
FIG. 4 is a diagram showing another example of candidate information output by the speech recognition unit 12
FIG. 5 is a diagram showing speech input text in an initial state
FIG. 6 is a diagram for explaining a case where a candidate display instruction is given for “Heading”;
FIG. 7 is a diagram showing a case where candidate display including candidates with different readings (pronunciations) (abnormal words and different words) in addition to the display contents of FIG. 6 is performed.
FIG. 8 is a diagram showing a selection operation for the display candidate shown in FIG.
9 is a diagram showing a selection operation for the display candidate shown in FIG.
FIG. 10 is a diagram for explaining a case where a candidate display instruction for “reading” is performed;
FIG. 11 is a diagram for explaining a case where candidates are displayed together with notation in addition to reading (pronunciation)
12 is a diagram showing a selection operation for the display candidate shown in FIG.
FIG. 13 is a diagram for explaining an embodiment in which only “reading” is displayed without displaying a heading;
14 is a diagram for explaining a case where an instruction to display a candidate for a certain “reading” is given with respect to the display of FIG. 13;
FIG. 15 is a diagram for explaining a case where a candidate is displayed together with the most appropriate notation in addition to reading (pronunciation)
FIG. 16 is a diagram showing a selection operation for the display candidate shown in FIG.
FIG. 17 is a flowchart illustrating an example of processing contents in the candidate display control unit 15;
18 is a flowchart showing the contents of candidate selection display processing (step S38) in the flowchart of FIG. 17;
[Explanation of symbols]
11 ... Voice input part
12 ... Voice recognition unit
13 ... Candidate information holding unit
14 ... Candidate selection operation section
15 ... Candidate display control unit
16 ... Display section
21 ... Voice input part
22 ... Sound processing section
23 ... Verification unit
24 ... Acoustic dictionary (HMM)
25 ... Language dictionary (n-gram)
26 ... Word dictionary
27 ... Language processor
28 ... Word dictionary
29 ... Candidate output section

Claims (5)

In a text input processing device that inputs and processes text using speech recognition,
A speech recognition result comprising a plurality of phrase information, wherein the phrase information of the same phrase includes a plurality of pronunciation character strings including a plurality of allophone words, and a plurality of homophones having different notations, and Storage means for storing a speech recognition result having a candidate for a heading character string corresponding to the candidate;
First display control means for performing control to display any one of a plurality of pronunciation character string candidates in the same phrase side by side for each phrase;
A first phrase that performs control such that when any one of the pronunciation character strings for each phrase displayed by the control by the first display control means is designated using the input means, any one phrase is selected. A selection means;
First candidate display control means for performing a control to display a list of a plurality of pronunciation character string candidates based on the phrase information for the phrase selected by the first phrase selection means;
First selection means for performing control so that any one abnormal word can be selected by the input means from among the candidates for the pronunciation character string displayed in a list ;
Second control is performed to display a heading character string candidate corresponding to the pronunciation character string candidate displayed by the control by the first display control means or selected by the first selection means together with the pronunciation character string candidate. Display control means;
Second control is performed so that when any one of the heading character strings for each phrase displayed by the control by the second display control means is designated using the input means, any one phrase is selected . Clause selection means;
For the phrase selected by the second phrase selection means, the pronunciation displayed by the control by the first display control means or selected by the first selection means among the plurality of heading character string candidates based on the phrase information Second candidate display control means for performing control to display a list of items corresponding to character string candidates;
Second selection means for performing control so that any one homophone can be selected by the input means from among the candidates for the heading character string displayed in a list ;
A text input processing device comprising:   The speech recognition is for Japanese, the phonetic character string is one of Hiragana, Katakana, or Romaji, and the heading character string is a character string mixed with Kanji characters. Text input processing device. 2. The text input processing device according to claim 1, wherein the first candidate display control unit performs control to display a character string mixed with kanji characters corresponding to a pronunciation character string candidate in a pronunciation character string candidate list. In a text input processing method for inputting text by using speech recognition,
A speech recognition result comprising a plurality of phrase information, wherein the phrase information of the same phrase includes a plurality of pronunciation character strings including a plurality of allophone words, and a plurality of homophones having different notations, and A storage step of storing a speech recognition result having a candidate of a heading character string corresponding to the candidate;
A first display control step for performing control to display any one of a plurality of pronunciation character string candidates in the same phrase side by side for each phrase;
When any one of the pronunciation character strings for each phrase displayed by the control in the first display control step is designated using the input unit, the first phrase is controlled so that any one phrase is selected. With selection step ,
A first candidate display control step for performing a control to display a list of a plurality of pronunciation character string candidates based on the phrase information for the phrase selected in the first phrase selection step;
A first selection step of performing control so that any one abnormal word can be selected by the input means from among the candidates for the pronunciation character string displayed in a list ;
A second control is performed to display a heading character string candidate corresponding to the pronunciation character string candidate displayed by the control in the first display control step or selected by the first selection step together with the pronunciation character string candidate. A display control step;
Second control is performed so that when any one of the heading character strings for each phrase displayed by the control in the second display control step is designated using the input means, any one phrase is selected . A phrase selection step;
For the phrase selected in the second phrase selection step, among the plurality of heading character string candidates based on the phrase information, the pronunciation displayed by the control in the first display control step or selected in the first selection step A second candidate display control step for performing control to display a list of items corresponding to character string candidates;
A second selection step for performing control so that any one homophone can be selected by the input means from among the candidates for the heading character string displayed in a list ;
A text input processing method comprising: A program that processes text input using speech recognition,
Computer
A speech recognition result comprising a plurality of phrase information, wherein the phrase information of the same phrase includes a plurality of pronunciation character strings including a plurality of allophone words, and a plurality of homophones having different notations, and Storage means for storing a speech recognition result having a candidate for a heading character string corresponding to the candidate;
First display control means for performing control to display one of a plurality of pronunciation character string candidates in the same phrase side by side for each phrase;
A first phrase that performs control such that when any one of the pronunciation character strings for each phrase displayed by the control by the first display control means is designated using the input means, any one phrase is selected. Selection means,
First candidate display control means for performing a control to display a list of a plurality of pronunciation character string candidates based on the phrase information for the phrase selected by the first phrase selection means;
First selection means for performing control so that any one abnormal word can be selected by the input means from among the candidates for the phonetic character strings displayed in a list ;
Second control is performed to display a heading character string candidate corresponding to the pronunciation character string candidate displayed by the control by the first display control means or selected by the first selection means together with the pronunciation character string candidate. Display control means,
Second control is performed so that when any one of the heading character strings for each phrase displayed by the control by the second display control means is designated using the input means, any one phrase is selected . Phrase selection means,
For the phrase selected by the second phrase selection means, the pronunciation displayed by the control by the first display control means or selected by the first selection means among the plurality of heading character string candidates based on the phrase information Second candidate display control means for performing control to display a list of items corresponding to character string candidates;
A program for functioning as second selection means for performing control so that any one homophone can be selected by the input means from among the candidates for the heading character string displayed in a list.

Images