JPH0944491A - Document preparation device and kana/kanji method - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a document preparing device and a KANA(Japanese syllabary)/ KANJI(Chinese character) conversion method by which a character to be the component of an other character in an input character string can be substituted by the character having the component if the character exists in the input character string and improving the conversion efficiency of the KANA character string mixed with KANJI in this substituted character string. SOLUTION: This device has a character recognition part 105 and a KANA/KANJI mixed with KANJI conversion part 108 obtaining a conversion candidate by performing a KANA/KANJI conversion for a recognized KANA character string mixed with KANJI, and is provided with a character substitution means 109 substituting an element character to be the component of an other character by a character having the element character as the component if the element character exists in a recognized character. If the element character to be the component of the other character exists in the recognized character, a KANA/KANJI conversion is performed for the character string substituted by the character having the element character as the component. When the part of a complicated character is inputted, the character is substituted by the character having the character as the component and the KANA/KANJI conversion is performed.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document creation apparatus for inputting a kana-mixed kana character string to create a kanji-mixed document, and more particularly to a document creation apparatus related to conversion candidate creation processing and a kana-kanji conversion method. .

[0002]

2. Description of the Related Art Conventionally, in a document creating apparatus, a simple kanji character such as "meeting → meeting" can be directly input, and a difficult kanji character can be converted into a kana-mixed kana character string. It was possible. In addition, even if it is converted into "Seikai → Correct answer, politics, Seikai" by entering in Kana, by entering Kanji, it is possible to narrow down the conversion candidates as "Correct → Correct answer". .

However, like "Giji → pseudo, proceedings",
When a word is composed only of complicated Kanji characters, it is impossible to utilize the effects of the conversion of Kanji characters as described above.

[0004]

As described above, in the conventional document creating apparatus, when a word is composed only of complicated Kanji characters such as "Giji → pseudo, proceedings". , It was not possible to utilize the effects of conversion of mixed Kanji characters as described above.

Therefore, the present invention has been made in view of the above circumstances, and if there is a character which is a constituent element of another character in the input character string, it is replaced with a character having that constituent element, An object of the present invention is to provide a document creation apparatus and a kana-kanji conversion method that can improve the conversion efficiency of kana-mixed kana character strings in this replaced character string.

[0006]

In order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of an input pattern and the like as a character and converts kana-kanji into a kana-kanji character string recognized. A document creation device having kanji-mixed kana-kanji conversion means for obtaining a conversion candidate by performing a character conversion, and if the recognized character has an element character that is a component of another character, a character having the component character is a character. The document creating apparatus is characterized by comprising character replacing means for replacing

Further, in order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of an input pattern and the like as a character, and performs kana-kanji conversion into a kana-kanji character string that has been recognized. A kanji-mixed kanji conversion method that obtains conversion candidates. If the recognized characters include element characters that are components of other characters, convert the character string that replaces the element characters with the characters that have the components. Kana-Kanji conversion method characterized by.

Further, in order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of an input pattern and the like as a character, and performs kana-kanji conversion into a kana-kanji character string which is recognized. A document creation device having kanji-mixed kanji conversion means for obtaining conversion candidates, which calculates an evaluation value indicating whether or not an input pattern is biased near the boundary of a character area, and is recognized based on this evaluation value. Element character determining means for determining whether or not the identified character is an element character input as a component of another character, and if there is a character among the recognized characters that is determined to be an element character by the element character determining means. According to another aspect of the present invention, there is provided a document creating apparatus including a character replacing unit that replaces an element character with a character having a constituent element.

Further, in order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of the input pattern and the like as a character, and performs kana-kanji conversion into a kana-kanji character string which is recognized. It is a kanji-mixed kanji conversion method that obtains conversion candidates, and calculates an evaluation value that indicates whether or not the input pattern is biased near the boundary of the character area, and the characters recognized based on this evaluation value are It is judged whether it is an element character input as a constituent element of the character of, and if there is a character that is judged to be an element character by the above element character judgment among the recognized characters, it is changed to the character that has the element character as a constituent element The kana-kanji conversion method is characterized by converting the replaced character string.

Further, in order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of the input pattern and the like as a character, and performs kana-kanji conversion into a kana-kanji character string that has been recognized. A document creation device having kanji-mixed kanji conversion means for obtaining conversion candidates, wherein when a user selects a conversion candidate as a result of converting a character string including a replaced character, an input pattern of characters before replacement and A pattern learning table means for storing the replaced character set, a pattern matching means for detecting a pattern matching the input pattern among the patterns stored in the pattern learning table means, and an input pattern near the boundary of the character area. An evaluation value that indicates whether or not they exist unevenly is calculated, and the character recognized based on this evaluation value is a component of other characters. Element character determining means for determining whether or not the input element character is present, and if there is a character determined to be an element character by the element character determining means among the recognized characters and the pattern stored in the pattern learning table If there is a pattern matching unit that matches the input pattern of the character, a character replacing unit that replaces the element character with the character stored in the pattern learning table in combination with the pattern is selected. When the conversion candidate is the result of converting the character string including the character replaced by the character replacing unit, the pattern learning unit that stores the input pattern and the replaced character set in the character before replacement in the pattern learning table. The document creating apparatus is characterized by being provided.

Further, in order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of an input pattern and the like as a character, and performs kana-kanji conversion into a kana-kanji character string which is recognized. It is a kanji-mixed kanji conversion method that obtains conversion candidates.If a conversion candidate that is the result of converting a character string that contains replaced characters is selected, learn the input pattern of the character before replacement and the set of replaced characters. , An evaluation value indicating whether or not the input pattern exists near the boundary of the character area is biased, and some of the characters recognized based on this evaluation value may match the input pattern of the character. For example, there is a kana-kanji learning method characterized in that kana-kanji conversion is performed on a character string in which element characters are replaced with characters learned in pairs with the pattern.

Further, in order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of an input pattern and the like as a character, and performs kana-kanji conversion into a kana-kanji character string which is recognized. A document creation device having kanji and kana-kanji conversion means for obtaining conversion candidates, wherein when a conversion candidate resulting from conversion of a character string including replaced characters is selected, an input pattern of characters before replacement and characters replaced A pattern learning table means for storing a set of patterns, a pattern matching means for detecting a pattern matching the input pattern among the patterns stored in the pattern learning table means, and the input pattern being biased near the boundary of the character area. It calculates an evaluation value that indicates whether or not the character is recognized, and the character recognized based on this evaluation value is input as a component of other characters. Element character determination means for determining whether or not an element character, and if there is a character determined to be an element character by the element character determination means among the recognized characters, the element character type is replaced with a character having an element character as a constituent element. When a pattern that is judged to match the input pattern by the character replacement means and the pattern matching means is in the pattern learning table, the character string replaced with the character stored as a pair is converted by the kanji-kana-kana-kanji conversion means. Marker adding means for adding a marker to the resulting conversion candidates, and a set of characters in which the conversion candidates to which the markers have been added by the marker adding means in the conversion candidates are replaced by the character replacement by the means are stored in the pattern learning table. And a pattern learning means for performing the pattern learning.

Further, in order to achieve the above object, the present invention has an online handwritten character recognition means for recognizing the coordinate information of an input pattern and the like as a character, and performs kana-kanji conversion into a kana-kanji character string that has been recognized. In a document creation device having kanji and kanji conversion means for obtaining conversion candidates, when a user selects a conversion candidate as a result of converting a character string including a replaced character, an input pattern of characters before replacement By learning the set of replaced characters, calculate the evaluation value that indicates that the input pattern is biased near the boundary of the character area, and based on this evaluation value, recognize other characters among the recognized characters. If there is an element character that is a constituent element, when the element character is replaced with a character that has a constituent element, if there is a pattern that matches the input pattern among the patterns learned with it. In addition, a marker is added to the conversion candidates obtained as a result of converting the character string that has been replaced with the characters learned in combination with it, and when outputting the conversion candidates, the one with the marker added is output with priority. There is a kana-kanji conversion method.

Further, in order to achieve the above object, the present invention provides the above-mentioned document creating apparatus, wherein learning is performed when a conversion candidate of a result of converting a character string including a replaced character is selected.
This is to learn a character string obtained by replacing the input input pattern string.

Further, in order to achieve the above object, the present invention provides the kana-kanji conversion method, wherein learning is input when a conversion candidate of a result of converting a character string including a replaced character is selected. This is to learn the character string that replaces the input pattern string.

Further, in order to achieve the above object, the present invention provides the above-mentioned document creating apparatus, wherein learning is performed when a conversion candidate obtained by converting a character string including a replaced character is selected.
This is to learn a character string in which the input pattern string is replaced with a word including the replaced character as a unit.

Further, in order to achieve the above object, the present invention provides the above kana-kanji conversion method, in which learning is performed when a conversion candidate resulting from conversion of a character string including a replaced character is selected. This is to learn a character string obtained by replacing the input pattern string with a word containing a character as a unit.

Further, in order to achieve the above object, the present invention provides the above-mentioned document creating apparatus, wherein learning is performed by selecting a conversion candidate as a result of converting a character string including a replaced character and inputting the replacement character. This is to learn when the evaluation value indicating whether or not it exists near the boundary of the character region of the pattern is biased and satisfies a preset threshold value.

Further, in order to achieve the above object, the present invention provides the kana-kanji conversion method, wherein learning is performed by selecting a conversion candidate as a result of converting a character string including a replaced character and inputting the replacement character. This is to learn when the evaluation value indicating whether or not it exists near the boundary of the character region of the pattern is biased and satisfies a preset threshold value.

Further, in order to achieve the above object, the present invention is such that, in the document creating apparatus, when there is an element character which is a constituent element of another character in the recognized kanji-mixed kana character string, the element character When converting a character string when an element character is replaced with a character that has as a constituent element, if there are multiple element characters, for all possible combinations, and for each element character, use the element character as a constituent element. This is to convert characters for all the characters that they have to create a character string, and to convert kana-kanji.

Further, in order to achieve the above object, the present invention provides the above-mentioned kana-kanji conversion method, in which there is an element character which is a constituent element of another character in a kana-kanji character string which has been recognized. When converting a character string when an element character is replaced with a character that has as a constituent element, if there are multiple element characters, for all possible combinations, and for each element character, use the element character as a constituent element. This is to convert characters for all the characters that they have to create a character string, and to convert kana-kanji.

Further, in order to achieve the above object, the present invention provides the above-mentioned document creating apparatus, wherein when replacing element characters,
Substitution only when the input value of each element character is biased near the boundary of the character area and the evaluation value satisfies the preset value and the character can be a constituent element of another character To do.

Further, in order to achieve the above object, the present invention provides the above kana-kanji conversion method, wherein when element characters are replaced, input patterns of various element characters are biased near the boundary of the character area. It is to substitute only when the evaluation value indicating that the character satisfies a preset value and the character can be a constituent element of another character.

[0024]

According to the above configuration, if the recognized character has an element character that is a constituent element of another character, the character string in which the element character is replaced by a character having a constituent element is converted into Kana-Kanji characters. If you enter a part of a complicated character,
Kana-Kanji conversion is performed by replacing the character with a character that has a component. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

Further, according to the above configuration, an evaluation value indicating whether or not the input pattern exists near the boundary of the character area is biased, and the character recognized based on this evaluation value is the other character. It is judged whether it is an element character input as a constituent element of, and if any of the recognized characters is judged to be an element character, a character string in which the element character is replaced with a character that has a constituent element is returned in Kana-Kanji. As a result, when a part of a complicated character is input, the character is replaced with a character having a component, and kana-kanji conversion is performed. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

Further, according to the above configuration, when the conversion candidate resulting from the conversion of the character string including the replaced character is selected by the user, the character set before replacement is learned as the set of characters replaced with the input pattern. However, there are some input patterns that are biased near the boundary of the character area, and a part of a complicated character was input by performing Kana-Kanji conversion for the character string in which the element characters were replaced with the learned characters. In this case, the kana-kanji is returned after being replaced with a character that has that character as a constituent element. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

Further, according to the above configuration, when the user selects a conversion candidate as a result of converting the character string including the replaced character, the input pattern of the character before replacement and the set of replaced characters are learned. However, if there is an element character that is a constituent element of another character in the recognized characters, the character string in which the element character is replaced with the character that is the constituent element is converted into Kana-Kanji character, and it is input in the learned pattern. If there is a match with the pattern, a marker is added to the conversion candidates obtained as a result of converting the character string that has been replaced with the characters learned as a pair, and there is a conversion candidate with the marker attached in the conversion candidates. For example, by preferentially outputting the conversion candidate, when a part of a complicated character is input, it is replaced with a character having the character as a constituent element and kana-kanji conversion is performed. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

[0028]

The outline of the present invention is as follows. That is, a document is prepared which has an online handwritten character recognition means for recognizing the coordinate information of the input pattern etc. as a character and has kana-kanji kana-kanji conversion means for obtaining conversion candidates by performing kana-kanji conversion on the recognized kana-kanji character string. In the device, if there is an element character that is a constituent element of another character among the recognized characters, the character replacement means for replacing the element character with the character that has the element character as a constituent element is provided. is there.

An on-line handwritten character recognition means for recognizing the input pattern coordinate information and the like as a character is provided, and kana-kanji kana-kanji conversion means for obtaining conversion candidates by performing kana-kanji conversion on the recognized kana-kanji character string. In the document creating device, the evaluation value indicating whether or not the input pattern is biased near the boundary of the character area is calculated, and the input character is a constituent element of another character based on the evaluation value. The element character determination means for determining whether or not the element character is input as, and the character replacement means for replacing the element character with a character having a constituent element if the recognized character is an element character. is there.

Further, an online handwritten character recognition means for recognizing the coordinate information of the input pattern etc. as a character is provided,
A kana-kanji conversion is performed on the recognized kanji-mixed kana-kanji character to obtain a conversion candidate. In a document creation device having kanji-mixed kana-kanji conversion means, a character set obtained by converting a character string including a replaced character is stored. A pattern learning table, a pattern matching means for detecting a pattern matching with the input pattern among the patterns stored in the pattern learning table, and an evaluation showing whether or not the input pattern is biased near the boundary of the character area. By the element character determination means for calculating a value and determining whether the input character is an element character input as a component of another character based on this evaluation value, and the element character determination means among the recognized characters. There is a character that is determined to be an element character, and the character is input by the pattern matching means in the pattern stored in the pattern learning table. If there is a match with the turn, the character replacement means for replacing the element character with the character stored in the pattern learning table in combination with the pattern, and the conversion candidate selected by the user by the character replacement means. In the case where the result is the result of converting the character string including the replaced character, the pattern learning unit stores the input pattern of the character before replacement and the set of the replaced character in the pattern learning table.

Further, there is an on-line handwritten character recognition means for recognizing the coordinate information of the input pattern as a character, and kana-kanji kana-kanji conversion means for obtaining conversion candidates by performing kana-kanji conversion on the recognized kana-kanji character string. In the existing document creation device, when the user selects a conversion candidate as a result of converting the character string containing the replaced character, the input pattern of the character before replacement and the set of the replaced character are stored. A pattern learning table, a pattern matching unit that detects a pattern that matches the input pattern among the patterns stored in the pattern learning table, and a pattern that the pattern matching unit determines to match the input pattern are stored in the pattern learning table. , The character string that is replaced with the character that is memorized with it is mixed with Kanji. When the conversion control unit that preferentially outputs the conversion candidate obtained as a result of the conversion by the conversion unit and the conversion candidate selected by the user is the result of converting the character string including the character replaced by the character replacement unit. In addition, the input pattern of the character before the replacement and the pattern learning means for storing the set of the replaced character in the pattern learning table are provided.

According to the above configuration, if the recognized character has an element character that is a constituent element of another character, the character string obtained by replacing the element character with a character having a constituent element is converted into Kana-Kanji characters. When a part of a complicated character is input, it is replaced with a character having the character as a constituent element and converted into kana-kanji. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

Further, according to the above configuration, an evaluation value indicating whether or not the input pattern exists near the boundary of the character area is biased, and the character recognized based on this evaluation value is the other character. It is judged whether it is an element character input as a constituent element of, and if any of the recognized characters is judged to be an element character, a character string in which the element character is replaced with a character that has a constituent element is returned in Kana-Kanji. As a result, when a part of a complicated character is input, the character is replaced with a character having a component, and kana-kanji conversion is performed. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

Further, according to the above configuration, when the conversion candidate resulting from the conversion of the character string including the replaced character is selected by the user, a character set in which the input pattern is replaced with the character before replacement is learned. However, there are some input patterns that are biased near the boundary of the character area, and a part of a complicated character was input by performing Kana-Kanji conversion for the character string in which the element characters were replaced with the learned characters. In this case, the kana-kanji is returned after being replaced with a character that has that character as a constituent element. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

Further, according to the above configuration, when the user selects a conversion candidate as a result of converting the character string including the replaced character, the input pattern of the character before replacement and the set of replaced characters are learned. However, if there is an element character that is a constituent element of another character in the recognized characters, the character string in which the element character is replaced with the character that is the constituent element is converted into Kana-Kanji character, and it is input in the learned pattern. If there is a match with the pattern, a marker is added to the conversion candidates obtained as a result of converting the character string that has been replaced with the characters learned as a pair, and there is a conversion candidate with the marker attached in the conversion candidates. For example, by preferentially outputting the conversion candidate, when a part of a complicated character is input, it is replaced with a character having the character as a constituent element and kana-kanji conversion is performed. As a result, the user's input operation can be reduced, and the document creation efficiency can be improved.

An embodiment of the present invention will be described below with reference to the drawings. In this embodiment, the bias will be described as an example of the element character. FIG. 1 is a block diagram showing the configuration of one embodiment of the present invention.

The input device 101 is a tablet or the like for inputting handwritten characters, and is a device for outputting input information such as coordinates and time of input characters. The display device 123 is a CRT
And a screen display device such as an LCD (Liquid Crystal Display) display, by which the user can interactively proceed with document creation work.

The input control unit 120 stores input information such as coordinates and time of characters input from the input device 101 in the input pattern buffer 103 while the pattern input is continued.

The character recognition unit 105 extracts input information of each pattern from the input pattern buffer 103, performs character recognition by referring to the recognition dictionary 106, and stores the recognized character in the input character buffer 107. Element character determination unit 11
2 is an evaluation value that indicates whether or not the input pattern is biased near the boundary of the character area, and the character recognized based on this evaluation value is the element character that is input as a component of another character. Determine whether or not.

The pattern learning table 119 stores the input pattern of the character before the replacement and the set of the replaced character when the user selects a conversion candidate as a result of converting the character string including the replaced character.

The pattern matching unit 113 detects a pattern that matches the input pattern from the patterns stored in the pattern learning table. The character replacement unit 109 is a kanji dictionary 11
If there is an element character that is a constituent element of another character among the characters stored in the input character buffer 107 with reference to 0, the element character is replaced with a character that has a constituent element. Especially,
If there is an element character determined by the element character determination unit 112, those elements are replaced. In addition, there is a character that is judged to be an element character by the element character judging unit 112, and a pattern stored in the pattern learning table 119 is judged to match the input pattern of the character by the pattern matching unit 113. If there is, the element character is replaced with the character stored in the pattern learning table 119 in a pair with the pattern.

The kanji-mixed kana-kanji conversion unit 108 creates a character string in which the character replacement unit 109 sequentially replaces all the replaceable characters of all the element characters that are constituents of other characters. Stored in the conversion buffer 111, the kana-kanji conversion dictionary 115 is referred to, kana-kanji conversion is performed, and the conversion result is stored in the candidate buffer.
It stores in 16.

The marker adding unit 114 is the pattern matching unit 11.
When the pattern learning table 119 determined to match the input pattern by 3 is used as a conversion candidate obtained as a result of converting the character string replaced with the character stored as a pair with the Kanji-kana-kana conversion unit. Add a marker.

The conversion candidate control unit 117 uses the candidate buffer 11
The conversion candidates stored in No. 6 are displayed to prompt the user to select candidates, and the conversion candidates selected by the user are stored in the document buffer 120. In particular, if there is a conversion candidate to which a marker is added by the marker adding means in the conversion candidates, the conversion candidates are rearranged so that the output order of the conversion candidates is prioritized.

When the conversion candidate selected by the user is the result of converting the character string including the character replaced by the character replacing unit 109, the pattern learning unit 118 determines the pattern of the character before replacement and the replaced character. Store the set in the pattern learning table.

The display control unit 121 expands the data stored in the document buffer 120 into the display buffer 122 according to the format preset by the format setting unit 104, and sends it to the display device 123.

FIG. 2 is a flow chart showing the flow of processing from character input to selection of conversion candidates in one embodiment of the document creating apparatus of the present invention. The input pattern of the handwritten character input from the input device 101 passes through the input control unit 102 and is buffered in the input pattern buffer 103 (steps 201 and 202).

The input control unit 102 repeats the processing of character input and storage in the input pattern buffer while the conversion instruction is given (steps 201 to 203). When the conversion instruction is issued, the character recognition unit 105 recognizes the input pattern stored in the input pattern buffer 103 (step 204), and the recognition result is stored in the input character buffer 107 (step 205). .

Next, the kana-kanji conversion unit 108 containing kanji
Instructs the character replacement unit 109 to enter the input character buffer 10
Characters that may be biased are searched from the characters stored in step 7 (step 206). Then, an arbitrary character among the characters that can be biased is selected (step 207), and each character is replaced with a character having each character as a constituent element and stored in the conversion buffer 111 (step 208).

Next, the kana / kanji conversion unit 108 that mixes kanji
Performs kana-kanji conversion of the character string stored in the conversion buffer 111 by referring to the kana-kanji kanji conversion dictionary 115 (step 209) and outputs the conversion result to the candidate buffer 116.
(Step 210). The process of replacing the character and storing the conversion candidate in the candidate buffer is repeated for all replaceable characters (step 21).
0). When the process of storing the conversion candidates in the candidate buffer is completed, the conversion candidate control unit 117 presents the conversion candidates to the user and waits for the user's selection process (step 21).
3). The data selected by the user is the display control unit 12
1 is displayed on the screen according to the format preset in the format setting unit 104 (step 214) (step 215).

Further, in preparation for the next input / conversion processing, the input pattern buffer 103, the input character buffer 107, the conversion buffer 111, and the candidate buffer processing 116 are initialized (step 216).

FIG. 3 is a flow chart showing the flow of processing from character input to selection of a variation candidate in one embodiment of the document creating apparatus of the present invention. The processing from the character input shown in steps 301 to 305 in the figure to the storage of the recognition result in the input character buffer is steps 201 to 205 in FIG.
And is exactly the same.

When the process of storing the recognition result in the input character buffer is completed, the element character determination unit 112 evaluates that the input pattern is biased near the boundary of the character area for each recognized character. Is calculated (step 3
06), a character input as a bias is searched from the characters recognized based on this evaluation value (step 30).
7). Then, in the character replacement unit 109, an arbitrary character is selected from the characters determined to be biased (step 308),
Each character is replaced with a character having each character as a component and stored in the conversion buffer 111 (step 30).
9).

Next, the kana / kanji / kanji conversion section 108
Performs kana-kanji conversion of the character string stored in the conversion buffer 111 by referring to the kana-kanji kanji conversion dictionary 115 (step 310), and outputs the conversion result to the candidate buffer 116.
(Step 311). The process of replacing a character and storing it in the conversion candidate buffer is repeated for all replaceable characters (steps 309 to 312).

Further, the process of selecting a character that can be biased and storing it in the candidate buffer of conversion candidates is repeated for all combinations of characters that can be biased (steps 308 to 31).
3). The process after the user candidate selection process shown in steps 314 to 317 after the process of storing the conversion candidates in the candidate buffer is the same as steps 213 to 216 in FIG.

FIG. 4 is a flow chart showing the flow of processing from character input to selection of a change method in the embodiment of the document creating apparatus of the present invention. The processing from the character input shown in steps 401 to 405 in the figure to the storage of the recognition result in the input character buffer is steps 201 to 205 in FIG.
And is exactly the same.

When the process of storing the recognition result in the input character buffer 107 is completed, the element character determination unit 112 evaluates that the input pattern of each recognized character is biased near the boundary of the character area. A value is calculated (step 406), and a character input as a bias is searched from the characters recognized based on this evaluation value (step 40).
7). Then, the pattern matching unit 113 searches whether the input pattern corresponding to the searched character is stored in the learning table (step 408).

Next, in the character input character replacement unit 109, an arbitrary character is selected from the characters retrieved by the pattern matching unit 113 (step 409), and a character having each character as a constituent element is selected for each character. Is stored in the conversion buffer 111 (step 410). And
The kanji and kana-kanji conversion unit 108 refers to the kanji and kana-kanji conversion dictionary to perform kana-kanji conversion of the character string stored in the conversion buffer 111 (step 411) and stores the conversion result in the candidate buffer 116 (step 41).
2). From the character replacement to the conversion candidate buffer 11
The process of storing in 6 is repeated for all replaceable characters (steps 410 to 413).

Further, the process of selecting a character that can be biased and storing it in the candidate buffer of conversion candidates is repeated for all combinations of characters that can be biased (steps 409 to 414).

When the process of storing the conversion candidate in the candidate buffer 116 is completed, the conversion candidate control unit 117 presents the conversion candidate to the user and waits for the user's selection process (step 415). When the conversion candidate selected by the user is a converted character string, the input pattern corresponding to the character before replacement stored in the input pattern buffer 103 in the pattern learning unit 118 The replaced character is stored in the pattern learning table 119 (steps 416 and 417).

Next, the conversion candidate selected by the user is stored in the document buffer 120 (step 418), and the data stored in the document buffer 120 is displayed by the display controller 12.
1 is displayed on the screen according to the format preset by the format setting unit 104 (step 419).

Furthermore, in preparation for the next input / conversion processing, the input pattern buffer 103, the input character buffer 107, the conversion buffer 111, and the candidate buffer 116 are initialized (step 420).

FIG. 5 is a flow chart showing the flow of processing from the character input to the selection of the conversion candidate in the embodiment of the document creating apparatus of the present invention. The processing from the character input shown in steps 501 to 505 to the storage of the recognition result in the input character buffer is shown in steps 201 to 205 in FIG.
And is exactly the same. When the process of storing the recognition result in the input character buffer 107 is completed, the element character determination unit 1
For each recognized character, 12 calculates an evaluation value indicating that the input pattern is biased near the boundary of the character area (step 506), and based on this evaluation value, the recognized characters are calculated. The character input as the bias is searched (step 506). Subsequently, in the character conversion unit 109, an arbitrary character is selected from the searched characters that may be biased (step 508), and each character is replaced with a character having each character as a constituent element, and the conversion buffer 1 is selected.
11 (step 509).

Next, in the kanji / kanji / kanji conversion section 108, kana / kanji conversion of the character string stored in the conversion buffer 111 is performed with reference to the kanji / kanji / kanji conversion dictionary (step 510). Here, the marker adding unit 114 instructs the pattern matching unit 113 to determine that the converted character string is a character string that has been replaced so as to match the set of patterns and replacement characters stored in the pattern learning table 119. ,
A marker is added to the conversion candidates obtained as a result of conversion (steps 512 and 513). Thereafter, the conversion result is stored in the candidate buffer 116 (step 514). The process of replacing the characters and storing the conversion candidates in the candidate buffer is repeated for all replaceable characters (step 50).
9 to 515).

The process of selecting a character that can be biased and storing it in the conversion candidate candidate buffer 116 is repeated for all combinations of characters that can be biased (step 508).
Through 516). When the storage process of the conversion candidates in the candidate buffer is completed, the conversion candidate control unit 117 rearranges the output order of the conversion candidates so that the conversion candidates are prioritized, if any. 517).
The processes after the selection of the candidate of the user shown in steps 518 to 523 after the rearrangement of the conversion candidates is completed correspond to the steps 415 to 420 of FIG. 4 and are completely the same.

FIG. 6 shows an example of conversion in the document creating apparatus of the present invention. When a pattern is input to the input pattern buffer 103 as shown in (a), (b)
As shown in FIG.
A kana-kana character string "kiito" recognized in 5 is stored.

The "thread" in the recognized characters is a character that can be biased, and is replaced by the Chinese character "Nao, Kumi ..." Having the component "Itohen" in the character replacement unit 109. , (C), the conversion buffer 111 sequentially stores a character string such as "Kino" and "Kikigumi". These are converted in the Kana-Kana-Kana conversion unit 108,
As shown in (d), the conversion candidate “induction” is obtained in the candidate buffer 116.

FIG. 7 shows an example of conversion in the document creating apparatus of the present invention. When a pattern is input to the input pattern buffer 103 as shown in (a), the character recognition unit 105 is added to the input character buffer 107 as shown in (d).
The character string "kiito" mixed with kanji recognized in is stored.

Next, in the element character determination unit 112, an evaluation value representing whether or not each input pattern is biased near the boundary of the character area is calculated, and other characters are calculated based on this evaluation value. Detect the element characters that are the constituent elements of. These are converted in the kanji / kanji conversion unit 108, and the conversion candidate “induction” is obtained in the candidate buffer 116 as shown in (d).

FIG. 8 shows an example of conversion in the document creating apparatus of the present invention. When a pattern is input to the input pattern buffer 103 as shown in (a), the character recognition unit 105 is added to the input character buffer 107 as shown in (b).
The character string "kiito" mixed with kanji recognized in is stored.

Next, in the element character determination unit 112, an evaluation value representing whether or not each input pattern is biased near the boundary of the character area is calculated, and other characters are calculated based on this evaluation value. Detect the element characters that are the constituent elements of. Here, the “thread” is biased to the left and is determined to be an element character.

Further, with respect to the detected element character, the pattern matching unit 113 searches the pattern learning table 119 for an input pattern matching "thread". When the result is stored in the pattern learning table 119 as shown in (e), the recognized character "thread" is replaced with "Sa" stored as a pair, and as shown in (C). To the conversion buffer 111. This is converted by the kanji / kanji conversion unit 108, and the conversion candidate “induction” is obtained in the candidate buffer 116 as shown in (d).

FIG. 9 shows an example of conversion in the document creating apparatus of the present invention. When a pattern is input to the input pattern buffer 103 as shown in (a), the character recognition unit 105 is added to the input character buffer 107 as shown in (b).
The kana-kana character string "Ki-kai" recognized in is stored.

Next, in the element character determining unit 112, an evaluation value indicating whether or not the input pattern is biased near the boundary of the character area for each input pattern is calculated, and other evaluation values are calculated based on this evaluation value. Detects element characters that are character components. Here, the “tree” exists on the left side and is determined to be an element character.

Subsequently, in the character replacing section 109, the "tree" which is judged to be biased in the element character judging section 112.
Are sequentially replaced with the Chinese characters “ki”, “ju”, etc. having “kihen” and are stored in the conversion buffer 111 as shown in (c). These are converted in the kanji / kanji / kanji conversion unit 108, and conversion candidates “machine, opportunity, jukai” are obtained in the candidate buffer 116 as shown in (d). Here, when the conversion is performed, if the converted character string is a replacement that matches the set of the pattern and the character stored in the pattern learning table 119 as shown in (e), The marker adding unit 114 adds a marker to the corresponding conversion candidate.

In this example, this corresponds to the replacement of “tree → tree”,
A marker is added to the conversion candidate “jukai”. The conversion candidate control unit 117 rearranges the conversion candidates with priority if there is a marker added, and the conversion candidates are finally stored in the candidate buffer 116 in the order shown in (f).

[0077]

As described in detail above, according to the present invention, if a character that is a constituent of another character is present in the input character string, it is replaced with a character that has that constituent, and the replaced character is replaced. The conversion efficiency of a kanji-mixed character string can be improved by reducing the input operation by the user by performing kanji-mixed kana-kanji conversion on a column and creating conversion candidates.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of an embodiment of the present invention.

FIG. 2 is a flowchart showing a processing flow from character input to selection of a conversion candidate in the document creation device according to the embodiment.

FIG. 3 is a flowchart in a state different from that of FIG. 2 showing a processing flow from character input to selection of a conversion candidate in the document creation device according to the embodiment.

FIG. 4 is a flowchart in a state different from that of FIG. 3 showing a flow of processing from character input to selection of a conversion candidate in the document creating apparatus according to the embodiment.

FIG. 5 is a flowchart showing a flow of processing from character input to selection of conversion candidates in the document creation apparatus according to the embodiment, which is different from FIG. 4;

FIG. 6 is a diagram showing a state of conversion in the document creating apparatus according to the embodiment of the present invention.

FIG. 7 is a diagram showing a state of conversion in the document creating apparatus in a state different from that of FIG. 6 according to the embodiment of the present invention.

FIG. 8 is a diagram showing an example of conversion in the document creation device of the present invention.

FIG. 9 is a view showing a conversion example in the document creating apparatus of the present invention in a state different from that of FIG.

[Explanation of symbols]

Reference numeral 101 ... Input device, 102 ... Input control unit, 103 ... Input pattern buffer, 104 ... Formatting unit, 105 ... Character recognition unit, 106 ... Recognition dictionary, 107 ... Input character buffer, 108 ... Kanji / Kana-Kanji conversion unit, 109 ... Character recognition unit, 110 ... Kanji dictionary, 111 ... Conversion buffer, 1
12 ... Element character determination unit, 113 ... Pattern matching unit, 11
4 ... Marker adding unit, 115 ... Kana-kanji conversion dictionary containing Kanji, 116 ... Candidate buffer, 117 77 ... Conversion candidate control,
118 ... Pattern learning unit, 119 ... Pattern learning table, 120 ... Document buffer, 121 ... Display control unit, 12
2 ... Display buffer, 123 ... Display device.

Claims (18)

[Claims] 1. A kana-kanji kana-kanji conversion means which has an online handwritten character recognition means for recognizing coordinate information of an input pattern as a character and obtains conversion candidates by converting kana-kanji characters into a kana-kanji character string which has been recognized. A document preparation apparatus having the above-mentioned document, characterized by comprising a character replacing means for replacing an element character with a character having a component if the recognized character has an element character which is a component of another character. Creation device. 2. A kana-kanji kana-kanji conversion method which has an online handwritten character recognition means for recognizing coordinate information etc. of an input pattern as a character, and obtains conversion candidates by performing kana-kanji conversion on a kana-kanji character string containing recognized kanji. A kana-kanji conversion method characterized by converting a character string in which a character having an element character as a constituent element is converted if the recognized character has an element character that is a constituent element of another character. 3. A kana-kanji kana-kanji conversion means which has an online handwritten character recognition means for recognizing coordinate information of an input pattern and the like as a character, and obtains conversion candidates by performing kana-kanji conversion on a kana-kanji character string which has been recognized. A document creating apparatus having the above, wherein an evaluation value indicating whether or not an input pattern is present near a boundary of a character region is biased, and a character recognized based on this evaluation value is a constituent element of another character. Element character determining means for determining whether or not it is an element character input as, and if there is a character that is recognized as an element character by the element character determining means among the recognized characters, it is changed to a character having an element character as a component. A document creating apparatus comprising: a character replacing unit for replacing. 4. A kana-kanji kana-kanji conversion method having an online handwritten character recognition means for recognizing input pattern coordinate information and the like as a character, and performing kana-kanji conversion on a recognized kana-kanji character string to obtain conversion candidates. Then, an evaluation value representing whether or not the input pattern is biased near the boundary of the character area is calculated, and the character recognized based on this evaluation value is input as a component of another character. It is judged whether it is an element character, and if there is a character that is judged to be an element character by the above element character judgment among the recognized characters, it is converted into a character string in which the element character is replaced with a character that has a component. Characteristic kana-kanji conversion method. 5. A kana-kanji kana-kanji conversion means that has an online handwritten character recognition means for recognizing coordinate information of an input pattern as a character, and obtains conversion candidates by performing kana-kanji conversion for a kana-kanji character string that has been recognized. A document creation device having the above, a pattern for storing the input pattern of the character before replacement and the set of the replaced character when the conversion candidate resulting from the conversion of the character string including the replaced character is selected by the user. A learning table means, a pattern matching means for detecting a pattern matching the input pattern among the patterns stored in the pattern learning table means, and whether or not the input pattern is biased near the boundary of the character area. Calculate the evaluation value to represent, and the character recognized based on this evaluation value is the element character that is input as a component of other characters. Element character determination means for determining whether or not there is a character determined to be an element character by the element character determination means among the recognized characters and the pattern stored in the pattern learning table is determined by the pattern matching means. If it is determined that there is a match with the character input pattern, the character replacement means for replacing the element character with the character stored in the pattern learning table in combination with the pattern, and the selected conversion candidate is the character replacement means. And a pattern learning means for storing a set of the input pattern and the replaced character in the pattern learning table in the case of a result of converting a character string including the character replaced by Document creation device. 6. A kana-kanji kana-kanji conversion method that has an online handwritten character recognition means for recognizing the coordinate information of an input pattern as a character, and obtains conversion candidates by performing kana-kanji conversion on a kana-kanji character string that has been recognized. And when the conversion candidate of the result of converting the character string including the replaced character is selected, the input pattern of the character before replacement and the set of replaced characters are learned, and the input pattern is near the boundary of the character area. An evaluation value indicating whether or not there is a bias is present, and if any of the characters recognized based on this evaluation value matches the input pattern of that character, it is learned as a set with that pattern. A kana-kanji learning method characterized by converting kana-kanji into a character string in which element characters have been replaced with kanji characters. 7. A kana-kanji kana-kanji conversion means which has an online handwritten character recognition means for recognizing coordinate information of an input pattern as a character and obtains conversion candidates by performing kana-kanji conversion on a recognized kanji-kana character string. A document creating apparatus having the pattern learning table means for storing an input pattern of a character before replacement and a set of replaced characters when a conversion candidate resulting from conversion of a character string including the replaced character is selected. And a pattern matching means for detecting a pattern matching the input pattern among the patterns stored in the pattern learning table means, and an evaluation value indicating whether or not the input pattern is biased near the boundary of the character area. And determine whether the recognized character is an element character input as a component of another character based on this evaluation value. If there is a character which is judged to be an element character by the element character judgment means among the recognized characters, a character replacement means for replacing the element character type with a character having the element character as a constituent element, When a pattern judged to match the input pattern by the pattern matching means exists in the pattern learning table, a conversion candidate obtained as a result of converting the character string replaced with the character stored as a pair with the kanji mixed kana kanji conversion means Marker adding means for adding a marker to the pattern candidate, and pattern learning means for storing in the pattern learning table a character set in which the conversion candidate having the marker added by the marker adding means in the conversion candidate is replaced by the character replacement means. A document creation device comprising: 8. A kana-kanji kana-kanji conversion means that has an online handwritten character recognition means for recognizing coordinate information of an input pattern as a character and obtains conversion candidates by performing kana-kanji conversion on a kana-kanji character string containing recognized kanji. If the user selects a conversion candidate resulting from the conversion of the character string containing the replaced character, the document creation device that has the input pattern learns the input pattern of the character before replacement and the set of the replaced character, and inputs it. Calculates the evaluation value that indicates that the pattern is biased near the boundary of the character area, and if there is an element character that is a constituent element of another character among the characters recognized based on this evaluation value, the element When a character is replaced with a character that has a component, if there is a pattern that is learned as a pair with it and matches the input pattern, it is placed in the character that is learned as a pair with that. Kana-Kanji conversion method and outputting the priority to which the marker is added to when and by adding a marker string to the conversion result obtained conversion candidates, and outputs the conversion candidates. 9. The learning is a learning of a character string obtained by replacing an input pattern string input when a conversion candidate resulting from conversion of a character string including a replaced character is selected. The document creation device according to claim 3, 5, or 7. 10. The learning is for learning a character string in which an input pattern string input is replaced when a conversion candidate resulting from conversion of a character string including a replaced character is selected. Claim 4, or Claim 6, or Claim 8
Kana-Kanji conversion method described. 11. In learning, when a conversion candidate resulting from conversion of a character string including a replaced character is selected, a character string in which the input pattern string is replaced is learned in units of words including the replaced character. The document creation device according to claim 3, 5, or 7, wherein the document creation device is a document creation device. 12. In learning, when a conversion candidate resulting from conversion of a character string including a replaced character is selected, a character string in which the input pattern string is replaced is learned in units of words including the replaced character. 9. The kana-kanji conversion method according to claim 4, 6, or 8. 13. The learning is performed by checking whether or not a conversion candidate obtained as a result of converting a character string including a replaced character is biased and exists near a boundary of a character region of an input pattern of the replacement character. 4. The learning is performed when the evaluation value to be expressed satisfies a preset threshold value.
Item, or the document creation device according to claim 5 or claim 7. 14. The learning is performed by checking whether or not a conversion candidate obtained as a result of converting a character string including a replaced character is biased and exists near the boundary of the character area of the input pattern of the replacement character. 5. The learning is performed when the evaluation value to be expressed satisfies a preset threshold value.
Alternatively, the kana-kanji conversion method according to claim 6 or claim 8. 15. A character string is converted when an element character is replaced with a character having an element character as a constituent element when there is an element character that is a constituent element of another character in the recognized kanji-mixed kana character string. When there are multiple element characters, all possible combinations are created, and for each element character, all the characters that have the element character as a component are replaced to create a character string, and kana-kanji conversion is performed. The document creation apparatus according to claim 1, claim 3, claim 5, or claim 7. 16. A character string is converted when an element character is replaced with a character having an element character as a constituent element when there is an element character that is a constituent element of another character in the recognized kanji-mixed kana character string. When there are multiple element characters, all possible combinations are created, and for each element character, all the characters that have the element character as a component are replaced to create a character string, and kana-kanji conversion is performed. The kana-kanji conversion method according to claim 2, or claim 4, or claim 6, or claim 8. 17. When substituting element characters, an evaluation value indicating that the input patterns of various element characters are biased near the boundary of the character area and satisfies a preset value,
The document creating apparatus according to claim 3, 5, or 7, wherein the character is replaced only when the character can be a constituent element of another character. 18. When substituting element characters, an evaluation value indicating that the input patterns of various element characters are biased near the boundary of the character area and satisfies a preset value,
The kana-kanji conversion method according to claim 4, 6, or 8, wherein the character is replaced only when the character can be a constituent of another character.