JPH09212499A - Character string predicting method and document prepartion device using the same - Google Patents

Abstract

PROBLEM TO BE SOLVED: To eliminate the need for correcting operation for misinput and to make character string prediction accurate by outputting a KANA(Japanese syllabary) character string, having its KANJI(Chinese character) part converted in a predicted character string at the head part, from a word dictionary to a predicted character string candidate. SOLUTION: If an input operator inputs wrong KANJI to a character input part 11 when inputting a character string, the character string is stored in a character string buffer 12 and then taken out and passed to a character string conversion part 13. Here, a KANJI/KANA character conversion part 131 converts the wrong input part into KANA characters by referring to a KANJI/ KANA character dictionary 13. A character string retrieval part 14 retrieves a predicted character string from a word dictionary 15 according to the converted KANA characters. Here, if the KANA characters are registered in the dictionary 15, the retrieval part 14 decides the character string as a candidate to outputs KANJI characters from a output part 16 and displays them on the screen of a display part 17.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a character string prediction method for predicting a character string subsequent to an input character string based on a character string input by an input device, and a method of using the character string prediction method. Document creation device.

[0002]

2. Description of the Related Art Conventionally, in a document creating apparatus such as a Japanese word processor, when inputting a character string using a handwritten character recognition apparatus or the like, one character is input unlike a keyboard input. Since it takes time, the burden on the input person was large.

Therefore, in order to reduce the burden on the input person, when a part of the character string to be input is input, the character string following the character string is predicted based on the input character string. That method has been considered.

This is because, for example, in inputting a character string "Kanji dictionary", when the input person inputs "Kanji", the "dictionary" is predicted as a character string following the "Kanji", and finally "Kanji". It outputs "Kanji dictionary".

[0005]

However, the above-described character string prediction method is for predicting a character string that follows, assuming that the character string to be predicted is correctly input. Therefore, if a character string is entered by mistake, for example, if you make a wrong Kana-Kanji conversion when entering a character string, or if you make a mistake in recognizing a handwritten character when you enter it, There was a problem that the prediction became difficult.

In order to accurately predict a character string, the input person checks each time whether or not the input character string is correctly converted or recognized, and if it is incorrect, the character string is checked. There was a problem that it required a troublesome work such as correction, and rather burdened.

The present invention has been made in consideration of the above circumstances. Even if a character string to be predicted is erroneously input, an accurate character can be obtained without requiring correction work by the input person. An object of the present invention is to provide a character string prediction method capable of predicting a string and a document creation apparatus using this character string prediction method.

[0008]

[Means for Solving the Problems]

(1) The present invention provides a character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, or a document creation apparatus using the same method. Converts the Kanji part of the character string input as to Kana characters, and based on the converted character string, searches the word dictionary for a character string that has that character string at the beginning as a predicted character string, and then searches that character string. The predicted character string candidates are output.

According to this structure, the kanji part in the input character string is converted into the kana character, and the word dictionary is searched based on the converted character string. Then, when the relevant character string is searched, all the searched character strings are extracted and output as a predicted character string group.

Accordingly, even if a character string having an incorrect kanji character is input as a prediction target by kana-kanji conversion, for example, the character string that the input person is about to input can be correctly predicted. Further, the input person can obtain the predicted character string without correcting the character string input by mistake, and the burden on the input can be reduced.

(2) The present invention provides a character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, or a document creating apparatus using the same method. , The different character type part in the character string input as the prediction target is removed, and based on the removed character string, a character string having the character string at the beginning is searched as a predicted character string from the word dictionary, and the This is to output the retrieved predicted character string candidates.

According to this structure, the different character type portion in the input character string is removed from the search target, and the word dictionary is searched based on the removed character string. Then, when the relevant character string is searched, all the searched character strings are extracted and output as a predicted character string group.

Thus, for example, even if a character string having a character erroneously recognized by character recognition is input as a prediction target, the character string that the input person is about to input can be correctly predicted. Further, the input person can obtain the predicted character string without correcting the character string input by mistake, and the burden on the input can be reduced.

(3) The present invention provides a character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, or a document creation apparatus using the method. , Search a character similar to the character in the character string input as the prediction target from the similar character dictionary, convert the corresponding character to a similar character, and based on the converted character string, start that character string Is searched for as a predicted character string from the word dictionary, and the candidate of the searched predicted character string is output.

With such a configuration, the characters in the input character string are converted into similar characters, and the word dictionary is searched based on the converted character string. Then, when the relevant character string is searched, all the searched character strings are extracted and output as a predicted character string group.

Thus, for example, even when a character string having a character erroneously recognized by character recognition is input as a prediction target, the character string that the input person is about to input can be correctly predicted. Further, the input person can obtain the predicted character string without correcting the character string input by mistake, and the burden on the input can be reduced.

(4) The present invention provides a character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by the input person, or a document creation apparatus using the same method. , Judge the recognition evaluation value of the character in the character string input as the prediction target, remove the character with the low recognition evaluation value, and based on the removed character string, the character that has that character string at the beginning part The column is searched as a predicted character string from the word dictionary, and the candidate of the searched predicted character string is output.

According to such a configuration, the character having a low recognition evaluation value in the input character string is removed from the search target,
The word dictionary is searched based on the removed character string. Then, when the relevant character string is searched, all the searched character strings are extracted and output as a predicted character string group.

Thus, for example, even when a character string having a character erroneously recognized by character recognition is input as a prediction target, the character string that the input person is about to input can be correctly predicted. Further, the input person can obtain the predicted character string without correcting the character string input by mistake, and the burden on the input can be reduced.

(5) The present invention provides a character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, or a document creation apparatus using the same method. , The recognition evaluation value of the character in the character string input as the prediction target is determined, a character similar to a character with a low recognition evaluation value is searched from the similar character dictionary, and the corresponding character is converted to the similar character, Based on the converted character string, a character string having the character string at the beginning is searched as a predicted character string from the word dictionary, and the searched predicted character string candidates are output.

With such a configuration, the characters having a low recognition evaluation value in the input character string are converted into similar characters, and the word dictionary is searched based on the converted character string. Then, when the relevant character string is searched, all the searched character strings are extracted and output as a predicted character string group.

Thus, for example, even when a character string having a character erroneously recognized by character recognition is input as a prediction target, the character string that the input person is about to input can be correctly predicted. Further, the input person can obtain the predicted character string without correcting the character string input by mistake, and the burden on the input can be reduced.

[0023]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

(First Embodiment) FIG. 1 is a block diagram showing the arrangement of a document creation apparatus according to the first embodiment of the present invention. This device is provided with a character string input unit 11 and a character string buffer 1.
2, character string conversion unit 13, character string search unit 14, word dictionary 1
5, an output unit 16, and a display unit 17.

The character string input unit 11 is for inputting a character string to be predicted. In the present embodiment, the key input system is assumed, and the keyboard 111 and the Kana-Kanji conversion unit 11 are assumed.
2. It is composed of the conversion dictionary 113.

The keyboard 111 is an input device for keying in a character string. This keyboard 111 has
In addition to various input keys such as "character" key and "number" key,
Here, the prediction key 11 for instructing the start of the prediction operation
1a.

The kana-kanji conversion unit 112 uses the keyboard 11
Kana-Kanji conversion is performed on the character string input through 1. The character string (conversion result) obtained by this kana-kanji conversion is temporarily stored in the character string buffer 12 and then cleared when the input person confirms the character string. However, when the start of the prediction operation is instructed by pressing the prediction key 111a, the kana-kanji conversion unit 112 causes the character string buffer 12
The character string stored in is extracted and passed to the character string conversion unit 13 as a prediction target character string.

The conversion dictionary 113 includes a kana-kanji conversion unit 112.
It is a dictionary used in, and stores various information necessary for kana-kanji conversion processing, such as reading and headwords corresponding to the reading.

The character string buffer 12 is used to temporarily store the character string before being determined. The character string conversion unit 13 performs a process for converting a prediction target character string into a predictable character string, and in the present embodiment, the Kanji / Kana character conversion unit 131 and the Kanji / Kana character dictionary 1
It consists of 32.

The kanji / kana character conversion section 131 converts the kanji part of the input character string into kana characters and passes the converted character string to the character string search section 14. The Kanji / Kana character dictionary 132 is a dictionary used in the Kanji / Kana character conversion unit 131, and stores information for searching a Kana character corresponding to the Kanji / Kana character.

The character string searching unit 14 searches the word dictionary 15 for a character string having the character string at the beginning based on the character string obtained by the character string converting unit 13, and if the search is successful, The output unit 16 is notified of the character string as a candidate for the predicted character string. The word dictionary 15 is a dictionary used for prediction processing, and stores various words in both Kanji and Hiragana notation.

The output unit 16 outputs to the display unit 17 the predicted character string candidates retrieved by the character string retrieval unit 14.
The display unit 17 is, for example, an LCD (Liquid Crystal Displa).
y), a CRT (Cathode Ray Tube), or other display device, and presents the predictor character string candidates received from the output unit 16 to the input person.

Next, the operation of the first embodiment will be described.

FIG. 2 is a flow chart showing the operation of the prediction process in the first embodiment. First, a character string is input by the character string input unit 11 (step 10). In this case, in the first embodiment, the character string input unit 11 includes a keyboard 111, a kana-kanji conversion unit 112, and a conversion dictionary 113, and the kana-kanji conversion result of the character string keyed by the input person is used as the input character string. .

This input character string (conversion result) is temporarily stored in the character string buffer 12 (step 11). The input character string (conversion result) stored in the character string buffer 12 is displayed on the display unit 17 through the output unit 16.

Here, when the input person instructs the start of the prediction operation by pressing the prediction key 111a provided on the keyboard 111 (Yes in step 12), the character string stored in the character string buffer 12 becomes the prediction target character string. And is given to the character string conversion unit 13.

The kanji / kana character conversion part 131 in the character string conversion part 13 takes out the kanji part of the input character string and converts the kanji into kana characters by referring to the kanji / kana character dictionary 132 ( Steps 13, 14, 15).

When the Kanji part of the input character string is converted into the kana character in this way, the character string conversion unit 13 passes the converted character string to the character string search unit 14. Character string search unit 14
Searches the word dictionary 15 for a character string having the character string received from the character string conversion unit 13 at the beginning, retrieves the character string as a predicted character string candidate when the search is successful, and outputs the character string. 16 (step 16).

The output unit 16 passes the predicted character string candidates passed from the character string search unit 14 to the display unit 17 (step 1).
7). The display unit 17 displays the character string group requested to be displayed on the screen (step 18).

Next, a specific example will be described.

For example, in the case of inputting a character string "Kanji dictionary", the input person intends to input "Kanji", but after inadvertently converting it as "substitution" by Kana-Kanji conversion, an instruction to start the prediction operation is given. And

When the character string "conversion" (conversion result) is input through the character string input unit 11, the character string is temporarily stored in the character string buffer 12 because it is used as a key for predictive character string retrieval. To be done. In response to an instruction to start the prediction operation, the character string input unit 11 takes out a character string used for the prediction process, here, “substitution” from the character string buffer 12 and passes it to the character string conversion unit 13.

The Kanji / Kana character conversion unit 131 in the character string conversion unit 13 refers to the Kanji / Kana character dictionary 132 to change the Kanji character of the input character string “Kanji” into a Kana character and searches for a character string. Hand it over to section 14. In this case, since each character of “substitution” is a Kanji, the Kanji / Kana character conversion unit 131 converts the “substitution” into “Kanji” and passes it to the character string search unit 14.

The character string search unit 14 searches the word dictionary 15 for a predicted character string for the input character string, using the converted character string "Kanji" as a key.

Here, assuming that words such as “Kanji dictionary” in Kanji and “Kanjijisho” in Hiragana are registered in the word dictionary 15, the character string search unit 1
4 extracts the character string of the word from the word dictionary 15 as a candidate of the predicted character string for the input character string, and outputs it to the output unit 16.

In this case, which of the kanji notation and the hiragana notation is used to output the candidate of the predicted character string,
It should match the notation of the input character string. In this example,
The input character string is "substitution", which is in Kanji. Therefore, when outputting the predicted character string, the “kanji dictionary” of kanji is output in accordance with the input notation.

The output unit 16 transfers this "Kanji dictionary" to the display unit 17. The display unit 17 displays "substitution" input by the input person.
Instead of, the "Kanji dictionary" is displayed on the screen, and the processing here ends.

By the above processing, the character string "Kanji dictionary" which the input person was trying to input can be obtained,
As a result, the work of inputting the "dictionary" can be omitted. At the same time, the work of correcting the erroneous input portion "substitution" can be omitted.

In the first embodiment described above, the character string input unit 11 assuming the key input method is used, and the Kana-Kanji conversion is performed on the result of Kana-Kanji conversion of the character string key-input by the input person. Although prediction processing is performed by changing to characters, this method is also applicable to handwriting input methods such as second to fifth embodiments described later.
In this case, a method similar to the above may be applied to the result obtained by recognizing a character string input by handwriting and converting the recognized character string into kana-kanji.

(Second Embodiment) Next, a second embodiment of the present invention will be described.

FIG. 2 is a block diagram showing the arrangement of a document creating apparatus according to the second embodiment of the present invention. The present apparatus includes a character string input unit 21, a character string buffer 22, and a character string conversion unit 2.
3, a character string search unit 24, a word dictionary 25, an output unit 26, and a display unit 27.

The character string input unit 21 is for inputting a character string to be predicted, and in the present embodiment, a handwriting input method is assumed and is composed of a tablet 211, a recognition unit 212, and a recognition dictionary 213.

The tablet 211 is an input device for handwriting input of a character string. This tablet 211 is
A pen for handwriting input is used, and handwriting input of a character string is performed by the writing operation of the pen. Further, here, a prediction button 211a for instructing the start of the prediction operation is provided.

The recognition unit 212 character-recognizes the character string input by handwriting through the tablet 211. The character string (recognition result) obtained by this character recognition is temporarily stored in the character string buffer 22 and then cleared when the input person confirms the character string. However, the prediction button 21
When the start of the prediction operation is instructed by pressing 1a, the recognition unit 212 extracts the character string stored in the character string buffer 22 and passes this to the character string conversion unit 23 as the prediction target character string.

The recognition dictionary 213 is a dictionary used in the recognition unit 212, and stores various kinds of information necessary for character recognition processing, such as information on each stroke forming a character.

Further, the character string buffer 22 is used to temporarily store the character string before being determined. The character string conversion unit 23 performs a process for converting the prediction target character string into a predictable character string, and is composed of the different character type removal unit 231 in the present embodiment.

The different character type removing unit 231 checks the character type of the input character string, removes characters of different character types such as symbol characters and numbers from the input character string, and retrieves the character string after the removal. Pass to.

The character string search unit 24 searches the word dictionary 25 for a character string having the character string at the beginning based on the character string obtained by the character string conversion unit 23. If the search is successful, The output unit 26 is notified of the character string as a candidate for the predicted character string. The word dictionary 25 is a dictionary used for prediction processing, and stores various words in both Kanji and Hiragana notation.

The output unit 26 outputs the predicted character string candidates retrieved by the character string retrieval unit 24 to the display unit 27.
The display unit 27 is, for example, an LCD (Liquid Crystal Displa).
y) or a CRT (Cathode Ray Tube) or the like, and presents the predictive character string candidates received from the output unit 26 to the input person.

Next, the operation of the second embodiment will be described.

FIG. 4 is a flow chart showing the operation of the prediction process in the second embodiment. First, a character string is input by the character string input unit 21 (step 30). In this case, in the second embodiment, the character string input unit 21 includes the tablet 211, the recognition unit 212, and the recognition dictionary 213,
The input character string is the result of character recognition of the character string handwritten by the input person.

This input character string (recognition result) is temporarily stored in the character string buffer 22 (step 31). The input character string (recognition result) stored in the character string buffer 22 is displayed on the display unit 27 through the output unit 26.

Here, when the input person instructs the start of the prediction operation by pressing the prediction button 211a provided on the tablet 211 (Yes in step 32), the character string stored in the character string buffer 22 is the prediction target character string. And is given to the character string conversion unit 23.

Different character type removing unit 2 in the character string converting unit 23
31 checks the character type of the input character string,
Remove different characters such as numbers from the search target (Step 3
3, 34, 35).

In this way, when the different character type is removed from the input character string, the character string conversion unit 23 passes the removed character string to the character string search unit 24. The character string search unit 24
The word dictionary 25 is searched for a character string having the character string received from the character string conversion unit 23 at the beginning, and if the search is successful, the character string is extracted as a candidate for the predicted character string, and is output to the output unit 26. Hand over (step 36).

The output unit 26 passes the predicted character string candidates passed from the character string searching unit 24 to the display unit 27 (step 3).
7). The display unit 27 displays the character string group requested to be displayed on the screen (step 38).

Next, a specific example will be described.

For example, when the character string “Haru no Ogawa” is input, the input person intends to input “Haruno”, but after the character recognition is mistakenly recognized as “Haruo”, the start of the prediction operation is instructed. It is assumed that

When the character string “recognition result” is input through the character string input unit 21, the character string is temporarily stored in the character string buffer 22 because it is used as a key for predictive character string retrieval. Stored in. In response to an instruction to start the prediction operation, the character string input unit 21 takes out a character string used for prediction processing, here, “Haruo”, from the character string buffer 22 and passes it to the character string conversion unit 23.

Different character type removing unit 231 of the character string converting unit 23.
Checks the character type of the input character string “Haruo”, removes the different character type from it, and passes it to the character string search unit 24.
In this case, since only “∘” is a symbol, the different character type removing unit 231 excludes “∘” from the search target and passes a character string such as “Haru” to the character string searching unit 24.

The character string search unit 24 searches the word dictionary 25 for a predicted character string corresponding to the input character string by using the removed character string "Haru" as a key.

Here, assuming that words such as "Spring Stream" in kanji and "Haru no Ogawa" in hiragana are registered in the word dictionary 25, the character string search unit 2
4 outputs the character string of the word from the word dictionary 25 as a candidate for the predicted character string for the input character string, and outputs it to the output unit 26.

In this case, which of the kanji notation and the hiragana notation is used to output the predicted character string candidate is determined by
It should match the notation of the input character string. In this example,
The input character string is "Haru ○", which is written in Hiragana.
Therefore, when the predicted character string is output, "Haru no Ogawa" in Hiragana is output in accordance with the input.

The output unit 26 passes this "Haru no Ogawa" to the display unit 27. The display unit 27 displays "Haru no Ogawa" on the screen instead of "Haru ○" input by the input person,
The processing here is finished.

By the above processing, the character string "Haru no Ogawa" that the input person was trying to input can be obtained, and as a result, the input work of the portion "Ogawa" and the correction of "○" to "NO" You have saved the trouble of doing it.

In the second embodiment, the case where the prediction process is performed on the result of character recognition of the character string input by handwriting has been described. For example, after the prediction process, the character string is converted into Kana-Kanji character and A configuration in which the result is finally output is also possible.

(Third Embodiment) Next, a third embodiment of the present invention will be described.

FIG. 5 is a block diagram showing the arrangement of a document creating apparatus according to the third embodiment of the present invention. This device is provided with a character string input unit 31, a character string buffer 32, and a character string conversion unit 3.
3, a character string search unit 34, a word dictionary 35, an output unit 36, and a display unit 37.

The character string input unit 31 is for inputting a character string to be predicted, and in the present embodiment, a handwriting input method is assumed and is composed of a tablet 311, a recognition unit 312, and a recognition dictionary 313.

The tablet 311 is an input device for handwriting input of a character string. This tablet 311
A pen for handwriting input is used, and handwriting input of a character string is performed by the writing operation of the pen. Further, here, a prediction button 311a for instructing the start of the prediction operation is provided.

The recognition unit 312 character-recognizes the character string input by handwriting through the tablet 311. The character string (recognition result) obtained by this character recognition is temporarily stored in the character string buffer 32, and then cleared when the input person confirms the character string. However, the prediction button 31
When the start of the prediction operation is instructed by pressing 1a, the recognition unit 312 takes out the character string stored in the character string buffer 32 and passes this to the character string conversion unit 33 as the prediction target character string.

The recognition dictionary 313 is a dictionary used in the recognition unit 312, and stores various kinds of information necessary for character recognition processing, such as information on each stroke forming a character.

The character string buffer 32 is used to temporarily store the character string before being determined. The character string conversion unit 33 performs processing for converting the prediction target character string into a predictable character string, and in the present embodiment, the similar character conversion unit 331, the similar character search unit 332, and the similar character dictionary 333. Composed of.

The similar character conversion unit 331 requests the similar character search unit 332 to search for a similar character for each character of the input character string, and if there is a similar character, replaces the character with the similar character. . Similar character search unit 332
Is a similar character dictionary 33 to determine whether or not there is a similar character that is easily misrecognized by character recognition in the input character string.
When the search is successful, the similar character is passed to the similar character conversion unit 331. Similar character dictionary 3
33 has a structure in which similar characters that are easily misrecognized can be searched.

The character string searching unit 34 searches the word dictionary 35 for a character string having the character string at the beginning based on the character string obtained by the character string converting unit 33, and if the search is successful, The character string is notified to the output unit 36 as a predicted character string candidate. The word dictionary 35 is a dictionary used for prediction processing, and stores various words in both Kanji and hiragana notation.

The output unit 36 outputs the predicted character string candidates retrieved by the character string retrieval unit 34 to the display unit 37.
The display unit 37 is, for example, an LCD (Liquid Crystal Displa).
y) or a CRT (Cathode Ray Tube) or the like, and presents the predictive character string candidates received from the output unit 36 to the input person.

Next, the operation of the third embodiment will be described.

FIG. 6 is a flow chart showing the operation of the prediction process in the third embodiment. First, a character string is input by the character string input unit 31 (step 50). In this case, in the third embodiment, the character string input unit 31 includes a tablet 311, a recognition unit 312, and a recognition dictionary 313,
The input character string is the result of character recognition of the character string handwritten by the input person.

This input character string (recognition result) is temporarily stored in the character string buffer 32 (step 51). The input character string (recognition result) stored in the character string buffer 32 is displayed on the display unit 37 through the output unit 36.

Here, when the input person instructs the start of the prediction operation by pressing the prediction button 311a provided on the tablet 311, the character string stored in the character string buffer 32 is the prediction target character string. And is given to the character string conversion unit 33.

Similar character search unit 332 of character string conversion unit 33
For each character of the input character string, the similar character dictionary 332 is searched to determine whether or not there is a similar character that is likely to be erroneously recognized by character recognition, and when the search is successful, the similar character is similar. It is passed to the character conversion unit 333 (steps 53 and 54).

Similar character converter 3 in the character string converter 33
For each character of the input character string, 31 requests the similar character search unit 332 to search for a similar character, and if a similar character exists, replaces that character with a similar character (steps 55 and 56).

In this way, when the characters in the input character string are converted into similar characters, the character string conversion unit 33 passes the removed character string to the character string search unit 34. Character string search unit 34
Searches the word dictionary 35 for a character string having the character string received from the character string conversion unit 33 at the beginning, retrieves the character string as a candidate for the predicted character string when the search is successful, and outputs the output unit. 36 (step 57).

The output unit 36 passes the predicted character string candidates passed from the character string search unit 34 to the display unit 37 (step 5).
8). The display unit 37 displays the character string group requested to be displayed on the screen (step 59).

Next, a specific example will be described.

For example, when the character string “preventive injection” is input, the input person intends to input “prevention”, but after the character recognition incorrectly recognizes “prevention”, the start of the prediction operation is instructed. I shall.

When the character string “recognition” (recognition result) is input through the character string input unit 31, the character string is temporarily used in the character string buffer 32 because it is used as a key for predictive character string retrieval. Is stored. In response to an instruction to start the prediction operation, the character string input unit 31 takes out a character string used for the prediction process, here, “end” from the character string buffer 32, and passes it to the character string conversion unit 33.

Similar character search unit 3 in the character string conversion unit 33
32 refers to the similar character dictionary 333 to search for each character in "endou" whether or not there is a similar character that is easily erroneously recognized by character recognition or the like. As a result, the similar character search unit 332 knows that “end” and “yo” are similar characters, and passes it to the similar character conversion unit 331. As a result, the similar character conversion unit 331 converts “end” into “pre” to obtain the character string “prevention”, and passes this to the character string search unit 34.

The character string search unit 34 searches the word dictionary 35 for a predicted character string for the input character string, using the converted character string "prevention" as a key.

Here, assuming that words such as “preventive injection” in kanji notation and “yobouchisha” in hiragana notation are registered in the word dictionary 35, the character string search unit 34 causes the character string of that word to be registered. Is extracted from the word dictionary 35 as a candidate for the predicted character string for the input character string, and is output to the output unit 36.

In this case, which of the kanji notation and the hiragana notation is used to output the candidate of the predicted character string,
It should match the notation of the input character string. In this example,
The input character string is “Ryo”, which is written in Kanji. Therefore, when the predicted character string is output, the Chinese character “preventive injection” is output according to the input.

The output unit 36 passes this "preventive injection" to the display unit 37. The display unit 37 displays “End” input by the input person.
Instead of, "Preventive injection" is displayed on the screen, and the process here ends.

By the above processing, the character string "preventive injection" which the input person was trying to input can be obtained,
As a result, it is possible to save the work of inputting "injection" and the trouble of correcting "end" to "preliminary".

In the third embodiment, the case where the predicting process is performed on the result of character recognition of the character string input by handwriting has been described. However, for example, after the predicting process, the character string is converted into kana-kanji characters, A configuration in which the result is finally output is also possible.

(Fourth Embodiment) Next, a fourth embodiment of the present invention will be described.

FIG. 7 is a block diagram showing the arrangement of a document creating apparatus according to the fourth embodiment of the present invention. The present apparatus includes a character string input unit 41, a character string buffer 42, and a character string conversion unit 4
3, a character string search unit 44, a word dictionary 45, an output unit 46, and a display unit 47.

The character string input unit 41 is for inputting a character string to be predicted, and in the present embodiment, a handwriting input method is assumed and is composed of a tablet 411, a recognition unit 412, and a recognition dictionary 413.

The tablet 411 is an input device for handwriting input of a character string. This tablet 411
A pen for handwriting input is used, and handwriting input of a character string is performed by the writing operation of the pen. Further, here, a prediction button 411a for instructing the start of the prediction operation is provided.

The recognition unit 412 character-recognizes the character string input by handwriting through the tablet 411. The character string (recognition result) obtained by this character recognition is temporarily stored in the character string buffer 42, and then cleared when the input person confirms the character string. However, the prediction button 41
When the start of the prediction operation is instructed by pressing 1a, the recognition unit 412 takes out the character string stored in the character string buffer 42 and passes this to the character string conversion unit 43 as the prediction target character string.

The recognition dictionary 413 is a dictionary used in the recognition unit 412, and stores various kinds of information necessary for character recognition processing, such as information on each stroke forming a character.

Further, the character string buffer 42 is used to temporarily store the character string before being determined. The character string conversion unit 43 performs a process for converting the prediction target character string into a predictable character string, and is composed of the unnecessary character removal unit 431 in the present embodiment.

The unnecessary character removing unit 431 checks the recognition evaluation value at the time of input by handwritten character recognition or the like, and removes the character having the lowest recognition evaluation value from the character string.

The character string search unit 44 searches the word dictionary 45 for a character string having the character string at the beginning based on the character string obtained by the character string conversion unit 43, and if the search is successful, The output unit 46 is notified of the character string as a candidate for the predicted character string. The word dictionary 45 is a dictionary used for prediction processing, and stores various words in both Kanji and hiragana notation.

The output unit 46 outputs the predicted character string candidates retrieved by the character string retrieval unit 44 to the display unit 47.
The display unit 47 is, for example, an LCD (Liquid Crystal Displa).
y), a CRT (Cathode Ray Tube), or the like, and presents the candidate of the predicted character string received from the output unit 46 to the input person.

Next, the operation of the fourth embodiment will be described.

FIG. 8 is a flow chart showing the operation of the prediction process in the fourth embodiment. First, a character string is input by the character string input unit 41 (step 70). In this case, in the fourth embodiment, the character string input unit 41 includes a tablet 411, a recognition unit 412, and a recognition dictionary 413,
The input character string is the result of character recognition of the character string handwritten by the input person.

This input character string (recognition result) is temporarily stored in the character string buffer 42 (step 71). The input character string (recognition result) stored in the character string buffer 42 is displayed on the display unit 47 through the output unit 46.

Here, when the input person instructs the start of the prediction operation by pressing the prediction button 411a provided on the tablet 411 (Yes in step 72), the character string stored in the character string buffer 42 becomes the prediction target character string. And is given to the character string conversion unit 43.

The unnecessary character removing unit 4 in the character string converting unit 43
Reference numeral 31 examines the recognition evaluation value at the time of input by handwritten character recognition or the like, and removes the character having the lowest recognition evaluation value from the search target (steps 73 and 74). The recognition evaluation value is given together with the recognition result from the recognition unit 412 at the time of character recognition.

In this way, when a character having a low recognition evaluation value is removed from the input character string, the character string conversion unit 43 passes the removed character string to the character string search unit 44. The character string search unit 44 searches the word dictionary 45 for a character string having the character string received from the character string conversion unit 43 at the beginning,
If the search is successful, the character string is extracted as a candidate for the predicted character string and is passed to the output unit 46 (step 76).

The output unit 46 passes the predicted character string candidates passed from the character string searching unit 44 to the display unit 47 (step 7).
7). The display unit 47 displays the character string group requested to be displayed on the screen (step 78).

Next, a specific example will be described.

For example, in the case of inputting a character string "Kanji dictionary", the input person intended to input "Kanji", but after the character recognition was mistakenly recognized as "Kanyu", the prediction operation was instructed to start. I shall.

When the character string “Kanyu” (recognition result) is input through the character string input unit 41, the character string is temporarily stored in the character string buffer 42 because it is used as a key for predictive character string retrieval. Is stored. In response to an instruction to start the prediction operation, the character string input unit 41 takes out a character string used in the prediction process, here, “Kanyu”, from the character string buffer 42 and passes it to the character string conversion unit 43.

The unnecessary character removing unit 4 in the character string converting unit 43
Reference numeral 31 examines the recognition evaluation value for each character "Kanyu" and removes the character having the lowest recognition evaluation value from the search target. In this case, the recognition evaluation value of "Kan" is "5",
If the recognition evaluation value of “U” is “1”, the unnecessary character removing unit 431 knows that the recognition evaluation value of “U” is the lowest,
"U" is removed from the search target and "Kan" is passed to the character string search unit 44.

The character string search unit 44 searches the word dictionary 45 for the predicted character string corresponding to the input character string by using "Kan" which is the character string after the removal as a key.

Here, assuming that words such as “Kanji dictionary” in Kanji and “Kanjijisho” in Hiragana are registered in the word dictionary 45, the character string search unit 4
4 outputs the character string of the word from the word dictionary 45 as a candidate of the predicted character string for the input character string, and outputs it to the output unit 46.

In this case, which of the kanji notation and the hiragana notation is used to output the candidate of the predicted character string,
It should match the notation of the input character string. In this example,
The input character string is "Kanyu", which is written in Kanji. Therefore, when outputting the predicted character string, the “kanji dictionary” of kanji is output in accordance with the input notation.

The output unit 46 transfers this "Kanji dictionary" to the display unit 47. The display unit 47 displays "Kanyu" input by the input person.
Instead of, the "Kanji dictionary" is displayed on the screen, and the processing here ends.

By the above processing, the character string "Kanji dictionary" which the input person was trying to input can be obtained,
As a result, it is possible to save the work of inputting the "dictionary" and the trouble of correcting "U" into "character".

In the fourth embodiment, the case where the prediction process is performed on the result of character recognition of the character string input by handwriting has been described. However, for example, after the prediction process, the character string is converted into kana-kanji characters, A configuration in which the result is finally output is also possible.

(Fifth Embodiment) Next, a fifth embodiment of the present invention will be described.

FIG. 7 is a block diagram showing the arrangement of a document creating apparatus according to the fifth embodiment of the present invention. This device is provided with a character string input unit 51, a character string buffer 52, and a character string conversion unit 5.
3, a character string search unit 54, a word dictionary 55, an output unit 56, and a display unit 57.

The character string input unit 51 is for inputting a character string to be predicted, and in the present embodiment, a handwriting input method is assumed and is composed of a tablet 511, a recognition unit 512, and a recognition dictionary 513.

The tablet 511 is an input device for handwriting input of a character string. This tablet 511
A pen for handwriting input is used, and handwriting input of a character string is performed by the writing operation of the pen. Further, here, a prediction button 511a for instructing the start of the prediction operation is provided.

The recognition unit 512 character-recognizes the character string handwritten and input through the tablet 511. The character string (recognition result) obtained by this character recognition is temporarily stored in the character string buffer 52 and then cleared when the input person confirms the character string. However, the prediction button 51
When the start of the prediction operation is instructed by pressing 1a, the recognition unit 512 takes out the character string stored in the character string buffer 52 and passes this to the character string conversion unit 53 as the prediction target character string.

The recognition dictionary 513 is a dictionary used in the recognition unit 512, and stores various kinds of information necessary for character recognition processing, such as information on each stroke forming a character.

The character string buffer 52 is used to temporarily store the character string before being determined. The character string conversion unit 53 performs a process for converting the prediction target character string into a predictable character string, and in the present embodiment, the low-evaluation similar character conversion unit 531 and the similar character search unit 53.
2 and a similar character dictionary 533.

The low-evaluation similar character conversion unit 531 checks the recognition evaluation value at the time of input by handwritten character recognition or the like, and replaces the character having a low recognition evaluation value with the similar character. The similar character search unit 532 refers to the similar character dictionary 533 to search the input character string for a similar character that is likely to be erroneously recognized by character recognition or the like. The character is passed to the low evaluation similar character conversion unit 531. The similar character dictionary 533 has a structure in which similar characters that are easily misrecognized can be searched.

The character string search unit 54 searches the word dictionary 55 for a character string having the character string at the beginning based on the character string obtained by the character string conversion unit 53, and if the search is successful, The output unit 56 is notified of the character string as a candidate for the predicted character string. The word dictionary 55 is a dictionary used for prediction processing, and stores various words in both Kanji and Hiragana notation.

The output unit 56 outputs the predicted character string candidates searched by the character string searching unit 54 to the display unit 57.
The display unit 57 is, for example, an LCD (Liquid Crystal Displa).
y) or a CRT (Cathode Ray Tube) or the like, and presents the predictive character string candidates received from the output unit 56 to the input person.

Next, the operation of the fifth embodiment will be described.

FIG. 10 is a flow chart showing the operation of the prediction process in the fifth embodiment. First, a character string is input by the character string input unit 51 (step 90). In this case, in the fifth embodiment, the character string input unit 51 includes a tablet 511, a recognition unit 512, and a recognition dictionary 513, and the character recognition result of the character string handwritten by the input person is used as the input character string.

This input character string (recognition result) is temporarily stored in the character string buffer 52 (step 91). The input character string (recognition result) stored in the character string buffer 52 is displayed on the display unit 57 through the output unit 56.

Here, when the input person gives an instruction to start the prediction operation by pressing the prediction button 511a provided on the tablet 511 (Yes in step 92), the character string stored in the character string buffer 52 becomes the prediction target character string. And is given to the character string conversion unit 53.

The low evaluation similar character conversion unit 531 in the character string conversion unit 53 checks the recognition evaluation value at the time of input by handwritten character recognition or the like, and passes the character having the lowest recognition evaluation value to the similar character search unit 532. (Steps 93, 94). The recognition evaluation value is given together with the recognition result from the recognition unit 512 at the time of character recognition.

The similar character search unit 532 searches the similar character dictionary 533 for a similar character in the low-rated character provided from the low-evaluated similar character conversion unit 531. , And passes the similar character to the low-evaluation similar character conversion unit 531 (step 95). The low-evaluation similar character converting unit 531 replaces the character with a similar character (step 96).

In this way, when a character having a low recognition evaluation value in the input character string is converted into a similar character, the character string conversion unit 53 passes the converted character string to the character string search unit 54.
The character string search unit 54 searches the word dictionary 55 for a character string having the character string received from the character string conversion unit 53 at the beginning, and if the search is successful, the character string is a candidate for a predicted character string. And outputs it to the output unit 56 (step 97).

The output unit 56 transfers the predicted character string candidates transferred from the character string search unit 54 to the display unit 57 (step 9).
8). The display unit 57 displays the character string group requested to be displayed on the screen (step 99).

Next, a specific example will be described.

For example, in the case of inputting a character string "Nihongoshori", the input person intends to input "Nihongogo", but instructs the start of the prediction operation after being mistakenly recognized as "Nihongogo". I shall.

When a character string "recognition" is input through the character string input unit 51, the character string is temporarily stored in the character string buffer 52 because it is used as a predictive character string search key. Are stored in the memory. In response to an instruction to start the prediction operation, the character string input unit 51 takes out a character string to be used for the prediction process, here, “Nihokugo”, from the character string buffer 52 and passes it to the character string conversion unit 53.

The low-evaluation similar character conversion unit 531 in the character string conversion unit 53 checks the recognition evaluation value for each character of "Nihokugo", and searches for the character having the lowest recognition evaluation value as a similar character. It is passed to the section 532. In this case, the recognition evaluation value of “ni” is “5”, the recognition evaluation value of “ho” is “4”, and the recognition evaluation value of “ho” is “ku”.
When the recognition evaluation value of "1" and the recognition evaluation value of "go" are "6", the low-evaluation similar character conversion unit 531 knows that the recognition evaluation value of "ku" is the lowest, and the similar character search is performed. Part 53
Hand over to 2.

The similar character searching unit 532 searches the similar character dictionary 533 to find out whether or not there is a similar character that is likely to be erroneously recognized by character recognition or the like for "ku". As a result, knowing that "ku" and "n" are similar characters,
This is passed to the low evaluation similar character conversion unit 531. As a result, the low-evaluation similar character conversion unit 531 converts “ku” into “n” to obtain a character string “Nihongo”, and passes this to the character string search unit 54.

The character string search unit 54 searches the word dictionary 55 for a predicted character string corresponding to the input character string by using the converted character string "Japanese" as a key.

If a word such as “Japanese processing” in kanji notation and “nihongogoshiri” in hiragana notation is registered in the word dictionary 55, the character string retrieving unit 54 makes the character of that word registered. The column is extracted from the word dictionary 55 as a candidate for the predicted character string for the input character string, and is output to the output unit 56.

In this case, which of the kanji notation and the hiragana notation is used to output the candidate of the predicted character string,
It should match the notation of the input character string. In this example,
The input character string is "Nihokugo", which is written in Hiragana. Therefore, when the predicted character string is output, the hiragana notation "Nihongoshori" is output in accordance with the notation at the time of input.

The output unit 36 passes this "Japanese language" to the display unit 37. The display unit 37 displays "Nihongo shori" on the screen instead of "Nihogogo" input by the input person, and ends the processing here.

By the above processing, the character string "Japanese language" which the input person was trying to input can be obtained, and as a result, the input work of the portion "SHORI" and "KU" are replaced by "N". You have saved the trouble of correcting to.

In the fifth embodiment, the case where the prediction process is performed on the result of character recognition of the character string input by handwriting has been described. However, for example, after the prediction process, the character string is converted into kana-kanji characters, A configuration in which the result is finally output is also possible.

The present invention is not limited to the above embodiments.

For example, instead of using the prediction button (prediction key) to start the prediction, the prediction process may always be executed by inputting a character string, or the prediction may be performed using some signal. You may start it.

Further, in the search of the predictive character string, the search may be completed by a preset number instead of ending the search when one candidate is successfully searched, or the search may be continued until the search fails. You can

The output destination of the prediction result may be another processing system instead of the screen.

In short, various modifications can be implemented without departing from the spirit of the present invention.

[0166]

As described above, according to the present invention, since the characters in the input character string are converted or excluded from the search target to search the word dictionary, the character string input as the prediction target Even if there is a mistake, the input person can correctly predict the character string that he is about to input. Further, the input person can obtain the predicted character string without correcting the character string input by mistake, and the burden on the input can be reduced.

[Brief description of drawings]

FIG. 1 is a block diagram showing the configuration of a document creation device according to a first embodiment of the present invention.

FIG. 2 is a flowchart showing an operation of a prediction process in the same embodiment.

FIG. 3 is a block diagram showing a configuration of a document creation device according to a second embodiment of the present invention.

FIG. 4 is a flowchart showing an operation of a prediction process in the same embodiment.

FIG. 5 is a block diagram showing the configuration of a document creation device according to a third embodiment of the present invention.

FIG. 6 is a flowchart showing an operation of a prediction process in the same embodiment.

FIG. 7 is a block diagram showing the configuration of a document creation device according to a fourth embodiment of the present invention.

FIG. 8 is a flowchart showing an operation of a prediction process in the same embodiment.

FIG. 9 is a block diagram showing the configuration of a document creation device according to a fifth embodiment of the present invention.

FIG. 10 is a flowchart showing an operation of a prediction process in the same embodiment.

[Explanation of symbols]

 11 ... Character string input unit 12 ... Character string buffer 13 ... Character string conversion unit 14 ... Character string search unit 15 ... Word dictionary 16 ... Output unit 17 ... Display unit 111 ... Keyboard 111a ... Prediction key 112 ... Kana-Kanji conversion unit 113 ... Conversion Dictionary 131 ... Kanji / Kana character conversion unit 132 ... Kanji / Kana character dictionary

Claims (10)

[Claims] 1. A character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, wherein a kanji part in a character string input as a prediction target is Convert to kana characters, and based on the converted character string, search for a character string that has that character string at the beginning as a predicted character string from the word dictionary, and output the searched predicted character string candidates. A character string prediction method characterized in that 2. In a character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, a different character type portion in the character string input as a prediction target. Is removed, based on the removed character string, a character string that has that character string at the beginning is searched as a predicted character string from the word dictionary, and the searched predicted character string candidates are output. A character string prediction method characterized by the above. 3. A character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, similar to a character in a character string input as a prediction target. Character is searched from the similar character dictionary, the corresponding character is converted to a similar character, and based on the converted character string, a character string that has that character string at the beginning is searched as a predicted character string from the word dictionary. A character string prediction method characterized in that the retrieved predicted character string candidates are output. 4. A character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, and recognizing a character in a character string input as a prediction target. Judge the evaluation value, remove the character with a low recognition evaluation value, and search the word dictionary as a predicted character string for a character string that has that character string at the beginning based on the character string after the removal, and then perform that search. A character string prediction method, characterized in that the predicted character string candidates are output. 5. A character string prediction method for predicting a character string having an input character string at the beginning as a character string to be input by an input person, and recognizing a character in a character string input as a prediction target. Judge the evaluation value, search the similar character dictionary for a character similar to the character with a low recognition evaluation value, convert the corresponding character to a similar character, and based on the converted character string, start that character string. A character string prediction method characterized in that a character string included in a part is searched as a predicted character string from a word dictionary, and the searched predicted character string candidates are output. 6. An input unit for inputting a character string to be predicted, a character string conversion unit for converting a Kanji part in the character string input by this input unit into a Kana character, and conversion by this character string conversion unit. Based on the searched character string, a character string search means for searching the word dictionary for a character string having the character string at the beginning as a predicted character string, and a predicted character string candidate searched by this character string search means are output. And a document output device for outputting the document. 7. An input means for inputting a character string to be predicted, a character string conversion means for removing a different character type portion in the character string input by the input means, and a different character type portion by the character string conversion means. A character string search means for searching a character string having the character string at the beginning as a predicted character string from the word dictionary based on the character string with the removed character string, and a predicted character string candidate searched by this character string search means. And a document output device for outputting the document. 8. An input means for inputting a character string to be predicted and a character similar to the character in the character string input by the input means are searched from a similar character dictionary to convert the corresponding character into a similar character. And a character string search means for searching the word dictionary as a predicted character string for a character string having the character string at the beginning based on the character string converted by the character string conversion means, An output device that outputs a candidate for a predicted character string searched by the string searching device. 9. An input unit for inputting a character string to be predicted, and a character string for determining a recognition evaluation value of a character in the character string input by the input unit and removing a character having a low recognition evaluation value. A conversion means and a character string search means for searching a character string having the character string at the beginning as a predicted character string from the word dictionary based on the character string from which the character having a low recognition evaluation value is removed by the conversion means. And a outputting means for outputting a candidate of the predicted character string searched by the character string searching means. 10. An input unit for inputting a character string to be predicted, a recognition evaluation value of a character in the character string input by this input unit, and a character similar to a character having a low recognition evaluation value is determined. A character string conversion unit that searches the similar character dictionary and converts the corresponding character to a similar character, and a character string that has the character string at the beginning based on the character string converted by this character string conversion unit, is a word dictionary. A document creating apparatus comprising: a character string search means for searching as a predicted character string from the output means; and an output means for outputting a predicted character string candidate searched by the character string search means.