JP5580767B2 - Handwritten character input method and handwritten character input device - Google Patents

Description

  The present invention relates to a handwritten character input method and a handwritten character input device.

In recent years, handwritten character input devices such as mobile terminals have kanji as conversion candidates not only when inputting “conference” in kanji, but also when inputting “kaigi” as a reading in hiragana. Some output the column "Conference".
In addition, for example, when the user cannot remember the kanji “meeting” of “meeting” or wants to shorten the input time and inputs “aigi” with kanji katakana, the kanji string “ There is also a handwritten character input device that outputs “conference” (for example, see Patent Document 1 or 2).

In addition, in the handwritten character input device of Patent Document 1, when input with kanji katakana mixed, the katakana portion is once converted into hiragana, and in the above example, “meeting” is converted, and a table for kanji katakana mixed input is created. Refer to and convert “meeting” to “meeting”.
In the handwritten character input device of Patent Document 2, as in Patent Document 1, the katakana part of the character string mixed with kanji katakana is converted to hiragana, the dictionary is referred to, the converted part is converted to kanji, and the kanji string is converted. get.

JP-A-8-202706 JP 7-129564 A

  However, in the handwriting input device disclosed in Patent Document 1, a combination of hiragana and kanji (for each conversion candidate) (converted from katakana) is used as a table for outputting an appropriate conversion candidate when input in a mixture of kanji and katakana. Need to have as many tables as possible. For example, a table having headings “Kaikai” and “Meeting” is required for “Conference” as a conversion candidate. It is unrealistic to prepare a table for inputting all kanji characters that contain all conversion candidates.

  In addition, in the handwriting input device disclosed in Patent Document 2, a user inputs handwritten kana kana characters that are different from reading in a word and reading only in the kanji that constitute a word that the user wants to obtain. In this case, an appropriate conversion candidate cannot be obtained. For example, when the user inputs “handbaki field” in order to obtain “habukino”, the kanji “曳” is “hiki” as the normal reading, so “biki” is registered in the dictionary as the reading. In this case, an appropriate conversion candidate cannot be extracted.

  In view of the above problem, the present invention does not require a special table for inputting kanji mixed with kana characters, and is a kanji that constitutes a word that the user wants to obtain, and reading in the word is different from reading in the kanji alone. The present invention provides a handwritten character input method and a handwritten character input device capable of obtaining appropriate conversion candidates even when a user inputs handwritten kanji characters using kana.

In order to solve the above problems, a first technical means of the present invention, there is provided a handwriting input method handwritten input apparatus acquires a kanji conversion candidate character string of kanji kana input in handwriting, the handwriting A recognition character control unit of the character input device generates a search target character string that is a character string in which the kana part of the character string mixed with the kana-kanji character is changed to a wild card; and a kana-kanji conversion engine of the handwritten character input device , Referring to the dictionary, obtaining as a conversion candidate a word that matches a part other than the wildcard of the search target character string, and the conversion candidate control unit of the handwritten character input device includes the acquired conversion candidates And a step of extracting conversion candidates in which the reading includes the kana of the character string mixed with the kana-kanji characters.

  The second technical means of the present invention is a handwritten character input device for acquiring a kanji conversion candidate for a character string mixed with kanji and kana inputted by hand, and changes the kana part of the character string mixed with kana and kanji to a wild card. A recognized character control unit that generates a search target character string that is a character string, a dictionary, and a kana-kanji conversion engine that acquires, as conversion candidates, a word that matches a part other than the wild card of the search target character string, A conversion candidate control unit that extracts a conversion candidate whose kana kanji mixed character string is included in the reading among the acquired conversion candidates.

  According to a third technical means of the present invention, in the second technical means, the recognized character control unit generates an alternative notation character string obtained by converting a katakana notation portion of the character string mixed with kana and kanji into a character string of hiragana notation. Then, the conversion candidate control unit extracts, as the conversion candidate, those in which the alternative notation character string exists in the reading.

According to a fourth technical means of the present invention, in the third technical means, the conversion candidate control unit is configured such that the character string length of the wildcard part is the same as or longer than the character string length of the alternative notation character string. Only short ones are extracted as the conversion candidates.

  According to the present invention, without using a special table for kanji kana mixed input, kanji that constitutes a word that the user wants to obtain, and the reading in the word and the reading in the kanji alone are different, Even when the user performs handwriting input with a kana, an appropriate conversion candidate can be obtained.

It is a block diagram of an example of the handwritten character input device of this invention. It is a flowchart explaining the process example of the handwritten character input device of FIG.

The handwritten character input device of the present invention will be described with reference to FIG.
The handwritten character input device of the present invention is configured as, for example, a mobile phone device with a handwritten character input function. In the illustrated example, the handwritten character input unit 2, the handwritten character recognition unit 3, and the recognized character control unit 4. A kana-kanji conversion engine 5, a morpheme analysis unit 6, a conversion candidate control unit 7, a conversion candidate display unit 8, and an output character selection unit 9. The handwritten character input device 1 uses a dictionary of the kana-kanji conversion engine 5 when a character string consisting only of hiragana is handwritten and input via the handwritten character input unit 2, and reads a kanji character string that reads the character string. Can be output as conversion candidates. However, since this point is a conventional general operation, a detailed description thereof will be omitted. In the following, a characteristic operation of the handwritten character input device 1, that is, input of a Chinese character string desired by the user will be described. The operation of the handwritten character input device 1 when handwritten input is performed with kana and kana mixed when a specific kanji cannot be remembered or the input time is desired to be shortened, and the processing of each part at that time will be mainly described.

The handwritten character input unit 2 is for a user to input characters, and includes a touch panel device or the like.
The handwritten character recognition unit 3 is a handwriting recognition engine that recognizes the character (figure) input by the handwritten character input unit 2, that is, converts the input character (figure) into a character code that can be processed by the kana-kanji conversion engine 5. When “National Diet Gigi-do” is input to the handwritten character input unit 2 mixed with kanji, the “National Diet Gigi-do” is output to the recognized character control unit 4 as a handwritten recognition character string.

  The recognition character control unit 4 converts only a search target character string that is a character string using a kana portion of a handwritten recognition character string mixed with kanji and kana as a wild card and a katakana notation portion of the recognition character string into a character string of hiragana notation. The alternative notation character string is generated. For example, when the handwritten recognition character string is “National Diet Gigi-do”, “National Diet * Do” is generated as a search target character string, and output to the Kana-Kanji conversion engine 5, “Giji” is generated as an alternative notation character string, and converted. Output to the candidate control unit 7. Note that “*” is a wild card.

  The kana-kanji conversion engine 5 has a dictionary to extract a kanji string from the hiragana character string as a conversion candidate. Based on the dictionary, the kana-kanji conversion engine 5 obtains conversion candidates for the handwritten input character string from the search target character string input from the recognized character control unit 4. For example, when “National Diet * Hall” is input as the search target character string, “National Diet Building”, “National Diet Restaurant”, etc. are acquired as conversion candidates and output to the morphological analysis unit 6.

The morpheme analysis unit 6 acquires a reading of the input conversion candidate. For example, when “National Diet Building” and “National Diet Restaurant” are input as conversion candidates, “Kokakai Gijido” and “Kokakaishokudo” are acquired and output to the conversion candidate control unit 7 together with the above conversion candidates. To do.
The conversion candidate control unit 7 discriminates among the conversion candidates that the alternative notation character string exists in the reading, and outputs it to the conversion candidate display unit 8. For example, when “Kokokai Gijido” and “Kokakaishokudo” are input from the morphological analysis unit 6 as conversion candidate readings, and “Giji” is input from the recognized character control unit 4 as an alternative notation character string, The “National Diet Building” including “Giji” in the reading is output to the conversion candidate display unit 8.

The conversion candidate display unit 8 is a display such as a liquid crystal display device, and displays conversion candidates such as “National Diet Building” input from the conversion candidate display unit 8.
The output character selection unit 9 is configured as, for example, a keyboard device, and is used by the user to select a desired one of the conversion candidates displayed on the conversion candidate display unit 8.

  As described above, in the handwritten input character input device 1, when a character string mixed with kanji characters such as “National Diet Gigi-do” is input, the kana part of the character string is converted into kanji characters, which is appropriate. The narrowed conversion candidates can be extracted and displayed without using a special table.

  FIG. 2 is a flowchart for explaining a processing example of the handwritten character input device 1 of FIG. 1 when “handbaki field” is input by handwriting.

When the handwritten character input unit 1 inputs “feather field” to the handwritten character input unit 2 (step S1), the handwritten character recognition unit 3 acquires the handwritten character string “feather field” (step S2). .
Then, the recognized character control unit 4 generates a search target character string “Han * field” and an alternative notation character string “Biki” from the handwritten recognition character string “Habiki field” mixed with kanji characters (step S3). From the generated search target character string “Han * Field”, the conversion candidates “Haba Wilderness”, “Habano”, and “Habarano” are acquired by the Kana-Kanji conversion engine 5 (step S4).

  In addition, the morphological analysis unit 6 acquires readings “Hanekoya”, “Habikino”, and “Hanekenya” of each acquired conversion candidate (step S5). And among the conversion candidates “Habarano”, “Habanono”, “Habarano”, there is an alternative character string “Biki” in the readings “Hanekoya”, “Habikino”, “Hanegenya” What is to be performed, that is, “Hakuno” is extracted by the conversion candidate control unit 7 (step S6) and displayed on the conversion candidate display unit 8 (step S7).

In addition to the above-described example, the handwritten character input device 1 can extract / display an appropriate conversion candidate even when a kanji / kana mixed character string starting / ending with katakana is input.
The handwritten character input device 1 acquires the handwritten recognition character string “Ifou dignity” by the handwritten character recognizing unit 3 when, for example, “ifif dignified” is input to the handwritten character input unit 2.
Then, the recognized character control unit 4 generates a search target character string “* Dodon” and an alternative notation character string “Ifou” from a handwritten recognition character string “Ifou Dodon” mixed with Kanji characters. From the generated search target character string “* Dojin”, conversion candidates “Dojin”, “Jododo Dodo”, and “Weifeng Dodo” are acquired by the Kana-Kanji conversion engine 5.

  In addition, the morphological analysis unit 6 acquires the acquired conversion candidate readings “How”, “Seisei” and “Ifudo”. And among the conversion candidates “Dozen”, “Jonaido Dodo”, “Weifeng Dodo”, the alternative notation character string “Ifou” exists in the reading “How”, “Seisei Dodo” and “Ifudodo” In other words, “dignified” is acquired by the conversion candidate control unit 7 and displayed on the conversion candidate display unit 8.

  Note that a character string as it is, that is, a handwritten recognition character string, may be output and displayed as a candidate in preparation for intentionally inputting katakana. For example, when “Sai Yin Sin” is written, “Sai Yin Sin” is output as a candidate together with “Sin Bribery Sin”, which corresponds to the case where Kanji characters that are difficult to read are intentionally mixed in Katakana. be able to.

In addition, when the Kana-Kanji conversion engine 5 obtains conversion candidates using the search target character string, the Kana-Kanji conversion engine 5 narrows down the wild card portion “*” that is shorter than the alternative notation character string length and excludes apparently unnecessary candidates. Good. In other words, it is preferable to extract only the character string length of the wild card portion that is the same as or shorter than the character string length of the alternative notation character string. This is because, for example, the kanji notation is usually shorter than the hiragana notation, and in the case of “National Diet Gigi-do”, the “*” part of the search target character string “National Diet * Do” This is because it cannot be longer than the character string length (2 characters).

In the above example, a case has been described in which a kana string mixed with katakana is input. However, a similar operation is performed when a character string mixed with hiragana and kanji (for example, “National Diet Gijudo”) is input. May be.
Note that the morpheme analysis unit 6 does not need a function for determining a paragraph break or part of speech, and includes a reverse conversion function for acquiring a reading instead of a morpheme analysis for analyzing a sentence. That's fine.

DESCRIPTION OF SYMBOLS 1 ... Handwritten character input device, 2 ... Handwritten character input part, 3 ... Handwritten character recognition part, 4 ... Recognition character control part, 5 ... Kana-kanji conversion engine, 6 ... Morphological analysis part, 7 ... Conversion candidate control part, 8 ... Conversion Candidate display part, 9 ... Conversion candidate selection part.

Claims (4)

Handwritten character input device is a hand-written character input method of acquiring kanji conversion candidate character string of kanji kana input in handwriting,
The recognized character control unit of the handwritten character input device generates a search target character string that is a character string obtained by changing the kana part of the character string mixed with kana and kanji into a wild card;
A kana-kanji conversion engine of the handwritten character input device refers to a dictionary, and acquires a word that matches a part other than a wild card of the search target character string as a conversion candidate;
The conversion candidate control unit of the handwritten character input device includes a step of extracting , from the acquired conversion candidates, conversion candidates whose reading includes the kana of the character string mixed with the kana-kanji characters. To input handwritten characters. A handwritten character input device for acquiring a kanji conversion candidate for a character string mixed with handwritten kanji characters,
A recognized character control unit that generates a search target character string that is a character string in which the kana part of the character string mixed with kana and kanji is changed to a wild card;
A kana-kanji conversion engine that refers to a dictionary and obtains a word that matches a part other than the wildcard of the search target character string as a conversion candidate;
A handwritten character input device comprising: a conversion candidate control unit that extracts a conversion candidate whose kana kanji mixed character string is included in the reading among the acquired conversion candidates. The recognized character control unit generates an alternative notation character string obtained by converting a katakana notation part of a character string mixed with kana and kanji into a character string of hiragana notation,
The handwritten character input device according to claim 2, wherein the conversion candidate control unit extracts, as the conversion candidate, one in which the alternative notation character string exists in the reading. The conversion candidate control unit extracts, as the conversion candidate, only a character string length of the wild card portion that is the same as or shorter than a character string length of the alternative notation character string. 3. The handwritten character input device according to 3.

Images