JPH04283852A - Chinese character converting method for japanese syllabary of described address - Google Patents

Abstract

PURPOSE:To convert input KANA(Japanese syllabary) into correct KANJI (Chinese character) data as to only place names of village section level of an address. CONSTITUTION:When there is no same KANA data with input KANA data in a character dictionary file, it is checked whether or not there is another reading applied to KANJI corresponding to the input KANA data; when so, the input KANA data is corrected into KANA data based upon the reading and the corrected KANA data is collated with KANA data in the character dictionary file. When they match each other, KANJI data corresponding to the KANA data is taken out of the character dictionary file.

Description

[Detailed description of the invention]

0001

[Field of Industrial Application] The present invention relates to a method for converting address kana to kanji, and in particular, a method for converting input kana at the character level into kanji correctly when the character level kanji for address (including residence) has multiple readings. Concerning how to convert address kana to kanji.

0002

2. Description of the Related Art Conventionally, for example, when an insurance company or the like writes an address and a name on a form, the address is written in kanji and the name is written in kana or kanji with kana. ANK this form
It is read with an optical character reader (hereinafter referred to as an OCR device) that can recognize handwritten kanji in addition to letters. When converting the read address kanji data into kana data, the conversion device uses the read kanji data and the kanji dictionary stored in the storage device (a kanji and its kana are registered as a pair). Verify the word with the registered kanji in the file. Here, word matching refers to the level of prefecture names (referred to as the first level), the level of place names such as cities and wards (referred to as the second level), and the level of place names such as towns and villages (referred to as the third level). This refers to dividing the place name into character level (corresponding to the place name level above the street address level, i.e., the last place name level of the address)), treating the place name at each level as one word, and performing matching for each word. .

[0003] When a match is found through word matching, the kana data for the registered kanji is extracted from the kanji dictionary file to generate kana data. However, converting the handwritten kanji of an address into correct kana data has a low accuracy rate and takes a long time to match words. On the other hand, when it comes to names written on forms, the handwritten kanji OCR device has a high recognition rate (accuracy rate) of kana data for names written in kana, and the advantage is that it does not take much time to collate. Therefore, in response to the user's request, we decided to have the address on the form written in kana, and a function that generates kanji data from kana data by matching words in the address (
I decided to add the Kana-Kanji conversion of addresses to the conversion device. In this case, since kana is not a complex font like kanji, the read data that is the recognition result of kana from the handwritten kanji OCR device is more accurate than when it is kanji, so the conversion device converts it into kana-kanji and outputs it. The kanji data that was converted was processed as if it had been converted normally.

0004

[Problems to be Solved by the Invention] However, the above-mentioned conventional address kana to kanji conversion method has the following problems. Only when the input kana data read by the handwritten kanji OCR device is exactly the same as the kana data registered in the dictionary file of the storage device, can it be converted to kanji data, and the place name kanji of the address can be read no matter which way it is read. Even when applied (e.g., town → “cho” or “machi”),
Anything different from the kana registered in the dictionary file was treated as an error and could not be converted to kanji data. Note that in the actual dictionary, ``yo'' is capitalized as ``yo'', and the voiced mark is one character, like in a telegram, but in this book, for convenience of explanation, it is written in the usual way. In view of such conventional problems, the purpose of the present invention is limited to place names at the character level of addresses, and if there are multiple readings that apply to the kanji of the place name, the kana data registered in the character dictionary file differs from the kana data registered in the character dictionary file. Correct kana data for the input kana data (the same kana data as the kana data registered in the character dictionary file)
The objective is to provide a method for converting address kana into kanji that can be corrected and converted into correct kanji data.

[0005]

[Means for Solving the Problems] The present invention verifies the input kana data of one word of a place name at the character level of an address with the kana data of a character dictionary file provided in a storage means, and when a match is found, the kana data is In the address kana kanji conversion method in which kanji data for the input kana data is extracted from the character dictionary file, when the same kana data as the input kana data is not in the character dictionary file, the kanji corresponding to the input kana data is used. If there is another reading, the input kana data is corrected to the other reading, and the corrected kana data is word-matched with the kana data in the character dictionary file. When a match is found, the kanji data corresponding to the kana data is extracted from the character dictionary file.

[0006]

[Operation] When the same kana data as the input kana data is not found in the character dictionary file, it is checked whether the input kana data has another reading that corresponds to the kanji, and if there is another reading, the input kana data is changed to that other reading. The corrected kana data is compared with the kana data in the character dictionary file, and when a match is found, the kanji data corresponding to the kana data is extracted from the character dictionary file. Therefore, input kana can be converted into correct kanji data only for place names at the character level of addresses.

[0007]

[Example] Next, an example of the present invention will be described. Figure 3
1 is a block diagram showing an embodiment of an OCR system according to the present invention. In the figure, reference numeral 1 denotes an OCR device (hereinafter referred to as a handwritten kanji OCR device) that can recognize handwritten kanji characters in addition to ANK character recognition. Kanji, kana, etc. are read, character recognition is performed, and the read data (recognition data) is passed to the control unit 2. 2 is memory 2-1
This control unit 2 is connected to a handwritten kanji OCR device 1, a keyboard (referred to as KB) 3 as an input device, a display device (referred to as CRT) 4, and a hard disk (referred to as HD) 5. ing. Control part 2
is for controlling the handwritten kanji OCR device 1, CRT 4, hard disk 5, etc. in an integrated manner.

[0008] Here, the CRT 4 is used to display input data, display data read from the handwritten kanji OCR device 1, and so on. In addition, the hard disk 5 has dictionary files, particularly dictionary files for checking words of place names in addresses (kana of place names and kanji of the kana are registered).
etc. are stored.

[0009] The dictionary file for word matching of place names in addresses includes a file at the prefecture name level (referred to as the first level file) and a file at the place name level such as ward, city, county, etc. (second level file).
It's called a level file. ) and place name level files (referred to as third level files or character level files) such as towns and villages. The second level files are provided classified by each prefecture of the first level. The third level file is further classified into each city, ward, and county corresponding to the second level. The present invention relates to a (character level file), and in particular, this (character level file) is hereinafter referred to as a character dictionary file. Further, reference numeral 6 denotes a conversion device, and this conversion device 6 converts input kana data into kanji data, converts kanji data into kana data, and the like. The conversion device 6 includes a control section 2, a keyboard 3, a CRT 4, a hard disk 5, and the like. In addition, 7 reads a document as a reading medium, and compares the read data (recognition data) with the data in the dictionary file. This is an OCR device that converts images into kana data. This OCR
The device 7 is also capable of comparing input data from the keyboard 3 with data in a dictionary file, and converting the input data into kana data, kanji data, or the like.

Next, an outline of the word matching process according to the present invention will be explained with reference to FIG. Note that FIG. 2 is an explanatory diagram illustrating an overview of word matching processing according to the present invention. A form 11 (in which the address is written in kana as shown) as a reading medium is input into the handwritten kanji OCR device 1. Or input the address using the keyboard 3 as shown in the figure. The control unit 2 receives read data from the handwritten kanji OCR device 1 and input kana data from the keyboard 3.
The data is stored in the memory 2-1. Next, the control unit 2 performs word matching (conversion of address kana to kanji). This word matching will be explained below. Since the kana data is separated by spaces, the control unit 2 takes out the kana data separated by spaces from the memory 2-1 as one word. In the illustrated example,
``Tokyo'', ``Minatoku'', and ``Shibaura'' each constitute one unit. Next, the control unit 2 sequentially extracts one word at a time from the first level to the third level of the place name of the address and performs verification. In the illustrated example, the control unit 2 first retrieves "Tokyo" from the memory 2-1, matches this with the prefecture name in the first level file of the dictionary file for word matching on the hard disk 5, and if they match, The kanji character "Tokyo" for "Tokyo" is extracted from the first level file and stored in the memory 2-1. Next, the control section 2
retrieves "Minatoku" from memory 2-1, checks this "Minatoku" with the ward name, city name, etc. in the second level file, and if there is a match, converts the kanji "Minato-ku" from "Minatoku". The data is extracted from the second level file and stored in the memory 2-1. Furthermore, the control unit 2 retrieves "Shibaura" from the memory 2-1, compares this "Shibaura" with the town name, village name, etc. in the character dictionary file, and if they match, the kanji "Shibaura" of the "Shibaura" is used. ” is taken out and stored in the memory 2-1. In this way, it is converted into the kanji for the address "Shibaura, Minato-ku, Tokyo."

Next, an embodiment of the present invention will be explained using the flowchart shown in FIG. Note that FIG. 1 is a flowchart showing one embodiment of the present invention. Further, FIG. 1 will be explained with reference to FIG. 4 as a specific example. FIG. 4 is an explanatory diagram showing an example of correction of kana data. First, kana data is input (step S1). Here, a form (with the address and place name written in kana) as a reading medium is transferred to the handwritten kanji OCR device 1.
Enter. Alternatively, the operator inputs the address using the keyboard 3 in kana characters. The kana data (recognition data) read by the handwritten kanji OCR device 1 is supplied to the control section 2 . The control unit 2 stores this kana data in the memory 2-1. Furthermore, kana input from the keyboard 3 is also stored in the memory 2-1 of the control section 2.

Here, the kana data stored in the memory 2-1 includes the first level place name and the second level place name.
Place names at the second level and place names at the third level (character level) are separated by spaces, and kana separated by spaces are treated as one word. Figure 4
In the example, there is a space between the second level place name and the third level place name. The control unit 2 extracts one word of kana data from the kana data of the address stored in the memory 2-1 (step S2). Note that the kana data of one word taken out from the memory 2-1 is referred to as input kana data. here,
When taking out one word of kana data from the memory 2-1, it is assumed that the data is taken out in order from the first level to the third level. Therefore, if there is no first level place name, the control unit 2 retrieves the second level place name from the memory 2-1.

Next, the control unit 2 checks whether the input kana is at the character level (step S3). If it is not at the character level, the conventional processing is performed (step S4). That is, if the control unit 2 is not at the character level, it first compares the input kana with the kana in the first level file on the hard disk 5, and if there is a match, it selects the kanji (prefecture name) for that kana from the first level file. Take it out. If the input kana and the kana in the first level file do not match, the control unit 2 performs word matching between the input kana and the kana in the second level file, and if they match, the kanji (place name of city, ward, etc.) of the kana is displayed. Extract from the second level file. In addition, in the example in Figure 4, "Kawagoeshi"
corresponds to this case, but the explanation is omitted in FIG. 4 because it is not related to the invention. The control unit 2 controls the hard disk 5
The kanji data of the place name extracted is stored in the memory 2-1. Furthermore, in reality, the number of place names at the first level and the second level is smaller than the number of place names at the character level, and generally no problem occurs. When the conventional processing is completed in step 4, the kana data for the next word is retrieved from the memory 2-1 (steps S5 and S6), and if it is not at the character level again, the conventional processing is performed in the same way as described above. (Step S4).

Next, the control unit 2 retrieves one word of kana data from the memory 2-1 (steps S5 and S6). If the input kana is at the character level, the control unit 2 performs word matching between the input kana and the kana in the character dictionary file on the hard disk 5, and if the word matching results in a match (the same kana as the input kana is in the character dictionary file). If there is one), the kanji data for that kana is extracted from the character dictionary file and kanji data is generated (steps S7, S13). This kanji data is stored in the memory 2-1. If in step S7, the input kana does not match the kana in the character dictionary file (if the same kana as the input kana does not exist in the character dictionary file),
The control unit 2 checks the dictionary file on the hard disk 5 to see if there is another commonly used reading (character) for the kanji corresponding to the input kana (steps S8 and S9), and if there is no other commonly used reading (letter), it determines as an error and does not generate kanji data. . In addition, the control unit 2
If the kanji corresponding to the input kana has another commonly used reading (character), correct the input kana to a character with another reading, and then match the corrected kana with the kana in the character dictionary file again. (matching) (steps S10, S
11). If there is a match as a result of word matching (if the corrected kana is in the character dictionary file), the control unit 2 extracts the kanji data for that kana from the character dictionary file and generates kanji data (steps S12, S13). . If the same kana as the corrected kana is not found in the character dictionary file, it will be treated as an error and the kanji data will not be generated.

In the example shown in FIG. 4, since "saiwaimachi" is not in the character dictionary file, a word matching program installed separately on the hard disk 5 is used to check whether there is another commonly used reading (letter) for the kanji corresponding to "machi", for example. ``Machi'' is searched from the table created in ``Cho'', and the corrected kana ``Saiwaicho'' is compared with the character dictionary file (steps S7 to S11). As a result of the comparison, since the corrected kana was found in the character dictionary file, the kanji data "Saiwai-cho" for that kana is extracted from the character dictionary file and the kanji data "Saiwai-cho" is generated as shown in the figure (step S12 , S13). As can be seen from the above explanation,
For place names at the character level of addresses, if there are multiple readings (characters) that can be applied to the kanji of the place name, the correct kana data (characters) will be displayed for input kana data that is different from the kana data registered in the character dictionary file. It is possible to correct the correct kana data (the same kana data as the kana data registered in the dictionary file) and extract the kanji data for the correct kana data from the character dictionary file. Therefore, even if the input kana is different from the kana in the character dictionary file, the input kana is corrected to the correct kana,
It can be converted into correct kanji data. The present invention is not limited to this embodiment, and various applications and modifications can be made without departing from the gist of the present invention.

[0016]

Effects of the Invention As described above, according to the present invention, only for a place name at the character level of an address, if there are multiple readings (characters) applicable to the kanji of the place name, the kana data registered in the character dictionary file This has effects such as being able to correct input kana data that is different from the original kana to the correct kana and converting it to correct kanji data.

[Brief explanation of the drawing]

FIG. 1 is a flowchart showing one embodiment of the present invention.

FIG. 2 is an explanatory diagram of an overview of word matching processing according to the present invention.

FIG. 3 is a block diagram showing an embodiment of an OCR device according to the present invention.

FIG. 4 is an explanatory diagram showing an example of kana data correction.

[Explanation of symbols]

1 Handwritten kanji OCR device 2 Control section 2-1 Memory 3 Keyboard 4 CRT 5 Hard disk 6 Conversion device 7 OCR system

Claims (1)

[Claims] Claim 1: The input kana data of one word of a place name at the character level of the address is word-checked with the kana data in the character dictionary file provided in the storage means, and when there is a match, the kanji data for the kana data is transferred to the character In a kanji conversion method for address kana that is extracted from a dictionary file, when the same kana data as the input kana data is not in the character dictionary file, it is determined whether there is another reading that can be used for the kanji corresponding to the input kana data. If there is a different reading, modify the input kana data to that other reading, then match the corrected kana data with the kana data in the character dictionary file. A method for converting address kana to kanji, characterized in that kanji data for kana data is extracted from the character dictionary file.