JPH0498363A - Knowledge processing system for continuous field - Google Patents

Abstract

PURPOSE:To divide a field on a slip and to process knowledge as to the divided field by connecting the continuous fields through the use of information on the connection of the continuous fields having an OCR definition body. CONSTITUTION:When a connection processing part 11 receives a reading result from a read part 6 which separately reads character strings in the continuous fields 52 and 53, it uses information on the optical character reader(OCR) definition body 9 and identifies that they are the read character strings from the continuous fields 52 and 53. Then, it connects them in a prescribed order, namely, the read character string of the continuous field 53 immediately after the read character string of the continuous field 52. Furthermore, a knowledge processing program 3 executes an address knowledge processing based on a KANA (Japanese syllabary) character string 51' as the knowledge processing and outputs address data as prescribed data. Thus, the field on the slip can be divided and knowledge as to the divided field concerned can be processed.

Description

[Detailed Description of the Invention] [Summary] Concerning a knowledge processing method for a continuous field that performs knowledge processing after combining a plurality of continuous fields divided into a plurality of fields on a form, it is possible to divide fields on a form, and , a reading unit that reads a character string written in a continuation field divided into multiple parts on a form with the purpose of enabling knowledge processing about the divided field, and a knowledge processing about the character string read by the reading unit. and a knowledge processing program that outputs predetermined data by providing a combination processing unit that combines character strings read from the continuation field, the knowledge processing program comprising:
The combination processing unit performs knowledge processing based on the combined character strings and outputs the predetermined data.

[Industrial application field]

The present invention relates to a knowledge processing method for a continuation field, and more particularly to a knowledge processing method for a continuation field that performs knowledge processing after combining a plurality of continuation fields divided on a 2g vote.

In recent years, the work of encoding addresses written on form 2, for example application forms for life insurance, etc., has been increasingly performed by OCR (optical character reader) processing systems. For example, a character string written in kana (katakana) is entered in a kana address field at a predetermined position on a form.
The recognition rate is improved by reading the information using CR and processing it with knowledge.

[Conventional technology]

The position of field (input column) 1 on the form, for example, the address field in kana, is determined in advance. The entries are also predetermined and usually consist of the prefecture, ``name of county or city,'' ``name of town, village, etc.,'' and address. Among these, "prefectural salt" is the first rule, "county name or city ward name"
is the second level, and the name of the town, village, etc. is the third level.

for example 1st level; Kanagawa Ken 2nd level: Kawasakinnakaharaku 3rd level, Kamiodana power Address: 1-1-1 It is.

In the kana address field, these items are entered, for example, as one handwritten character.

Conventionally, as shown in FIG. 6, after reading such a handwritten character string with OCR, corresponding candidate characters are generated,
This was used to access the address dictionary and perform knowledge processing. That is, the address dictionary is accessed using the postal code or the [candidate characters of] level. For example, in the case of a postal code, the corresponding first to third levels are read out from the dictionary, compared with the candidate characters, and outputted if the candidate character is defeated. At this time, kana-kanji conversion is also performed.

Note that the first postal code is entered in the postal code field on the form, read by OCR, and used for the above-mentioned access.

Incidentally, such a field cannot be divided into one part for convenience of knowledge processing. Therefore, for example, the kana address field also needs to be provided as a single (one line) field 1'. Here, field division refers to one field extending over multiple lines on a form, or one field being separated into multiple lines within the same line.

[Problem to be solved by the invention]

According to the above-mentioned conventional technology, since it is only possible to divide fields, there is a problem in that there are large restrictions on the shape of the form and the arrangement of fields on the form Ii.

For example, looking at the kana address field, as mentioned above, it is necessary to fill in all of the first to third levels and the street address.
The field becomes extremely wide.

In this case, the form of the form 5 should be horizontally long so that the kana address field can be placed in one line.

Alternatively, the paper needs to have a width equal to the length of the kana address field. Also, on the rA vote.

By arranging this horizontally long kana address field in one line, the arrangement of other fields is restricted.

Note that if a field, for example, a kana address field, is divided for reasons such as the shape of the form or the arrangement of fields on the form, the following problems will occur.

That is, by splitting the kana address field, 3 second or third
The character string of the level is divided at an unspecified part. Therefore, as it is.

Knowledge processing cannot be performed sufficiently (therefore, in this case, the operator must input all data from the keyboard).

SUMMARY OF THE INVENTION An object of the present invention is to provide a continuous field knowledge processing method that allows fields to be divided on a form and knowledge processing for the divided fields.

[Means for Solving the Problems] FIG. 1 is a diagram showing the principle configuration of the present invention.
An R processing system is shown.

In the first domain, 1 is a processing device, 3 is a knowledge processing program, 5 is a form, 51 is a kana address field, 52 and 53 are continuation fields, 6 is a reading unit 9 is an OCR definition body,
10 is an address dictionary, and 11 is a combination processing section.

Form 5 is a processing target of the OCR processing system, and has a continuation field F 52.53 at a predetermined position. "Continuation field)" 52.53 is a field that is originally one, for example, the kana address field 51, is divided into multiple fields. Therefore, the continuation fields 52 and 53 are filled in as originally one character string. The string that should be
The unspecified part is divided into multiple parts and filled in.

The reading unit 6 reads the information written in the form 5, particularly the character strings written in the continuation fields 52 and 53.

The combination processing unit 11 combines the character strings read from the continuation fields 52 and 53 by the Vf and extraction unit 6 in a predetermined order to form two original character strings.

The knowledge processing program 3 performs predetermined knowledge processing based on the strings combined by the combination processing section 11 and outputs predetermined data.

[For production]

FIG. 2 is an explanatory diagram of the operation of the present invention.

For example, as shown in Figure 1, it is assumed that the continuation field 5253 is obtained by dividing the kana address field 51 into two. In this case, in the second part of the continuation field 5253, a character string in katakana indicating the address is entered by the person filling out the form. At this time, the third level "Kami-odanaka" is divided into "r-kami" and "Odanaka" in the continuation fields 52 and 53.

When the combination processing unit 11 receives the result of reading from the reading unit 6 which has read the character strings in the continuation field 5253 separately, the combination processing unit 11 converts them into the continuation field '52 using the information of the OCR definition body 9, for example. .53. Then, the read character string of the continuation field 53 is combined with the read character string of the continuation field 53 immediately after the read character string of the continuation field 52 after 1 m in a predetermined order. As a result, one kana address character string 51', which should originally have been entered in one kana address field F51, is obtained.

At this time, the previously divided third level "Kamiodanaka" is treated as one third level character string.

The knowledge processing program 3 performs address knowledge processing based on the kana address character string 51' as knowledge processing and outputs address data as predetermined data. That is, kana address character string 51”
The address dictionary 10 is accessed based on the third level (town/village name, etc.) included in the address string 51'', and the address string corresponding to the kana address string 51'' is output.

As described above, according to the present invention, even if one field is divided into a plurality of continuation fields 52 and 53, knowledge processing can be performed after combining these fields. Therefore,
It is possible to eliminate restrictions on the shape of the Ii vote, the arrangement of fields on the @ vote, etc., and it is possible to easily design the @ vote. Furthermore, even if the field is divided, the data obtained is the same as if it were not divided, so the burden on the operator, user program, etc. who use this data can be eliminated.

[Embodiment] FIG. 3 is a block diagram of an embodiment, showing an OCR processing system.

In FIG. 3, 2 is a form processing unit, 4 is an access table, 7 is a display device, and 8 is a data file.

The form processing unit 2 is a central processing unit 1i1! ! The reading section 6 is provided in the processing device 1 consisting of a CPU (CPU) and a memory, and performs various processes on the form 5.
The information read from is sent to the form processing section 2. On the other hand, in order to process the information from the reading section 6, the form processing section 2 reads the OCR definition body 9.

The OCR definition body 9 stores various information for processing the corresponding form 5 in a predetermined format.

This information may be 8, for example, the paper size of form 5, It! It consists of information about the position, size, and entry contents (address, etc.) of various fields (input columns) on the form 5.

The form processing unit 2 reads the information optically read by the reading unit 6 (
The read character string) is interpreted using the OCR definition body 9. For example, according to the OCR definition program 9, the continuation field 52.
53 and also know the location of these continuation fields 52.
53 is obtained by dividing the kana address field 51 into two, and from the above, recognizes that the read character string read at that position is a character string representing an address (kana address character string).

For this purpose, the OCR definition body 9 stores the continuation relationship with other fields for each field on the form 5. The format of the OCR definition body 9 shown in FIG. 3 is an example thereof. That is, for each filtration, corresponding to the field name, information such as the address of the field on the form 5, attribute information of the field, and continuation field name as information regarding the combination of the field. Stored 0 e.g. continuation fields 52 and 53
If the field names are 5A and B, respectively, the continuation field name B for field name A is stored, and a message indicating that there is no continuation field for field name B (END
) information is stored. Note that the fact that the address string is an address character string is stored as fi property information.

When the combination processing section 11 receives the read character strings from the continuation fields 52 and 53 from the form processing section 2.

These are for continuation fields 52,53 and identify those to be combined. Furthermore, the continuation fields 52 and 53 of the OCR definition body 9 are sent from the @vote processing unit 2.
Obtain information about the combination of , and combine the corresponding strings. That is, the read character string of the continuation field 53 is joined immediately after the read character string of the continuation field 52 without placing a blank or the like.

The form processing unit 2 processes read character strings for various fields on the form 5.
Regarding 2.53, the combination processing unit 11 generates candidate character strings from the read character strings subjected to the combination processing, passes these read character strings and candidate character strings to the knowledge processing program 3, and requests knowledge processing.

Upon receiving the read character string and candidate character string, the knowledge processing program 3 performs knowledge processing using various dictionaries.

Now, continuation fields 52 and 53 are kana address field 5.
If 1 is divided, (address) knowledge processing program 3 is divided into 4? Address knowledge processing using the address dictionary 10 is performed on the A character string (kana address character string).

For this purpose, the knowledge processing program 3 creates an access table 4. That is, only the third level of each of the read character string and candidate character string is selectively extracted and the third level of the candidate character string is stored in a form corresponding to the third level of one read character string. At this time, if there are two or more candidate character strings for one read character string, the third level candidate character strings are arranged in descending order of priority. The knowledge processing program 3 is executed in priority order according to the created access table 4.

The address dictionary 10 is accessed using the third level of the candidate character string (third level access is performed).

Then, if there is an address character string corresponding to the third level, it is read out.

For this purpose, the address dictionary 10 stores address strings in a format that allows searches at the third level.

The knowledge processing program 3 creates address data using address character strings obtained from the address dictionary 10 through third level access. That is, first, the second level of the address string,
and the second level of the candidate character string. If there is a match, address data is created using the address string. For example, kanji data is created, and this is associated with kana data to create the kanji and kana data for the first rubel.

These are combined into one address data.

The display bag W7 is for displaying predetermined data to the user. That is, the knowledge processing program 3 displays predetermined data, such as address data, on the display device 7. The user who sees this uses an input device such as a keyboard (not shown)
Predetermined data is corrected by input from.

The data file 8 is a file for storing predetermined data, and is a file owned by a user program that processes the form 5. That is, the knowledge processing program 3 outputs (corrected) predetermined data to the data file 8.

FIG. 4 is a diagram showing a knowledge processing flow for a continuation field, and FIG. 5 is a diagram showing an example of knowledge processing for a continuation field.

Hereinafter, address knowledge processing will be explained according to FIG. 4 with reference to FIG. 5.

(2) The reading unit 6 reads the continuation field 5253 and the like on the form 5 and sends the result to the form processing unit 2.

The continuation fields 52 and 53 are obtained by dividing the kana address field 51. Therefore, among these there are 2
As shown in FIG. 5(A), the second level, third level, and street address of the address are entered.

Further, the third level "Kamiodanaka" is divided into two continuous fields 52 and 53. Furthermore, between the second level and the third level, and between the third level and the address,
A blank space is inserted.

It is used to distinguish between these.

(2) The form processing section 2 reads the OCR definition body 9 and interprets the read character string sent from the reading section. This recognizes that the character strings corresponding to the continuation fields 52 and 53 should be combined (field 5253 is a continuation field) and that it is information about an address.

(2) The combination processing unit 1 combines the read character string corresponding to the 1 continuation field 53 immediately after the read character string corresponding to the continuation field 52 according to instructions from the form processing unit 2. As a result, the third level "
A kana address character string 51' containing "Kamiodana Chikara" as one character string is obtained.

The form processing unit 2 reads this kana address character string (read character string) 5.
Generate a candidate character string corresponding to 1゛.

(2) The knowledge processing program 3 receives read character strings and candidate character strings from the form processing unit 2, and uses them to perform predetermined processing. First, in order to perform third level access, the knowledge processing program 3 creates an access table 4.

That is, the read character string and candidate character string are searched, and the character string "Kamiodanaka" between the first blank and the second blank is extracted as the third level. Then, the third level candidate character strings are arranged in order of priority. Such processing is possible because the third level is not divided.

The knowledge processing program 3 processes the third level of candidate character strings (“
The address dictionary 10 is accessed using the address dictionary 10, and it is checked whether or not a residential address string corresponding to the third level is stored.

■ If the address dictionary 10 contains the corresponding address character string (knowledge processing is successful), the knowledge processing program 3 first performs kana-kanji conversion processing. For example, the corresponding address string is a kana (katakana) string, but the address dictionary 10 stores the corresponding kanji data or character string, and by reading this along with the corresponding address string, the conversion is performed. Perform processing.

The knowledge processing program 3 compares the second level of the address string read from the address dictionary 10 with the second level of the candidate string. Here, the read address string is 1
For example, as shown in Figure 5 (E), "Kanagawa Ken
Kawasakinna Chikara Haraku Chikara Miodana Chikara'. That is,
It consists of the first to third levels. Of these, the second level is
The character string between the first blank and the second blank [is recognized as Kawasakinnakaharaku J]. On the other hand, the second level of the candidate character string is recognized as the character string "Kawasaki Nonaka Haraku" from the beginning to the first blank.

The number of duplicate kana readings at level 2 is about 30 (27 to be exact), and even in this case, there is no overlap between both level 2 and level 3, so 2nd level If there is a match as a result of the comparison, this is adopted as the address character string used to create the output (address data). At this point, the address character string has been corrected to consist of a kana character string and a kanji character string, each consisting of a first to third level and a street address, as shown in FIG. 5(E).

The above processing is also made possible by combining the nine-divided continuation field 5253 (character strings corresponding thereto).

(2) The knowledge processing program 3 generates address data, ie, kanabuta and kanji data, from the kana character string and kanji character string.

(2) The knowledge processing program 3 draws a frame for displaying kanji data, etc. at a predetermined position on the screen of the display device 7, as shown in FIG. 5(F), and displays the kanji data within this frame. This screen is a correction screen for the user to correct the displayed kanji data etc. by inputting from a keyboard or the like. Also, on the correction screen, this frame shows data in one line (kana character strings and kanji character strings are shown in one line each) so that the operator can easily see the data. If you look at it from above, there are two fields (two lines) 5.
2.53 is displayed.

■ The knowledge processing program 3 displays the following information within the frame of the correction screen.
Display ANK (Alpha Numesotsutakana) data. As a result, the correction screen becomes as shown in FIG. 5(F).

Note that even if there is no corresponding address string in the address dictionary 10, the frame is opened on the correction screen 9 and, for example, the read string or candidate string is displayed.

(2) After the modification by the user, the knowledge processing program 3 outputs the address data as predetermined data to the data file 8, for example. At this time, the address data is shown in Figure 5 (E).
As shown in the figure, the record information is output in the form of a record consisting of kana data and kanji data. This data is handled as one data (record) in the data file 8 so that the user program can easily handle it, but when viewed from the form 5, it is divided into two (two lines). This is a combination of the character strings in fields 52 and 53.

〔Effect of the invention〕

As explained above, according to the present invention, the combination processing unit combines the continuation fields using the information regarding the combination of continuation fields possessed by the OCR definition, thereby making it possible to perform knowledge processing on the continuation fields as well. This makes it possible to set up a continuation field on the form,
I! It is possible to eliminate restrictions on the shape of the form and the arrangement of fields on the form.

[Brief explanation of drawings]

FIG. 1 is a diagram illustrating the basic structure of the present invention. FIG. 2 is a diagram illustrating the operation of the present invention. FIG. 3 is a diagram illustrating the configuration of an embodiment. FIG. 4 is a flowchart of knowledge processing for a continuation field. FIG. 5 is a diagram showing an example of knowledge processing for a continuation field. 1 is a processing device, 2 is a form processing unit, 3 is a knowledge processing program, 4 is an access table, and 5 is a form. 51 is a kana address field, 52 and 53 are continuation fields, 6 is a reading section, 7 is a display device, 8 is a pig file, 9 is an OCR definition body, and 10 is an address dictionary. Patent applicant: BFN Co., Ltd. Agent: Patent attorney Hiroshi Morita (2 others) Principle configuration diagram of the present invention Figure 1 Explanation diagram of the action of the present invention Figure 2 Ill processing of continuation field 4th V (＾) Document reading Figure 5 (Part 1) of knowledge processing of character string + candidate character string continuation field Diagram showing an example (G) Figure 5 (Part 2) of knowledge processing of continuation field Diagram showing an example

Claims (3)

[Claims] (1) A reading unit (
6), and a knowledge processing program (6) that performs knowledge processing on the character string read by the reading unit (6) and outputs predetermined data.
3), a combination processing unit (11) for combining character strings read from the continuation field (52, 53) is provided, and the knowledge processing program (3) (11) A knowledge processing method for continuous fields, characterized in that knowledge processing is performed based on the combined character strings and the predetermined data is output. (2) An OCR definition body (9) storing information for processing the form (5) is provided, information regarding the combination of the continuation fields (52, 53) is stored in this, and the combination processing unit (11 ) obtains information regarding the combination of the continuation fields (52, 53) of the OCR definition body (9), and uses this to combine the character strings read from the continuation fields (52, 53). The continuous field knowledge processing method according to claim 1, characterized in that: (3) Provide an address dictionary (10) for storing address strings,
The reading unit (6) reads the continuation field (52, 53)
The combination processing unit (11) reads the character string entered in the kana address field (51) that is
53) are combined to form a kana address string, and the knowledge processing program (3) accesses the address dictionary (10) based on the third level in the kana address string. , an address character string corresponding to the third level is read from the address dictionary (10), and based on this, address data is output as the predetermined data. Knowledge processing method.