JPH08241314A - Document filing system - Google Patents

Abstract

PURPOSE: To provide a document filing system which reads out information entered into a document by utilizing features of the object document to be recognized and efficiently perform input and itemizing operation. CONSTITUTION: This document filing system which reads out characters having information on plural items entered into a document according to a certain format, converts them into character codes, and stores them is equipped with a character recognizing means 111 which recognizes the characters on the basis of dot patterns showing characters included in an image corresponding to the document and outputs corresponding character codes as recognition results, a decomposing means 112 which decomposes text information obtained as the recognition results into parts of speech as its constituent elements, and a classifying means 113 which analyzes the relativity of a series of parts of speech obtained by the decomposing means 112, classifying information represented with the series of parts of speech by items, and stores it.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a document filing system to which a character recognition device for recognizing characters from an image of a document is applied, and in particular, it is expected that the recognition rate by the character recognition device such as family register data is low. The present invention relates to a document filing system corresponding to a document to be created.

A character recognition device encodes the content of a document by collating a character pattern included in an image of an optically read document with a built-in character pattern, and is suitable for horizontal writing. Many products have been commercialized for documents in which typeface characters are arranged at intervals. Therefore, by applying such a character recognition device to a document filing system, various materials and reports printed using print, including books,
The contents of a huge amount of documents in the office can be encoded and saved as a compact size file.
In addition, information can be shared because searches are easy.

By the way, in recent years, with the spread of word processors and the like, the ratio of documents printed in print is overwhelming, but for example, handwritten documents such as original family registers managed by local governments nationwide. There is a considerable amount of documents in which handwritten parts and typed parts depending on the type are mixed, and there is an urgent need to store these materials in a coded form.

[0004] In particular, since the original family register is required to be preserved for 80 years after the removal of all persons listed in the family register, the ratio of the original family register organized before the introduction of the type is considerably large, When encoding and storing family register data, it is absolutely necessary to consider the existence of handwritten characters.

[0005]

2. Description of the Related Art As described above, a conventional character recognition apparatus corresponds to a document in which characters of a typeface are arranged at a constant interval, and a template with ruled lines is vertically extended by a writing brush. It is very difficult to recognize each character from the image of the document in which the characters are described in the state of being jammed.

For this reason, conventionally, as shown in FIG. 8, a microfilm 301, which is an original copy of a family register, is placed on a microfilm reader 302 to print an image of the original family register on paper to create a copy 303. Based on this, the operator reads the information described in the original family register, and inputs the reading result through the input device such as the keyboard 311 provided in the document filing system 310.

Further, in response to the input of the reading result,
The edit processing unit 312 extracts information of each item described in the family register to create a family register data file 313, and in order to confirm the information of each of these items, the collation list creation unit 314 separates each item. The described collation list 304 was printed out, and the collation list 304 and the above-mentioned copy 303 were collated manually.

At this time, when an error is found, the terminal operator again operates the keyboard 311 to correct the corresponding portion, and when the error is not found by the above-described collation processing, each item is not displayed for the first time. The code information corresponding to is stored in the family register data file 313.
Here, in order to support the work of reading the information described in the original family register, a work (hereinafter referred to as a marking work) in which a mark indicating a portion corresponding to a necessary item is preliminarily applied to the copy 303 may be performed. is there.

In this marking operation, if the division of each item is clearly specified, the above-mentioned edit processing unit 312
Can receive information divided for each item, so it is only necessary to determine whether each item of information is suitable for the item and reflect the determination result in the editing process. On the other hand, when the marking work indicates the range to be input as the necessary information, the edit processing unit 312 needs to perform a process of extracting a portion corresponding to each item from the input information. The labor required for can be reduced significantly.

As described above, conventionally, the marking work and the processing by the editing processing unit 312 have been used to slightly improve the efficiency of the work for inputting family register information.
The information reading work from 3, the input work, and the work of confirming the result of this work are all performed manually, and no attempt has been made to automate these works.

[0011]

As described above, if all the information reading work, the input work and the collation work are processed manually, the operator's burden is too great, and therefore, the reading error or the reading work is not performed. There is a high possibility that many mistakes will be triggered at various stages, such as simple typos when inputting information, and check mistakes during collation work.

In addition, in order to digitize all of the enormous family register originals by the above-mentioned method relying on manpower, enormous manpower is required, which requires astronomical cost. Therefore, in order to file the family register information, the information reading and input work should be automated and
It is necessary to support the collation work.

By the way, by the approach specialized for a special document in which a character string is arranged vertically in a template having ruled lines, each character is recognized to some extent from image data corresponding to a document such as an original family register. With the aim of recognizing by rate, it became possible to automate the work of encoding each character contained in such a document.

However, if the information is simply coded in this way, information that is unnecessary as information for each item such as the item name will also be coded. Therefore, the information of each item is extracted from the coded text information. It is necessary to devise the processing. In addition, handwritten characters may cause recognition omissions at a high frequency, so it is necessary to consider the recognition omissions by the character recognition device.

The present invention supports the document filing system capable of automatically performing the information reading process by utilizing the characteristics of the document and the collation process between the coded information and the original document. The purpose is to provide a document filing system capable of performing.

[0016]

FIG. 1 is a block diagram showing the principle of the document filing system according to claims 1, 5, and 6. According to the invention of claim 1, in a document filing system for reading a character in which information on a plurality of items is described in a manuscript according to a certain format, converting the character into a character code and storing the character code, an image corresponding to the manuscript is included. A character recognizing means 111 for recognizing each character based on a dot pattern representing the existing character and outputting a corresponding character code as a recognition result, and a decomposing means for decomposing the text information obtained as a recognition result into its parts of speech. The classifying unit 1 that analyzes the relationship between 112 and the series of parts of speech obtained by the disassembling unit 112, classifies information represented by the series of parts of speech for each of a plurality of items, and provides the storage processing.
And 13 are provided.

FIG. 2 is a principle block diagram of the document filing system of claims 2 to 4. Claim 2
In a document filing system that reads characters written in a manuscript, converts them into character codes and saves them, the invention recognizes each character based on a dot pattern that represents the characters included in the image corresponding to the manuscript. , Character recognition means 11 for outputting the corresponding character code as a recognition result
1 and the decomposing means 112 for decomposing the text information obtained as a recognition result into its parts of speech, and the series of parts of speech obtained by the decomposing means 112 are analyzed to obtain information represented by characters. And a correction unit 115 for correcting the text information of the recognition result on the basis of the detection result by the mismatch detection unit 114 and for the storage processing. .

According to the invention of claim 3, in the document filing system according to claim 1 or 2, the recognition result of the dot pattern which cannot be recognized by the character recognition means 111 is complemented and sent to the decomposition means 112. The input means 12 for inputting a candidate character corresponding to an unrecognized dot pattern.
1 and a character pattern representing a candidate character is arranged in an area adjacent to an unspecified dot pattern in the image corresponding to the original, and is combined with the image corresponding to the original.
Corresponding to the candidate character input via the input unit 121 in response to the input of the confirmation instruction, and the first display unit 123 for displaying the image obtained by the first combination unit 122. It is characterized in that it is provided with a confirming means 124 for confirming as a dot pattern recognition result and sending it to the disassembling means 112.

The invention of claim 4 is the document filing system according to claim 3, wherein the first synthesizing means 122 is used.
Is an image replacement unit 126 that replaces the image information of the area adjacent to the unrecognized dot pattern with the character pattern in response to the input of the character pattern indicating the candidate character, and the candidate character according to the input of the instruction to specify the typeface. Is converted into a designated typeface and is sent to the image replacing means 126.

According to a fifth aspect of the present invention, in the document filing system according to the first or second aspect, the character recognition means 111 stores a pattern dictionary 131 which stores a character pattern representing a corresponding character corresponding to a character code. According to the input of a dot pattern representing each character included in the image corresponding to the manuscript, each of the character patterns stored in the pattern dictionary 131 is matched, and the character code corresponding to the character pattern matching the dot pattern is generated. The configuration is provided with a collating unit 132 for outputting as a recognition result, and a registering unit 133 for giving a new character code to a character pattern corresponding to a dot pattern and registering it in a pattern dictionary in accordance with the collating result by the collating unit 132. It is characterized by

According to a sixth aspect of the invention, in the document filing system according to the first aspect, based on the position occupied by the information of each classified item in the text information of the recognition result, each information corresponds to the original document. Position specifying means 141 for specifying the dot pattern range as an item area
A second synthesizing means 142 for synthesizing a character pattern representing the corresponding information in an area adjacent to each item area of the image data corresponding to the original, and a second synthesizing means 1.
And a second display unit 143 for displaying the image data obtained at 42.

[0022]

According to the invention of claim 1, the character string recognized by the character recognizing means 111 is decomposed into morphemes by the disassembling means 112 and subjected to the classification processing of the classifying means 113.
It is possible to automate the work of reading the information of each item described in the manuscript and the itemization work.

According to the second aspect of the present invention, the recognition result by the character recognition means 111 is decomposed into a series of morphemes by the decomposition means 112, and is provided to the detection processing of the mismatch detection means 114. According to the detection result, the correction means 115 is provided. By operating the, the recognition result can be corrected in consideration of the consistency as information. The invention of claim 3 is the complementing means 120.
By the operation of the first combining means 122 of
Since the candidate character and the unrecognized dot pattern input via the can be displayed side by side on the display screen of the first display unit 123, the candidate character and the unrecognized dot pattern can be sufficiently compared and recognized. The results can be modified. In addition, when it is determined that the input candidate character is appropriate as the recognition result of the above-described unrecognized dot pattern, the operator inputs a confirmation instruction, and the confirmation means 124 operates in response to the confirmation instruction. The character code indicating the selected candidate character is decomposed as the corresponding recognition result 112.
And the recognition result of the character recognition means 111 can be complemented with the operator's judgment.

According to a fourth aspect of the present invention, the conversion means 125 converts the candidate characters into a character pattern of a designated typeface and provides them to the processing of the image replacement means 126, so that the candidate characters are recognized as the recognition target described in the manuscript. It can be displayed on the display unit 123 using a typeface similar to the character. According to the invention of claim 5, in the character recognition means 111, the matching means 1
In accordance with the collation result by 32, the registration means 133 creates a character pattern corresponding to the new character code in the pattern dictionary 13
By registering in 1, the new character pattern can be used for the character recognition process thereafter. As a result, the recognition rate of the character recognition unit 111 can be improved by flexibly coping with so-called “letters” that frequently appear in personal names and the like.

According to the sixth aspect of the present invention, the second synthesizing means 142 operates in response to the instruction from the position designating means 141 to correspond to the image corresponding to the document and each item obtained by the classifying means 113. The information to be combined with the second
Since they can be displayed side by side on the display screen of the display means 143, it is possible to proceed with the collation work while sufficiently comparing the information of each item and the image representing the characters written in the manuscript.

[0026]

Embodiments of the present invention will now be described in detail with reference to the drawings. FIG. 3 is a block diagram of an embodiment of a family register information filing system to which the inventions of claims 1 and 3 are applied. In FIG. 3, the microfilm reader 302 sends the image data corresponding to this image to the character recognition device 210 via the image buffer 201, instead of printing the image of the microfilm obtained by photographing the original family register on paper.

In response to this, the area extraction unit 211 of the character recognition device 210 cuts out an area corresponding to each character from the image data held in the image buffer 201,
The pattern matching unit 212 is configured to recognize the character indicated by the dot pattern in each area by matching the dot pattern in each area with the character pattern in the pattern dictionary 213.

The pattern matching unit 212, as the matching result of the dot patterns of each area, if the matching rate of the dot pattern and the character pattern is equal to or more than a predetermined threshold value, as the recognition result of the dot pattern of the area. ,
A character code corresponding to the character indicated by the corresponding character pattern may be output, and if the matching rate is less than or equal to a predetermined threshold value, it may be output that the recognition result is undetermined.

Here, the above-mentioned pattern dictionary 213 is
In addition to character patterns corresponding to standard typefaces such as Mincho typeface used in normal type printing, it also has a character pattern of a writing brush.Furthermore, for each typeface, not only the corresponding kanji but personal name kanji and old kanji It also has a character pattern. The recognition complementation processing unit 220 corresponding to the complementing unit 120 operates according to the input of the recognition result for each area obtained in this way.

In the recognition complement processing section 220 shown in FIG. 3, the code holding section 221 holds the code indicating the recognition result by the character recognition device 210, and sends it to the image synthesizing section 223 via the pattern converting section 222. Then, the image synthesizing unit 223 synthesizes the image data in the image buffer 201 and a series of character patterns representing the recognition result received from the pattern converting unit 222, and provides the display device 202 with a display operation. There is.

The image synthesizing section 223 corresponds to the first synthesizing means 122. First, as shown in FIG. 4 (a), the regions of the images corresponding to the respective characters described in the original family register are respectively displayed. A character pattern representing a character obtained as a recognition result or a mark indicating indetermination (indicated by a symbol? In FIG. 4) may be synthesized in the area adjacent to.

Further, in FIG. 3, the candidate input unit 224
Is a character code corresponding to the input means 121, and is a character code indicating a candidate of a character corresponding to the dot pattern of the undetermined area in response to an instruction from the operator input via a keyboard (not shown) or the like. Is input and the corresponding character code of the code holding unit 221 is rewritten via the replacement processing unit 225.

In this case, for example, when the candidate character “edit” is input through the candidate input unit 224 for the dot pattern indicated by the symbol in FIG. 4A, the replacement processing unit 224
By this, the corresponding recognition result of the code holding unit 221 (“?” Indicating undetermined in this case) is replaced with the character code indicating the candidate character “edit”, and accordingly, the pattern conversion unit 222 causes the candidate character to be changed. A character pattern representing a “composition” is obtained and is used in the combining process by the image combining unit 223. As a result, as shown in FIG. 4B, the dot pattern in the corresponding area is rewritten with the character pattern representing the candidate character “edit”.

In this way, when the replacement processing unit 225 replaces the contents of the code holding unit 221, and the replacement of the dot patterns of all the undefined areas is completed, the contents held by the code holding unit 221. Is output as the recognition result, the function of the determining means 124 described in claim 3 is realized, the character corresponding to the undetermined dot pattern is confirmed, and the recognition result by the character recognition device 210 can be complemented. .

In this case, since the undetermined dot pattern and the character pattern corresponding to the candidate character are displayed side by side by the image synthesizing unit 223, the operator recognizes the dot pattern to be recognized and the character pattern of the candidate character. Can be compared sufficiently, and the coincidence / non-coincidence of two dot patterns can be intuitively and accurately determined.
Thereby, like the handwritten character, the character recognition device 210
It can flexibly support characters whose recognition rate tends to be low and can support accurate character recognition, and converts documents containing handwritten characters such as the original family register into character codes,
It is possible to obtain text information indicating the content of the manuscript.

The text information thus obtained is
The analysis processing unit 230 corresponding to the disassembling unit 112 and the classifying unit 113 is provided with the processing. In the analysis processing unit 230 shown in FIG. 3, the decomposition processing unit 231 realizes the function of the decomposition unit 112 by decomposing the input text information into morphemes based on the morpheme dictionary 232.
The series of morphemes is analyzed by the syntactic analysis unit 233 and the semantic analysis unit 23.
It is configured to be provided for the processing of No. 4.

The morpheme dictionary 232 described above has areas corresponding to the types of information contained in the family register, and is configured to store the corresponding morphemes. In this morpheme dictionary 232, for example, a place name such as a municipality name may be stored in the address area, and a family name and a first name may be stored separately in the name area. In addition, the analysis processing unit 230 includes a syntax rule holding unit 235 that holds a rule relating to the connection of each morpheme in the sentence written in the family register, and the syntax analysis unit 233 is held in this syntax rule holding unit 235. The configuration is such that the connection of a series of morphemes obtained by the decomposition processing unit 231 is analyzed with reference to the rule.

At this time, the syntactic analysis unit 233 may associate a series of morphemes with each item according to syntactic rules. For example, the six morphemes “Tokyo”, “Marunouchi”, “One”, “Chome”, “One”, and “No.” obtained by decomposing the recognition result of the original family register shown in FIG. The item name “main domicile” may be associated with each other as a group of morphemes to be represented. Similarly, various item names such as “name” and “composition date” may be associated with a group of corresponding morphemes.

On the other hand, the semantic analysis unit 234 is configured to analyze the meaning of a set of morphemes associated with each item and determine whether each meaning matches the corresponding item. The semantic analysis unit 234 holds information about the range of information associated with each item, and for example, the date and item name “indicated by a set of morphemes corresponding to the item name“ composition date ””. The consistency of the relevant information may be determined by comparing the date range corresponding to the “organization date”.

Further, the analysis control unit 23 of the analysis processing unit 230
6 is stored in the family register data file 203 by combining each item and a group of morphemes when the semantic analysis unit 234 determines that each item and corresponding information match each other, and the output processing unit 240 also stores the combination. It is configured to be sent out and output for the collation work. On the other hand, when the syntax analysis unit 233 or the semantic analysis unit 234 detects an inconsistency, the analysis control unit 236 activates the correction processing unit 237, and the correction processing unit 237 corrects the correction required by the operator. The instruction may be received, the morpheme delimiter position or the original text information itself may be corrected, and the morpheme delimiter position and the original text information may be processed again by the syntactic analysis unit 233 and the semantic analysis unit 234.

As described above, the syntactic analysis unit 233, the semantic analysis unit 234, and the correction processing unit 237 operate according to the instruction from the analysis control unit 236, so that the classification unit 11
By implementing the function of 3, the information of each item included in the text information obtained as the recognition result by the character recognition device can be automatically classified and stored as coded information.

As a result, the reading work of the family register information and the information input work can be automated, and the conventional manual marking work can be omitted to greatly reduce the burden on the operator. Further, by combining the processing of the analysis processing unit 230 with the processing of the character recognition device 210 and the recognition complementation processing unit 220 described above, most of the manual information input work is omitted and the burden on the operator is greatly reduced. It is possible.

That is, if the invention of claim 1 is applied,
It is possible to omit the manual marking work and information input work, automate the conversion process of family register data, and use it for the final matching work. For example, together with the contents of the family register data file 203 described above, the image buffer 201
The image data in the image is sent to the output processing unit 240 described above, and the output processing unit 240 arranges the image of the original family register represented by this image data and the information classified into items in the new family register format. It is sufficient to print out the registered family register side by side.

Thus, it is possible to perform the collation work while comparing the original family register and the new format family register on the same sheet. By the way, the recognition complement processing unit 22
Despite the correction processing by 0, an error may remain in the text information of the recognition result. Such a recognition error is
Recognition result of analysis by the analysis processing unit 230
It is possible to solve by feeding back to 0.

FIG. 5 shows a block diagram of an embodiment of a family register information filing system to which the invention of claim 2 is applied. In this case, the decomposition processing unit 231 of the analysis processing unit 230 responds to the input of the character string that is not registered in the morpheme dictionary 232.
This character string may be temporarily decomposed as a proper noun and provided to the processing of the syntactic analysis unit 233 and the semantic analysis unit 234 together with other decomposition results.

Further, the syntactic analysis unit 233 classifies the temporarily decomposed proper noun into appropriate items based on the relation between the preceding and following morphemes and the syntactic rule. , It is determined whether or not the item in which the proper noun is classified is an item in which the proper noun is allowed,
It is sufficient to detect the inconsistency between the item and its contents. For example,
Proper nouns are not allowed for items with place names such as permanent domiciles or reporting locations or items with dates, so the provisional proper nouns described above were included in the information corresponding to these items. If this is detected as an inconsistency,
In response to this, the analysis control unit 236 causes the correction processing unit 237 to
Instead of, the recognition complement processing unit 220 may be activated to instruct the retry of the correction processing of the recognition result.

Here, the information including the character string which is erroneously recognized by the processing of the above-mentioned syntax analysis unit 233 is
Since the items are classified into appropriate items, the analysis control unit 236
With the retry instruction of the correction process, the information indicating the reason for the inconsistency with the location where the inconsistency is detected may be notified to the effect that it does not match the information range of the corresponding item.
In response to the input of the retry instruction described above, the image synthesis unit 2
23, again, the image corresponding to the original family register and the character pattern representing the recognition result held in the code holding unit 221 may be combined and displayed via the display device 202.

At this time, the image composition section 223
In the image synthesized as described above, highlights the portion of the recognition result that is inconsistent and the dot pattern corresponding to this portion to indicate the portion that needs to be corrected, and Display data representing information indicating the reason may be created and sent to the display device 202, respectively.

As described above, the functions of the correction means 115 are realized by the operation of the respective parts of the recognition complementation processing part 220 in accordance with the instruction from the analysis control part 236, and the mismatching is caused by the syntactic analysis process and the semantic analysis process. It is possible to correct the recognition result of the location where is detected. In this case, the recognition result of the part where the inconsistency is detected is treated in the same way as the part where the recognition result is undetermined, and the part where the character recognition result is likely to be erroneous is recognized along with the recognition results before and after it. It is displayed in the area adjacent to the corresponding dot pattern, and the reason for the mismatch is also displayed.

Therefore, the operator uses the range of the type of information that matches the item indicated by the corresponding inconsistency reason, the image data of the corresponding area, and the recognition results before and after the clue as the correct character string. Can be guessed. This estimation result is input as a candidate character to the candidate input unit 223 via a keyboard or the like, and the replacement processing unit 224 replaces the corresponding code of the code holding unit 223 with the character code indicating this candidate character, thereby the corresponding portion. The character recognition result of can be corrected.

As a result, it is possible to selectively correct a portion having a high possibility that the reading result is erroneous and also based on the diversified information. In particular, by providing the reason for inconsistency, the operator can narrow down the range that can be considered as the characters that should correspond to the image data of the corresponding area, so that the operator can correct the work and be more accurate. It is possible to obtain various reading results.

Further, in the address area of the morpheme dictionary 232, information about the change of the place name such as the change of the town name is held together with the place name of each generation, and the semantic analysis unit 234 stores the information classified into the items in which the place name is described. When determining the consistency, if the configuration is such that the information on the item in which the corresponding date is described and the information on the place name change described above is referred to, more precise determination is possible.

In this case, for example, the morpheme corresponding to the erroneous recognition result is searched from the address area of the morpheme dictionary 232 based on the information about the place name before and after or the place name change.
It may be provided to the recognition complement processing unit 220 as an example of the candidate character string. This makes it possible to more strongly support the character recognition process from the image by utilizing the contents of the morpheme dictionary 232 and the contents of the related description written in the original family register.

Further, by devising the structure of the pattern dictionary 213, it is possible to flexibly deal with characters made up other than common kanji and common kanji (hereinafter referred to as "kanji") and used for subsequent character recognition. It is possible to In FIG. 6, claim 5
FIG. 3 is a configuration diagram of an embodiment of a family register information filing system to which the invention of FIG.

In FIG. 6, the family register information filing system has a registration processing unit 250 corresponding to the registration means 133 added to the family register information filing system shown in FIG.
The registration processing unit 250 responds to the instruction from the operator.
A new character code is associated with the dot pattern in the designated area, and the pattern dictionary 213 of the character recognition device 210 is associated.
It is configured to be registered in the character formation area 214 provided in the.

In the registration processing section 250, the image cutout section 251 reads out a designated dot pattern from the image buffer 201 in response to a registration instruction from the user, and the pattern creating section 252 uses the dot pattern as a basis. Then, a character formation pattern to be newly registered as a character pattern is created. The pattern creating unit 253 may extract a pattern in which at least one line segment is arranged in a specific positional relationship by performing a thinning process on the dot pattern in the designated area, for example. Then, this pattern is used as a collation character pattern corresponding to the character represented by the dot pattern described above,
It may be sent to the writing processing unit 254. At this time, when the original dot pattern is an image of a character by a writing brush, this dot pattern may be used as a collating character pattern for a writing brush.

In addition, the code determination unit 253 retrieves an unregistered character code from the character formation area 214 in response to the input of the registration instruction described above, and outputs this character code as a character code corresponding to the new character pattern. The writing processing unit 254 corresponds to this character code,
The character formation pattern and the dot pattern described above are written in the character formation area 214 of the pattern dictionary 213.

In this way, when the matching processing unit 212 corresponding to the matching means 132 determines that the corresponding character pattern does not exist in the pattern dictionary 213, a new character pattern is used as a character forming pattern as necessary. You can register. For example, the operator performs matching with various candidate characters by the process of the recognition complementation processing unit 220 for the dot pattern that has been undetermined by the matching processing unit 212, and as a result, the operator matches the corresponding dot pattern. When is determined to correspond to the typesetting, the registration instruction may be input by operating the keyboard or the like to activate the above-described registration processing.

In this case, the decomposition processing unit 231 of the analysis processing unit 230 decomposes the character string including this character code as a proper noun according to the input of the character code for character formation, and the syntax analysis unit 233. It may be provided to the processing of the semantic analysis unit 234. Thereby, the text information obtained as the recognition result can be classified into each item by the processing of the analysis processing unit 230 regardless of the presence or absence of the typesetting.

By registering a new character in the pattern dictionary 213 as described above, the collation processing unit 212 and the recognition complementation processing unit 220 thereafter perform character recognition including this character. Therefore, the recognition rate can be improved. Furthermore, the analysis processing unit 230
In, in the case where a proper part-of-speech (for example, surname, first name, etc.) for a proper noun including a character is determined based on the processing result of the syntactic analysis unit 233 and registered as a new element of the corresponding part-of-speech in the morphological dictionary 232. After that, this proper noun can be treated like other morphemes.

In this way, the character recognition device 2 flexibly responds to the characters that often appear as characters representing a person's name.
Since the recognition process by 10 can be strongly supported and the recognition result including the character formation can be used for the itemization process by the analysis processing unit 230, the filing work of the family register information can be performed more efficiently. Further, instead of comparing the new format family register with the original family register on paper, it is also possible to compare both on the display screen of the display device 202.

FIG. 7 is a block diagram showing an embodiment of a family register information filing system to which the invention of claim 6 is applied. In FIG. 7, the family register information filing system includes a collation data creation unit 261 instead of the output processing unit 240 illustrated in FIG. 3, and the collation data created based on the contents of the family register data file 203 is converted into the pattern conversion unit 222. The image is sent to the image synthesizing unit 223 of the recognition complementation processing unit 220 via the image compensating processing unit 220, and is used for the synthesizing process with the image of the original family register held in the image buffer 201.

The collation data creation unit 261 is, for example,
The contents of the family register data file 203 are compared with the text information obtained as a recognition result, all the character codes other than the overlapping portions are converted into character codes indicating blanks, and the information corresponding to each item is the original. It is only necessary to create collation data that is the position occupied by the text information and that the other part is blank.

In this case, the position of the character code other than the blank in the collation data indicates the position where the information corresponding to each item is to be displayed. The function is realized. Further, the pattern conversion unit 222 receives the above-mentioned collation data instead of the recognition result from the code holding unit 221, retrieves the corresponding character pattern from the pattern dictionary 213 of the character recognition device 210, and sequentially performs the image synthesis unit 223. You can send it to.

In response to this, the image composition section 223
Similar to the process of synthesizing with the recognition result, in the image corresponding to the original family register, the received character patterns are sequentially arranged and synthesized in an area adjacent to the range in which the dot patterns representing each character are distributed, Display device 202
You can send it to. In this way, the pattern converting unit 262 and the image synthesizing unit 223 operate according to the input of the collation data to realize the function of the second synthesizing unit 142, and to obtain the image of the original family register and the information of each item. A series of character patterns that are represented can be combined and displayed on the display device 202 corresponding to the second display unit 143.

As a result, the operator can proceed with the collation work while closely comparing the information recorded in the original family register with the information classified into the items, so that the original family register corresponding to the information of each item can be advanced. The information can be intuitively and accurately grasped, and the work can be efficiently performed, and the work load on the operator can be significantly reduced. Further, if the image composition unit 223 highlights the dot pattern area corresponding to each item in the image of the original family register, it is possible to more easily grasp the information of the original family register corresponding to the information of each item. .

Furthermore, the pattern conversion section 222 searches for a character pattern of a writing brush instead of searching a character pattern of a standard typeface from the pattern dictionary 213 for a designated item in accordance with an instruction from the operator. Then, the function of the conversion unit 125 described in claim 4 is realized, and for the portion described by the brush in the original family register, the information of the corresponding item can be displayed by the brush body.

As described above, by displaying the information of the corresponding item adjacent to the original image data by using the typeface similar to the original family register, the operator can display the information and items described in the original family register. It is possible to intuitively collate the classified information with the dot pattern matching.
As a result, it is possible to further reduce the work load on the operator during the collation work.

Further, as described above, since the collation work is performed on the screen, when the inconsistency is detected in the collation work, the process of the recognition complement processing section 220 can be directly performed. Become. For example, the operator inputs a candidate character string to the candidate input unit 222 of the recognition complementation processing unit 220 via the keyboard 202, operates the replacement processing unit 223, and causes the information of the corresponding item to correspond to the candidate character string. Replace with the character code.

In this way, while proceeding with the collation work,
Since it is possible to successively correct detected errors, it is possible to dramatically improve the operability of the collation work and the final correction work associated therewith, and to efficiently proceed the filing work of family register information. it can. As described above, according to the present invention, by the processing of the recognition complementation processing unit 220 and the analysis processing unit 230 and the collation processing using the recognition complementation processing unit 220,
Since the recognition result by the character recognition device can be complemented, it can be applied to the filing work of various documents which could not be the target of such filing work in the past.

For example, as the pattern dictionary 213 provided in the character recognition device 210, ones corresponding to cursive writing and line writing are prepared, and morphemes and syntactic rules in old sentences are stored in the morpheme dictionary 222 and the syntactic rule holding unit 235, respectively. You can leave it. As a result, the system of the present invention can be applied to filing of old documents and the like, which can greatly contribute to preservation and utilization of valuable cultural assets.

[0072]

As described above, the invention of claim 1 is
By adopting morphological analysis, it is possible to automatically classify the recognition results for each item, and it is possible to automate the itemization work as well as the reading and inputting work of the character information written in the manuscript. The work load on the user can be significantly reduced.

Further, in the invention of claim 2, when the recognition result is corrected by considering the morphological analysis result, when the recognition error occurs, the connection of the character strings before and after it is used as a clue to detect the recognition error. I can fix it,
The recognition processing by the character recognition means can be complemented.
Further, according to the invention of claim 3, by displaying the unrecognized dot pattern and the candidate character side by side, the result of the correction work can be taken as the recognition result while sufficiently comparing them, and therefore the character recognition means. A more accurate recognition result can be obtained by complementing the recognition processing by.

In particular, by converting the candidate pattern into a typeface corresponding to an unrecognized character and displaying the typeface, the recognition result can be corrected in consideration of the characteristics of the character shape of the typeface. The recognition processing by means can be supported more strongly. Further, if the invention of claim 5 is applied and a new character pattern is registered in the pattern dictionary as needed, further improvement in the recognition rate can be expected.

Further, according to the invention of claim 6, the information classified for each item can be displayed by being superimposed on the image of the original document, whereby the information of each item can be displayed in a corresponding portion of the document. Therefore, it is possible to strongly support the collation work and significantly reduce the burden on the user.

[Brief description of drawings]

FIG. 1 is a principle block diagram of a document filing system according to claims 1, 5, and 6;

FIG. 2 is a principle block diagram of the document filing system according to claims 2 to 4;

FIG. 3 is a configuration diagram of an embodiment of a family register information filing system to which the inventions of claims 1 and 3 are applied.

FIG. 4 is a diagram illustrating an image combining process.

FIG. 5 is a configuration diagram of an embodiment of a family register information filing system to which the invention of claim 2 is applied.

FIG. 6 is a configuration diagram of an embodiment of a family register information filing system to which the invention of claim 5 is applied.

FIG. 7 is a configuration diagram of an embodiment of a family register information filing system to which the invention of claim 6 is applied.

FIG. 8 is a diagram showing a configuration example of a conventional family register information filing system.

[Explanation of symbols]

 111 character recognition means 112 decomposition means 113 classification means 114 inconsistency detection means 115 correction means 120 complementing means 121 input means 122 first combining means 123 first display means 124 confirming means 125 converting means 126 image replacing means 131, 213 patterns Dictionary 132 Collating means 133 Registering means 141 Position specifying means 142 Second synthesizing means 143 Second displaying means 201 Image buffer 202 Display device (display) 203, 313 Family register data file 210 Character recognition device 211 Area extracting part 212 Pattern matching part 214 character formation region 220 recognition complementation processing unit 221 code holding unit 222 pattern conversion unit 223 image synthesis unit 224 candidate input unit 225 replacement processing unit 230 analysis processing unit 231 decomposition processing unit 232 form Elementary dictionary 233 Syntax analysis unit 234 Semantic analysis unit 235 Syntax rule storage unit 236 Analysis control unit 237 Correction processing unit 240 Output processing unit 250 Registration processing unit 251 Image cutout unit 252 Pattern creation unit 253 Code determination unit 254 Write processing unit 261 Collation data creation unit 301 Microfilm 302 Microfilm reader 303 Copy 304 Collation list 310 Document filing system 311 Keyboard 312 Edit processing unit 314 Collation list creation unit

Continuation of the front page (51) Int.Cl. ⁶ Identification number Office reference number FI Technical display location 9194-5L G06F 15/403 310C 15/62 330D (72) Inventor Kyosuke Hirono 2 Kitayuki Nishi-ku, Yokohama-shi, Kanagawa 8-4 chome, Fujitsu Keihin System Engineering Co., Ltd. (72) Inventor Hideyuki Yoshida 2-8-4 Kitayuki, Nishi-ku, Yokohama City, Kanagawa Prefecture Fujitsu Keihin System Engineering Co., Ltd.

Claims (6)

[Claims] 1. A document filing system for reading a character having information on a plurality of items according to a certain format on a manuscript, converting the character into a character code, and storing the character code, which is included in an image corresponding to the manuscript. Character recognition means for recognizing each character based on the dot pattern representing the character and outputting the corresponding character code as a recognition result, and decomposition means for decomposing the text information obtained as a recognition result into its parts of speech, A series of parts of speech obtained by the disassembling means are analyzed to classify the information represented by the series of parts of speech for each of the plurality of items, and a classifying means for storage processing is provided. And document filing system. 2. A document filing system for reading characters written on a manuscript, converting the characters into character codes and storing the character codes, wherein each character is converted based on a dot pattern representing characters included in an image corresponding to the manuscript. Character recognition means for recognizing and outputting the corresponding character code as a recognition result, decomposing means for decomposing the text information obtained as a recognition result into its parts of speech, and a series of parts of speech obtained by the decomposing means. An inconsistency detecting unit that analyzes the relevance and detects an inconsistency as information represented by characters, and based on the detection result by the inconsistency detecting unit, corrects the text information of the recognition result and saves it. A document filing system, characterized by comprising: 3. The document filing system according to claim 1, further comprising a complementing unit that complements a recognition result of a dot pattern that could not be recognized by the character recognizing unit and sends the result to the disassembling unit. Complementing means, the input means for inputting a candidate character corresponding to the unrecognized dot pattern, and the character pattern representing the candidate character in the area adjacent to the unspecified dot pattern in the image corresponding to the document, A first synthesizing unit for synthesizing the image corresponding to the original, and a first synthesizing unit for displaying the image obtained by the first synthesizing unit.
And a confirmation means for confirming the candidate character input via the input means as a recognition result of the corresponding dot pattern in response to the input of the confirmation instruction and transmitting it to the decomposition means. A document filing system characterized by the fact that there is. 4. The document filing system according to claim 3, wherein the first synthesizing unit outputs image information of an area adjacent to an unrecognized dot pattern to the character in response to input of a character pattern indicating a candidate character. An image replacement means for replacing with a pattern, and a conversion means for converting a candidate character into a specified typeface according to an input of an instruction for designating the typeface and sending it to the image replacement means. Document filing system to do. 5. The document filing system according to claim 1 or 2, wherein the character recognizing means stores a pattern dictionary storing a character pattern representing a corresponding character corresponding to a character code, and an image corresponding to a document. In accordance with the input of a dot pattern representing each character included in, the collation that collates with each of the character patterns stored in the pattern dictionary, and outputs the character code corresponding to the character pattern that matches the dot pattern as a recognition result. A document characterized by comprising: means for registering a new character code to a character pattern corresponding to the dot pattern and registering the character pattern in the pattern dictionary in accordance with a matching result by the matching means. Filing system. 6. The document filing system according to claim 1, wherein, based on the position occupied by the classified information in the text information of the recognition result, the range of the dot pattern on the original document corresponding to each information is determined. A position designating means for designating as an item area; a second synthesizing means for synthesizing a character pattern representing corresponding information in an area adjacent to each item area of the image data corresponding to the document; A document filing system comprising: second display means for displaying the image data obtained by the means.