JP4872285B2 - Document management apparatus, document management system, and document management method - Google Patents

Description

  The present invention relates to a technology for centrally managing electronic documents generated by various application software by converting them into electronic documents of a predetermined format.

  In general, in an electronic document created by application software for creating a document, each character is basically represented by a character code, which is convenient when performing processing such as correction, processing, and search. Some applications form and store or output an electronic document as image data. In this case, an original character code string is also attached to the image data. Therefore, the image data of the electronic document can be browsed and various processes can be performed on the character code string.

  However, when the character is modified, the character code of the modified part may be replaced with a special code and the character code may be lost. In addition, in the case of an application that expresses a modification by combining character codes, there is a case where the character code does not make sense as a sentence even if it remains. Further, when a character image is pasted, the character code corresponding to the character image is not originally provided. In these cases, there arises a problem that various processes using the character code are hindered.

  Therefore, in Patent Document 1, character recognition such as OCR is performed on image data of an electronic document, and a character code string obtained as a result is stored or output as a new electronic document.

JP 2004-178438 A

  However, in the invention described in Patent Document 1, the original character code string originally attached to the electronic document is discarded, and the character code string obtained by character recognition is adopted. When there is a recognition error, an incorrect character code is generated, which is different from the original character code.

  The present invention has been made in view of the above-described circumstances. A document management apparatus, a document management system, and a document management method for generating an appropriate character code string from an electronic document including a character code string and image data obtained by imaging the character code string. The purpose is to provide.

In order to achieve the above object, the document management apparatus adopted by the present invention comprises: input means for receiving image information including information obtained by imaging character code information and an electronic document having the character code information; and Document image generation means for developing the received image information of the electronic document into an image, layout analysis means for analyzing the layout of the image developed by the image generation means and recognizing a region, and recognized by the layout analysis means Determination means for determining whether the area is a character area, character recognition means for generating character code information by performing character recognition processing on the area determined as a character area by the determination means, and the input means About the received character code information and the character code information generated by the character recognition means, Evaluated by document analysis according to algorithms, evaluation results on the basis, a character code selection means for selecting one of the character code information, by joining the parts of the character code information selected by the character code selection means And character code information creating means for character code information of the electronic document.

  With this configuration, the character code information creating unit generates the character code information by connecting the character code information obtained by the layout analysis unit, the character recognition unit, and the determination unit. In the character code information attached to the image information, such as when the character does not make sense as a sentence or when the character image is pasted on the image information and there is no character code, The obtained character code can be complemented, and appropriate character code information can be obtained.

  The document management apparatus preferably includes an electronic document storage unit that converts the image information into another format and stores the converted image information together with the character code information created by the character code information creation unit.

  In the document management apparatus, the predetermined algorithm is preferably an algorithm according to natural language analysis.

The document management system employed by the present invention includes a terminal device and a management server connected to the terminal device via a network,
The management server develops image information including information obtained by imaging character code information and an electronic document having the character code information, and image information of the electronic document received by the input means into an image. Document image generation means, layout analysis means for analyzing the layout of the image developed by the image generation means and recognizing the area, and determination means for determining whether or not the area recognized by the layout analysis means is a character area And character recognition means for generating character code information by performing character recognition processing on the area determined as the character area by the determination means, and character code information received by the input means and the character recognition means For character code information, the corresponding parts are analyzed according to a predetermined algorithm and evaluated, Based on the value result and a character code selection means for selecting one of the character code information, and the character code character code information of the electronic document by joining the parts of the character code information selected by the selecting means Character code information creating means,
The terminal device includes electronic document transmission means for transmitting an electronic document to the input means of the management server.

The document management method employed by the present invention includes an image information including information obtained by imaging character code information and an electronic document having the character code information, and image information of the electronic document received in the input step. A document image generation step of expanding the image into an image, a layout analysis step of analyzing the layout of the image expanded in the image generation step to recognize the region, and whether the region recognized by the layout analysis step is a character region A character recognition step for generating character code information by performing character recognition processing on the region determined as the character region by the determination step, the character code information received in the input step, and the character The character code information generated in the recognition step Portions to each other was evaluated by the document analysis according to a predetermined algorithm, the evaluation results on the basis of a character code selection step of selecting one of the character code information, each character code information selected by the character code selection step And a character code information creating step for joining the portions to obtain character code information of the electronic document.

  According to the document management apparatus of the present invention, the character code information creating means generates the character code information by connecting the character code information obtained by the layout analysis means, the character recognition means, and the determination means. Even if the code does not make sense as a sentence, or even if a character image is pasted and there is no character code, this part of the character code can be complemented, making use of the character code information of the electronic document However, appropriate character code information can be obtained.

<A. Embodiment>
A first embodiment according to the present invention will be described below with reference to the drawings. FIG. 1 is a diagram showing an overall configuration of a document management system 1 according to the present embodiment. The system 1 includes a user terminal 100, a management server 200, and a database 300 connected via a network 400.
The user terminal 100 has a function of generating an electronic document using various application software. The electronic document generated by these application software has image information including information obtained by imaging character code information (for example, JIS code, EUC, shift JIS) and the character code information.

The management server 200 converts image information of an electronic document generated by a certain kind of application software transmitted from the user terminal 100 by processing to be described later into image information of a predetermined format, and converts the image information of the predetermined format into character code information. At the same time, it is stored in the database 300.
Specifically, the format of the image information is specified from the file name of the transmitted electronic document, and if it is a convertible format, it is converted into a predetermined format (for example, PDF (Portable Document Format)).

As described in the prior art, the electronic document generated by the application software is accompanied by a character code corresponding to each character of the electronic document, and this character code information is stored even when converted into a predetermined format. Attached. For this reason, as shown in FIG. 2, the electronic document stored in the database 300 stores image information and character code information corresponding to each character of the image information for the file name of the electronic document. .
Specifically, the image information “001-D” / character code information “001-C” for the file name “001” of the electronic document, and the image information “002-D” / character for the file name “002”. The code information is “002-C”.

<A-1. Configuration of Management Server 200>
FIG. 3 is a block diagram illustrating a functional configuration of the management server 200. The management server 200 includes a data transmission / reception unit 210, a document image generation unit 220, a layout analysis unit 230, a determination unit 240, an OCR processing unit 250, a character code information selection unit 260, a character code information creation unit 270, and a control unit that controls them. 280. The control unit 280 includes a CPU (Central Processing Unit), a ROM (Read Only Memory), and a RAM (Random Access Memory) (not shown), and has a function of controlling various processing operations.

The data transmission / reception unit 210 exchanges data with an external device (not shown) connected to the network 400 such as the user terminal 100.
The document image generation unit 220 converts the image information of the electronic document generated by the application software into an image. That is, the electronic data is converted into image data to be printed on paper. The layout analysis unit 230 performs layout analysis on the image data imaged by the document image generation unit 220 and divides the image data into character regions and graphic regions (drawings and photographs). The determination unit 240 determines whether the area recognized by the layout analysis unit 230 is a character area. When the determination unit 240 determines that the region recognized by the layout analysis unit 230 is a character region, the OCR (Optical Character Reader) processing unit 250 performs character recognition of the region and character codes for each character of the document data Confirm the information.

The character code information selection unit 260 compares the character data information attached to the electronic document with the character code information determined by the OCR processing unit 250, and selects the optimum character code information by natural language analysis. Then, the character code information creation unit 270 creates character code information of the electronic document by connecting the portions of the character code information selected by the character code information selection unit 260. This character code information is stored in the database 300 as an electronic document together with image information.
Here, natural language analysis is a technique for analyzing sentences using a known morphological analysis or syntactic analysis technique, and in this embodiment, analyzes whether the meaning is understood as words. .

<A-2. Operation of document management system>
Next, the character code information creation processing operation of the document management system 1 according to the present embodiment will be described based on the flowchart of FIG.
The management server 200 starts its operation when an electronic document generated by predetermined application software is transmitted from the user terminal 100 (step S1; YES).
In this description, a case where there is one electronic document transmitted from the user terminal 100 will be described, but it is a matter of course that the user may transmit the electronic documents collectively.

  The control unit 280 temporarily stores the character code information C1 attached to the image information D1 of the transmitted electronic document (step S2). Further, the control unit 280 transfers the image information D1 received by the data transmission / reception unit 210 to the document image generation unit 220. The document image generation unit 220 generates image information D2 obtained by rasterizing the transferred image information D1 (step S3).

Next, the control unit 280 transfers the image information D <b> 2 generated by the document image generation unit 220 to the layout analysis unit 230. The layout analysis unit 230 divides the image information D2 into a document area and a graphic area (step S4).
Further, the control unit 280 causes the determination unit 240 to determine whether or not these regions are character regions among the regions divided by the layout analysis unit 230. Information serving as a character area is transferred to the OCR processing unit 250. The OCR processing unit 250 determines the character code for the transferred character area character and generates character code information C2 (step S5).

  The control unit 280 reads the character code information C1 stored in step S2, and transmits it to the character code information selection unit 260 (step S6). Further, the control unit 280 transmits the character code information C2 determined for the character by the OCR processing unit 250 to the character code information selection unit 260. The character code information selection unit 260 compares the character code information C1 and the character code information C2, and selects the optimum character code information by natural language analysis (step S7).

  Further, the control unit 280 creates character code information of the electronic document by connecting the character code information selected by the character code information selection unit 260 with the character code information creation unit 270 (step S8).

  Thereafter, the image information D1 is converted into other image information D2 in a predetermined format, and the created character code information is stored in the database 300 together with the image information D2.

<B. Specific example>
Next, a specific example of this embodiment will be described with reference to FIGS.
As shown in FIG. 5, in the electronic document, a sentence X1 “This is a shaded character.” Is decorated with a shaded character, and a sentence X2 “This is a outlined character.” Decorated. Here, the application software is F1 and F2.
The sentence X1 in the application software F1 is “This is a shaded character.” → “This is a shaded text character.” Character code information is assigned to these 22 characters. Will be. On the other hand, the sentence X1 in the application software F2 is “This is a shaded character”, and accurate character code information is assigned to these 11 characters.
In addition, since “this is a hollow character” in the sentence X2 is captured as a character image, character code information is not allocated to both the application software F1 and F2.

Here, the operation | movement of this embodiment is demonstrated referring FIG. 6 about the text X1.
As mentioned above, the input sentence X1 is “This is a shaded character.” Decorated with a shaded character. In the application software F1, “This is a shaded letter character. . " For this reason, character data information in which character codes are assigned to these 22 characters is attached to the image information.

Here, to explain in correspondence with the processing of FIG. 4, in the processing of step S2, character code information corresponding to 22 characters is stored. In the process of step S3, image data (center in FIG. 6) obtained from the input document data of the character X1 is obtained. Through the processing in steps S4 to S6, an 11-character code shown on the right side in FIG. 6 is obtained. In the processing in step S7, natural language analysis is performed on the 22-character code and the 11-character code, and an optimal character code is selected. Of course, an 11-character code is selected here.
Then, in the process of step S8, the 22 character code of the character code information attached to the image information is replaced with 11 character code information, and the replaced character code information is attached to the image information.
Incidentally, JIS codes corresponding to “This is a shaded character” are represented as 2433, 246C, 244F, 3146, 4955, 242D, 4A38, 3B7A, 2447, 2439, and 2123.

  In addition, since no character code is assigned to “This is a hollow character” in the sentence X2, the character code information obtained in the processing of steps S3 to S7 is attached to the image information.

As described above, in the present embodiment, while utilizing the accurate character code information originally attached to the electronic document generated from various application software, there is no character code information corresponding to the character or the meaning of the text. Character code information can be obtained by complementing the parts that are not formed using the character codes obtained in the processes of steps S3 to S6. In addition, since the natural code analysis is performed in the character code selection process, the correct character code is selected even if the character recognition process (OCR process) has caused erroneous character recognition. Can do.
Furthermore, image information with accurate character code information attached can also be attached to other image information converted into a predetermined format. Even if the image information is in other formats, the character code information is used. Search and translation functions and re-editing are possible.

<C. Modification>
As mentioned above, although embodiment of this invention was described, this invention is not limited to each embodiment mentioned above, A various aspect is possible.
In the above-described embodiment, the document management system has been described in which the user terminal 100 and the management server 200 are connected via the network 400. However, the present invention is not limited to this, and the function of the management server 200 is built in the user terminal 100, The terminal 100 may be a document management device. The management server 200 may be installed inside or outside the company to perform document management. Furthermore, the management server 200 may be provided with a translation unit and a translation dictionary unit, and the management server 200 may be used as the management server.

  Even when the annotation data such as annotations and sticky notes is included in the document data, if the management server 200 of this embodiment is used, character code information is created from the characters written by the annotations. It is also possible to search for annotations as keywords.

  In the embodiment, only one user terminal 100 and one management server 200 are shown in FIG. 1, but the document management system 1 may include a plurality of user terminals 100 and management servers 200. is there.

  In the above-described embodiment, the user terminal 100 is realized by the control unit and the management server 200 is realized by the control unit 280 executing the program. However, a part or all of this is realized by hardware. Also good.

It is a figure which shows the structure of the document management system which concerns on embodiment of this invention. It is a figure which shows typically the data stored in the database used for embodiment of this invention. It is a block diagram which shows the structure of the management server used for this invention embodiment. It is a flowchart which shows operation | movement of the document management system by embodiment. It is a figure which shows the specific example by embodiment. It is a figure which shows the specific example by embodiment with FIG.

Explanation of symbols

DESCRIPTION OF SYMBOLS 1 ... Document management system 100 ... User terminal 200 ... Management server 210 ... Data transmission / reception part 220 ... Document image generation part 230 ... Layout analysis part 240 ... Determination part 250 ... OCR processing part 260 ... Character code Information selection unit, 270 ... character code information creation unit, 280 ... control unit, 300 ... database, 400 ... network.

Claims (5)

Input means for receiving image information including information obtained by imaging character code information and an electronic document having the character code information;
Document image generation means for developing image information of the electronic document received by the input means into an image;
Layout analysis means for recognizing a region by analyzing the layout of the image developed by the image generation means;
Determining means for determining whether the area recognized by the layout analyzing means is a character area;
Character recognition means for generating character code information by performing character recognition processing on an area determined to be a character area by the determination means;
With respect to the character code information received by the input means and the character code information generated by the character recognition means, the corresponding parts are analyzed by document analysis according to a predetermined algorithm, and one of the character codes is determined based on the evaluation result. A character code selection means for selecting information;
A document management apparatus comprising: character code information creating means for joining character code information selected by the character code selection means to form character code information of the electronic document. The document management apparatus according to claim 1.
The converted image information to another format, the document management apparatus, wherein the image information after conversion that you comprise an electronic document storage means for storing together with the character code information the character code information producing means. The document management apparatus according to claim 1 or 2,
The document management apparatus, wherein the predetermined algorithm is an algorithm according to natural language analysis. A terminal device, and a management server connected to the terminal device via a network,
The management server is
Input means for receiving image information including information obtained by imaging character code information and an electronic document having the character code information;
Document image generation means for developing image information of the electronic document received by the input means into an image;
Layout analysis means for recognizing a region by analyzing the layout of the image developed by the image generation means;
Determining means for determining whether the area recognized by the layout analyzing means is a character area;
Character recognition means for generating character code information by performing character recognition processing on an area determined to be a character area by the determination means;
With respect to the character code information received by the input means and the character code information generated by the character recognition means, the corresponding parts are analyzed by document analysis according to a predetermined algorithm , and one of the character codes is determined based on the evaluation result. A character code selection means for selecting information ;
Character code information creating means for connecting character code information pieces selected by the character code selection means to form character code information of the electronic document,
The terminal device is
An electronic document transmission unit that transmits an electronic document to the input unit of the management server. An input step of receiving image information including information obtained by imaging character code information and an electronic document having the character code information;
A document image generation step of developing image information of the electronic document received in the input step into an image;
A layout analysis step for recognizing a region by analyzing a layout of the image developed by the image generation step;
A determination step of determining whether or not the region recognized by the layout analysis step is a character region;
A character recognition step of generating character code information by performing character recognition processing on the region determined to be a character region by the determination step;
For the character code information received in the input step and the character code information generated in the character recognition step, the corresponding portions are analyzed by document analysis according to a predetermined algorithm, and either character code is determined based on the evaluation result. A character code selection step for selecting information;
A document management method comprising: a step of creating character code information by connecting portions of the character code information selected in the character code selection step to obtain character code information of the electronic document.

Images