JPH0922442A - Electronic management system for image document data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve the utility value of an electronic data base by separating and extracting the character and segment information out of the dot information of an image data form and converting these extracted information into the codes and managing these codes by a text file of high utility value. SOLUTION: This system consists of an electronic file medium 30, a paper medium 20, a scanner 8, a segment/character separation part 13, an editing device 2, a recording medium 9 and a retrieval device 11. The part 13 receives the image data from the medium 30 or the scanner 8, normalizes with thinning of lines the image data into an (N-dot×M-dot) size, allocates the normalized image data to 1000 types of vertical, horizontal, upper right oblique and upper left oblique line elements, divides these allocated image data into (A×B) areas to compare them with a character symbol dictionary J5 which is previously produced and stored, and then separates in sequence the divided areas into the segments and the character symbol/photo images by extraction of features in order to discriminate the directional line element feature value by the multi- dimensional vector out of the image data.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image / document material for facilitating information retrieval and management by identifying and extracting character codes from image data format information such as graphic information, photograph / image information and character information. Regarding electronic management system.

[0002]

2. Description of the Related Art Document materials include character information, drawings, photographic images, and audio data. Conventionally, there is no device for identifying and extracting a character code from information in an image data format in which graphic information, photograph / image information, character information and the like are mixed among them. Here, the information of the image data format is, for example, dot image information of a dot image read by an image scanner.

Document materials include paper media and electronic file media. When a paper medium is taken in as document data, it is generally read by an image scanner and converted into an electronic file. The data format includes a character code format and an image data format. The character code format data is, for example, a text format data group expressed in JIS kanji code, which can be directly searched by an arbitrary character code, and the image data format data is a figure, an image, a photograph, or the like. It is a data group consisting of dot information of, and cannot be searched by the character code.

FIG. 4 shows the configuration of a conventional electronic management system for document materials. The system configuration includes an electronic file medium 30, a paper medium 20, a scanner 8, an editing machine 2,
It comprises a recording medium 9 and a searcher 11.

The electronic file medium 30 is a recording medium such as a floppy disk or MT, communication data from a communication line, etc., and is a medium which can be directly input to the editing machine 2. Further, this medium is stored in both a character code format and an image data format.

The paper medium 20 is a medium that cannot be directly input to the editing machine 2, and is a medium such as a paper drawing, a printed book, a photograph, or an image. The scanner 8 uses the paper medium 2
0 is converted into dot information in the image data format that can be input by the editing machine 2 and supplied.

The editing machine 2 receives the input data from the electronic file medium 30 or the scanner 8 and stores it in the recording medium 9 for each unit data, adds this data format information and the index corresponding thereto, and creates an index.・ Save as a heading for searching. By this, an electronic database of all information is realized.

The searcher 11 receives the input parameter 12 of the search condition from the outside, and displays and outputs the data group satisfying the search condition in the recording medium 9 in the data display format corresponding to the data group. As a result, it is possible to search / view data that is needed at any time.

[0009]

In the above description, character information is included in the image data format information captured by the scanner 8 or the like. If the character information of this dot expression can be converted into a character code, the utility value will be significantly improved as a more effective electronic database. 2. Description of the Related Art Conventionally, there is a reading device that converts a fixed-size sentence into a character code from the surface of a book or newspaper.

However, some information in the image data format, such as a circuit drawing, has a mixture of circuit symbols, symbol character information, and connecting wiring line segment information, and among these dot information. It is difficult to separate and extract the target character information from, and it is not realized yet. For this reason, it was not possible to make a character code database file and arbitrarily convert it into a highly useful database for searching, copying, processing, etc. for storage.

Therefore, the problem to be solved by the present invention is to separate and extract character information and line segment information from the dot information in the image data format, convert the code, and manage the text file with high utility value. The purpose is to do.

[0012]

In order to solve the above-mentioned problems, in the structure of the present invention, in the image data 100,
In order to identify the direction line element feature amount by a multidimensional vector, it is normalized to a size of N dots × M dots, thinned, and assigned to four types of line elements: vertical, horizontal, diagonally upper right, diagonally upper left. After that, this is further divided into an A × B area, and compared with a character / symbol dictionary J5 created and stored in advance, and a line segment and a character / symbol / photo image are sequentially separated by feature extraction. The recognition means 13 is provided as a constituent means. As a result, the utility value is obtained by receiving data in the image data format including dot information such as figures and images, which are electronic document materials, and separating and extracting from the image data 100 into character information, line segment information, and image information. To realize an electronic management device for document materials converted to a high code data format.

In addition to the above configuration, the line segment character separation recognition unit 1
An electronic management device for image / document data, which receives the character code, the line segment data, and the photographic image data separated by 3 and adds the association information to be restored to the diagram corresponding to the original image data 100 and saves it in the recording medium. To be realized. As a result, the electronic management device has a higher utility value as a database file for arbitrary searches by character codes, copying and processing of character strings and sentences, and the like.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows the configuration of an electronic management system for document material of the present invention. The implementation outline of separation and extraction will be described.
The method of separating line segments, character symbols, and photographic images is basically based on feature extraction. That is, the direction line element feature amount of the information read by the scanner 8 is identified by a multidimensional vector. First, one piece of information is normalized to a size of N dots × M dots and thinned. After thinning the line, vertical, horizontal,
It is assigned to four types of line elements, diagonally upper right and diagonally upper left. This is further divided into an A × B area, and compared with a representative vector of the recognition target line segment / character symbol / photo image, which is created and saved in advance, for example, the character / symbol dictionary J5 to compare the line segment with the character symbol / photo image. Further, the character symbol and the photographic image are sequentially separated.

The system configuration is an electronic file medium 30.
A paper medium 20, a scanner 8, a line segment character separation recognition unit 13, an editing machine 2, a recording medium 9, and a search machine 11. The present invention has a configuration in which a line segment character separation recognition unit 13 is added.

The line segment character separation recognizing unit 13 operates the image data 10 from the electronic file medium 30 or the scanner 8.
Upon receiving 0, the character / symbol code is extracted from this image data by separating it into a character / symbol, a drawing and a photographic image. As shown in FIG. 2, the functional structure of the line segment character separation recognition unit 13 includes a first line segment separation unit A1, a second line segment separation unit A2, a third line segment separation unit A3, and a first character segment. Character recognition comparison between the symbol separation unit B1, the second character symbol separation unit AB1, the photographic image separation unit C1, the fourth character symbol separation unit C2, the fifth character symbol separation unit C3, the photographic image separation unit D1 Part C5
And a character / symbol dictionary J5, a character storage unit C6, an unidentifiable output unit C7, and a manual identification unit C8.

The first line segment separating unit A1 is used for the image data 1
00, the line segment and the character / symbol / photograph image area data are separately output. The second line segment separating unit A2 receives the line segment area data from the first line segment separating unit A1, and receives the line segment and the character symbol /
Separate output to photographic image. The first character symbol separator B1 is
The character / symbol / photo image from the first line segment separating unit A1 is received, and the character / symbol and the photo image are separated and output.

The photographic image separating unit C1 receives the photographic image data from the first character / symbol separating unit B1, and further separates and outputs the character / symbol and the photographic image. The fifth character symbol separator C3
Upon receiving the photographic image from the photographic image separating section C1, the character symbols remaining therein are extracted, and the photographic image and the photographic image are separated and output. The photographic image separating section D1 outputs the photographic image data extracted by the fifth character / symbol separating section C3.

The second character / symbol separating section AB1 receives the extracted character data from the second line segment separating section A2 and the first character / symbol separating section B1, synthesizes the two and stores them in the character / symbol storing section C4. . Further, the extracted character data from the fourth character / symbol separation unit C2 and the fifth character / symbol separation unit C3 is also temporarily stored in the character / symbol storage unit C4.

The character recognition / comparison unit C5 receives the character image data from the character / symbol storage unit C4 and sequentially compares it with the character dictionary / character size dictionary comparison dictionary, which is a character / symbol dictionary J5 data group. Of the character image data is discriminated as the character code of the character image data, and the character code is supplied to the editing machine 2 via the character storage section C6. If the matching rate is lower than the desired level, it is output to the manual identification section C8 as an unidentifiable character. The manual identification unit C8 displays an ambiguous character shape that cannot be determined on the screen, identifies a person, and supplies the identified character code to the editing machine 2 via the character storage unit C6.

An example of the results obtained by the above means is shown in FIG. FIG. 3A shows the original image data 100.
FIG. FIG. 3B shows the result of separating only the line segments. In FIG. 3C, the character codes extracted separately from the line segments are merged again and printed.

The editing machine 2 receives the character code, the line segment (graphic data) and the photographic image which have been separated and extracted as described above, and stores them in an associated form so that they can be restored as an image diagram, and An index for storage is attached to and stored in the recording medium 9 in association with a heading for index / search. As a result, the extracted character code can be used as a keyword at the time of search, and a database of more valuable information can be realized. In particular, in the case of image data including many character codes such as books and drawings, it is possible to use search and the like very effectively.

Further, the keywords obtained by the above extraction may be registered in the recording medium 9 as an index for retrieval, or may be added or changed at any time later to facilitate the reference retrieval of the database. . Further, the separated character code, graphic data, photographic image, etc. may be edited / corrected / written in as needed to make a highly useful database.

[0024]

Since the present invention is configured as described above, it has the following effects. The line segment character separation recognition unit 13 receives an electronic file in an image data format or image data read by a scanner 8 from a paper medium, and separates a character symbol, a drawing, and a photographic image from the image data to obtain a character symbol. By extracting the code, it becomes possible to use the character code in the editing machine 2 as a keyword at the time of searching, and an effect that a database of information of higher utility value can be realized can be obtained. Especially, it is effective for separating and extracting image data including many character codes such as circuit diagrams and architectural drawings.

The index stored in the recording medium 9 can be used to add and change keywords for search later to facilitate reference search of the database, and separate character codes, graphic data, photographic images, etc. can be edited / edited at any time. It is also possible to make corrections / notes and make a highly useful database.

[Brief description of drawings]

FIG. 1 is a block diagram of an electronic management system for document material according to the present invention.

FIG. 2 is a functional structure diagram of a line segment character separation recognition unit 13 of the present invention.

FIG. 3A of the present invention is image data 100 which is an original drawing. (B) is a diagram in which only line segments are separated from the image data 100. (C) is a diagram in which the separated line segment and the extracted character code are combined again.

FIG. 4 is a block diagram of a conventional electronic management system for document materials.

[Explanation of symbols]

 A1 First line segment separation unit B1 First character symbol separation unit AB1 Second character symbol separation unit C1, D1 Photo image separation unit 2 Editing machine A2 Second line segment separation unit C2 Fourth character symbol separation unit A3 Third line segment Separation unit C3 Fifth character symbol separation unit C4 Character symbol storage unit J5 Character symbol dictionary C5 Character recognition comparison unit C6 Character storage unit C7 Unidentifiable output unit 8 Scanner C8 Manual identification unit 9 Recording medium 11 Search machine 13 Line segment character separation recognition Part 20 Paper medium 30 Electronic file medium 100 Image data

Claims (2)

[Claims] 1. An image data format data consisting of dot information of a drawing which is an electronic document material is received,
In the electronic document management device for document data that separates and extracts from this image data (100) into character information, line segment information, and image information, in the image data (100), the direction line element feature amount is a multidimensional vector. In order to discriminate by using, the size is normalized to N dots × M dots, the line is thinned, and assigned to four types of line elements of vertical, horizontal, diagonally upper right, and diagonally upper left. Line segment and character symbol / photograph image are sequentially separated by feature extraction by comparing with a character / symbol dictionary J5 created and stored in advance.
An electronic management system for image and document materials characterized by having 3) and having the above. 2. The line segment character separation recognition unit (13) receives the separated character code, line segment data, and photographic image data, and adds association information to be restored to the drawing corresponding to the original image data (100). An electronic management system for image / document data according to claim 1, which is stored in a recording medium.