JPS59135576A - Registering and retrieving device of document information - Google Patents

Abstract

PURPOSE:To shorten the time needed for production of a summary edition with a simple method by extracting automatically the information of an area among the printed document pictures which is marked by a fluorescent pen, etc. CONSTITUTION:A document 11 is scanned at a picture information input part 14 and then fed into a region discriminating part 15. The part 15 segmentes a fixed region centering on an area in the document 11 which is marked by a fluorescent pen. This segment region is stored in a memory 18 together with the entire document picture. For the region stored in the memory 18, a character train region is segmented by a segmenting part 16 and then stored again in the memory 18. The character train sent from the part 16 is summarized at a summary producing part 17 and then stored in the memory 18 in the form of a page of summary edition. Thus a summary edition of the printed document can be obtained automatically.

Description

DETAILED DESCRIPTION OF THE INVENTION The present invention relates to a document information registration and retrieval device that stores printed documents and other document information and retrieves them.

<Prior Art> Conventional devices for storing document information such as printed documents store image information of the entire document as it is on a magnetic disk, etc., and input management information of the document, such as the title, from another device. It was common. In addition, if there is a part of a printed document etc. where a human has attached marks 13 such as an angline 12 or a fluorescent pen mark to a document 11 such as a document as shown in FIG. There is a drawback that when the image information is stored in a magnetic tape or the like and the image is later read out, the document image becomes blurred.

In addition, when managing printed document images, existing devices manage document images sequentially, so it is necessary to output each document in its entirety and then check whether each document image is necessary for searching. The drawback was that it could not be recognized otherwise.

<Summary of the Invention> In order to eliminate these drawbacks, the present invention identifies and extracts the mark area of a document based on the difference in mark and color information, and creates abstract information from the information of the mark area. υ, this summary information and corresponding document information are stored in pairs. Therefore, only the summary information can be read and displayed, and it can be quickly determined whether the information is necessary or not, and the search can be performed quickly. Furthermore, the summary information is created automatically and does not require any human intervention.

<Embodiment> FIG. 2 shows an embodiment of this invention. The document 11, such as a document lifespan, is scanned by the image information input section 14 and input as image information. The image information is input to a discrimination processing section 15 that discriminates the area. Discrimination processing unit 15, extraction processing unit 16, summary creation unit 17, memory 18 in which document images and programs that perform storage processing are temporarily stored, central processing unit (CPU) 19, magnetic disk in which document images are actually stored. and a storage unit 21 such as an optical disk;
The document image display device 22 and the image output device 23 are connected to the common bus 2.
Connected to 4.

The image input unit 14 is an input device such as a television camera or a facsimile, and the input image signal is sent to the discrimination processing unit 15, where the image signal is centered on the part of the printed document 11 marked with a fluorescent pen. A certain area is carved out. For example, as shown in Figure 3, "1 purpose" and "L
If a mark is attached to "ED Printer", the discrimination processing unit 15 divides an area 25 and 26 including the vicinity of the mark. In the actual discrimination process, a sensor for color information may be used in the case of a facsimile, and a color filter may be used in the case of a color camera.

Specifically, while scanning the entire document 11 with a color camera or the like, the positional coordinates (x, y) of 1 minute in the document marked with a fluorescent pen are sampled, and that area is first extracted. Thereafter, a certain area around the area drawn with the fluorescent pen is extracted and stored in the memory 18 based on the area 25, 26 and the entire document image. This area 2
5.26 extraction can be realized by software. 9. The part marked by scanning the document 11 with a color camera, ? When the output is obtained, the X and Y coordinates of the marked area are detected from the scanning signal at that time, and the X component is x1 to X2, and the Y component is y, to y.
If it is 2, expand △X and △y horizontally and vertically in the X and Y directions, respectively, and xl - △x - X2 +
The area limited by △X and yl - △y - y2 + △y is defined as an area divided by the discrimination processing unit 15 (for example, 26), and the image information contained in this area 26 is extracted from the entire image information as summary information. taken out.

Determined area 2 of document image stored in memory 18
5.26, the cutout processing unit 16 cuts out the character string area and deletes half-finished characters. That is, specifically, the peripheral distribution of areas 25 and 26 in the memory 18 is examined in the vertical and horizontal directions, and a character string area is cut out from the peripheral distribution and half-finished characters are deleted. A specific example is shown in FIG. That is, for example, in the region 26, find the maximum value of the image level in each point of xl-△x-x2+△X for each scanning line in the lateral force direction, that is, for each y value of Dy+-△Y42+△y. Then, as shown in FIG. 4A, the portion corresponding to line 1 27 of the document becomes zero. The point of each central note in each zero region is ¥1. Take out the amount within 2 yen, divide its outer value by υt, and examine the image shown in Figure 4B. Same (as a brush, x in area 26, 1-△x-
For each straight line of X of x2+△X, y, 7△Y”-Y2
For each of +△y, find the maximum a-direction in -, respectively.
You will be asked if it is the distributed force ball shown in Figure 4C.

From this, @6 minutes corresponding to space 28 is detected, and x1 to x
Take out the inside of 2. As a result, a half-finished portion as shown in the fourth diagram is removed.

The mark area 29 of the bright light pen in the document surface 1 image created by the above processing is inputted into the memory 18 again,
In that case, the area 26 (25) is divided by 1lIJ.

The area 29 created by the cutout processing section 16 is then aggregated by the summary creation section 17) to complete a summary version for one page. A summary of one page is stored in memory 18. In that case, it may be possible to perform character recognition processing and encode the summarized version. The information of the summarized document image stored in the memory 18 and the corresponding original image (the entire binary information) are stored in the storage section 21 through the common bus 24. At that time, the summary information is stored so that it is read out first. FIG. 5 shows a schematic flow of the above processing.

When searching for stored document images, first, only the summarized version of the registered document image is displayed on the display device 22, and it is determined whether or not it is necessary. If necessary, a condensed version follows,
Original image (including its summarized version if necessary) output device 2
3. Output as a hard copy.

<Effects> As explained above, the document information registration and retrieval device of the present invention automatically extracts information from a portion of a printed document image marked with a fluorescent pen or the like, and uses that information to extract a summary of the printed document. It is possible to automatically create a version (abstract).

This has the advantage of shortening and simplifying the time required to create an abridged version. Even if the document information is still being searched, by using the summarized version of the information, the search time can be shortened and the operation can be simplified.

In addition, by changing the discrimination processing section, it becomes possible to extract a portion of a document image based on various marks.

[Brief explanation of the drawing]

FIG. 1 is a diagram showing an example of adding a mark to an existing document image using a fluorescent pen, FIG. 2 is a block diagram showing an embodiment of the present invention, and FIG. FIG. 4 is a diagram for explaining the processing in the cutout processing section, and FIG. 5 is a flowchart showing the order of the overall processing. 11: Document, 14: Image information input section, 15: Discrimination processing section, 16: Cutting processing section, 17: Abstract creation section, 18: Document image storage/program storage memory, 19: Central processing unit ( CPU), 21: Document image storage unit, 22: Document image display device, 23: Document image output device, 24: Common bus, 26: Area cut out in area discrimination processing, 29: Area cut out in cutting processing . Patent Applicant: Representative of Nippon Telegraph and Telephone Public Corporation Takashi Kusano No. 1 Figure 2 Figure 11 no Random Z O 3 Figure 4 Figure 5 (¥1

Claims (1)

[Claims] (1) An input unit that reads document information as an electrical signal, a discrimination processing unit that discriminates a marked area in the document from the read information, and a character string area is cut out from the discriminated area. an extraction processing unit, a summary creation unit that creates summary information of the document from the extracted information, a storage processing unit that stores the read document information and the created summary information, and each such document information. and a storage section for storing a plurality of sets of summary information, a document image information display section for reading out and displaying the information in the storage section, and a document image output section for reading out the information in the storage section and outputting it as a hard copy. A document information registration and retrieval device comprising: