JPS5820992Y2 - Kanji document search device - Google Patents

Description

[Detailed Description of the Invention] The present invention relates to a search device for kanji documents using a surname or a corporate name (hereinafter referred to as a name) as a keyword.

More specifically, the present invention relates to a kanji document retrieval device suitable for kanji document management such as bills, checks, customer and specific person business management, etc.

Currently, in kanji document processing, kanji documents are generally stored in original documents and searches are performed manually.

As an advanced process, there is a method in which the entire text of a Kanji document is registered in Kana and the search is performed using the Kana character name, or a Kanji processor is used to register the document and then use Kanji directly in the search function.

However, when using kana characters, the recall rate is high but the precision rate is low.

When using the Toda kanji, both recall and precision are high, but
The disadvantage is that it takes too much time to register documents and input kanji keywords.

Even when registering kanji documents, inputting kana or kanji codes has the disadvantage that input errors occur and the registration process takes too much time and money.

Furthermore, there is a lot of waste and waste of time and manual intervention such as microfilm processing and development processing.

An object of the present invention is to provide a kanji document retrieval device that solves the above-mentioned drawbacks of the prior art and facilitates registration and retrieval.

Therefore, in the present invention, a Kanji document is registered as an image in a video file, and when searching for it, a name originally composed of Kanji characters, that is, a keyword, is performed in Kana characters.

When a kana character keyword is input, the corresponding kanji notation group is displayed on the display device, and the correct notation is specified from the displayed kanji notation group using a suitable means such as a light pair.
Based on the specified kanji notation, the corresponding kanji document is read from the video file and displayed on the display device.

According to the present invention, since the kanji document is filed as video information in a video file, registration can be performed using a television camera, facsimile, etc., so erroneous input can be prevented and registration can be performed easily.

Furthermore, since the method uses a method of displaying a group of kanji notations and specifying the correct notation by inputting a kana character keyword, it is possible to search for kanji documents with the same effect as inputting a kanji keyword, that is, with a high recall rate and precision rate.

Next, the present invention will be explained in detail with reference to one embodiment.

FIG. 1 is a block diagram showing one embodiment of the present invention, KIZIJ.

In the figure, 1 is a Kanji document to be processed, and the Kanji document 1 is imaged by a television camera 2 and stored as video information in a video information memory 4 via a buffer memory 3.

The buffer memory 3 stores video signals captured by the camera 2,
It has a function of temporarily storing video signals read out from the video information memory 4.

The address at which the video signal is stored in the video information memory 4 is specified by the register 21 of the central processing unit 6.

The address at which the video signal is read from the video information memory 4 is specified by the register 22 of the central processing unit 6.

The central processing unit 6 further includes n register groups 23-1 to 23-1.
23-n is provided.

This register group 23 is for processing codes of kana character names which will be explained in detail later, and although a plurality of registers are required, it is usually sufficient to provide about 10 registers.

A keyboard 5 is connected to the central processing unit 6, through which data such as the start and end of processing, as well as names in kana characters, are input.

As memories, a search information memory 7, a video information memory, and a reference information memory 13 are further provided.

The search information memory 7 stores index information that combines the kana character name input from the keyboard 5 and the address of the video signal of the document in the video information memory 4.

The video information memory 11 stores video signals for each screen of a kanji notation group corresponding to a kana character keyword.

Further, the reference information memory 13 stores data in which the code corresponding to each kanji notation in the video information memory 11 is compared with the screen number and the display position on the screen, and the register 21 of the central processing unit 6. Stores the address data to be set to 22.

The video combining device 10 includes a buffer memory 3 and a central processing unit 6.
is connected to the TV monitor 9 to simultaneously display various images.

The light pen 8 selects a selection on the screen of the television monitor 9 and sends a signal to the central processing unit 6.

Randomizing device 12 connected to central processing unit 6
is a device for determining the address of the search information memory 7 from the contents of the register group 23.

Next, the operation will be explained.

When the kanji document 1 is collected, the line 106 from the keyboard 5
A processing start signal is sent to the central processing unit 6 via.

In response to this processing start signal, the central processing unit 6 reads the start address signal of the empty area of the video information memory 4 from the reference information memory 13 via the line 114, and reads the address signal.
Set in register 21.

When this address signal is set in the register 21, a set completion signal is sent to the buffer memory 3 via the line 103, and the video signal of the kanji document 1 is sent through the camera 2 to the address in the video information memory 4 indicated by the register 21. Process 1-1 is stored.

Each time one Kanji document 1 is processed, address register 2
The contents of 1 are counted up one by one, and the kanji document information is read one after another.

When reading of the kanji document 1 is completed, a reading sales signal is inputted from the keyboard 5 to the central processing unit 6, and the contents of the address register 21 are saved in the reference information memory 13.

Next, the start address signal of the unprocessed record in the video information memory 4 is similarly read out from the reference information memory 13 and set in the address register 22.

The video signal of the kanji document is read out from the video information memory 4 according to the address signal of the register 22, and the video combination device 1
0 on the monitor television 9.

This is the upper half 31 of the screen of the monitor television 9 shown in FIG.
will be displayed.

When the user looks at this screen, it may be the name part 34 of the kanji document.
The user inputs the kana reading keyword of the name using the keyboard 5.

This kana character keyword is displayed in the center 32 of FIG. 2 by the video combining device 10.

When a space signal is sent from the keyboard 5, the screen 32
The kanji notation group corresponding to the kana displayed in is read from the video information memory 11 and displayed in the first part 33 of FIG. 2 by the video combining device 10.

The user uses the light pen 8 to select the same kanji notation as the name part 34 from the kanji notation group 33 displayed on one part of the monitor television 9, and transmits the position signal to the central processing unit via the line 108. 6, and the codes corresponding to the kanji notation are sequentially set in the register group 23 from the reference information memory 13.

FIG. 3a shows a specific display example.

The name of the kanji document here is "Asahi Kogyo Co., Ltd."
The kana name displayed on the kana display section 32 is also input using the keyboard 5.

And in part 1 33, the kanji notation corresponding to "Asahi" is written "
"Asahi" and "Asahi" correspond to "Kogyo" in "Industrial"
and ``mining'', and ``corporation'', which corresponds to ``Kabushiki Gaisha'', are displayed.

Looking at the screen of this monitor television 9, the user uses the light pen 8 to write the correct kanji characters "Asahi", "Kogyo" and "
Ltd.".

FIG. 3 shows how the codes corresponding to the selected kanji notation are set in the registers 23-1 to 23-3.

When all the codes for one name have been set in the register 23, the contents of the register 23 are input to the randomizing device 12 to obtain the address of the search information memory 7.

This situation is shown in FIG.

When this address information is input to the central processing unit 6 via the line 113, this address information, the kana character keyword displayed on the screen 32, the kanji correspondence code in the register 23, and the video information set in the register 22. Index information as shown in FIG. 5 is created based on the address of the memory 4.

This index information is stored in the search information memo IJ 7 using the address information on line 113.

When this process is completed, the contents of the address register 22 are counted up by one, the video signal of the next kanji document in the video information memory 4 is set on the monitor television 9, and the same process is performed.

Address, the contents of register 22 are address register 2
If the content of address register 22 becomes a dog, the registration process ends and the content of address register 22 is temporarily saved in reference information memory 13.

To search for a video signal of a registered kanji document, the name of the object to be searched for is input in kana characters from the keyboard 5.

This kana character keyword is displayed on the monitor television 9 by the video combination device 10, and at the same time, the kanji notation group corresponding to this kana is also read from the video information memory 11 and displayed on the monitor television 9 by the video combination device 10.

When a target kanji notation to be searched for is selected from a group of kanji notations using the light pen 8, the code of the corresponding kanji notation is obtained from the reference information memory 13.

These codes are sequentially set in the register 23, and when all the codes for one name have been set, the contents of the register 23 are input to the randomizing device 12 to obtain address information of the search information memory 7.

Based on this address information, the search information memory 7 is searched to obtain the corresponding index information (FIG. 5).

Using the address of the video information memory 4 included in this index information, the video signal of the corresponding Kanji document is obtained from the video information memory 4 and displayed on the monitor television 9.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing an embodiment of the present invention, FIG. 2 is a diagram showing the screen configuration of the monitor television shown in FIG. 1, FIG. 3a is a diagram showing a specific display example of the monitor television, Figure 3 is a diagram showing a register in which a specific code corresponding to the kanji notation displayed from the monitor TV screen in Figure 3a is set, and Figure 4 is for explaining the randomizing device in Figure 1. and FIG. 5 are diagrams showing the structure of search information stored in the search information memory of FIG. 1. 1... Kanji document, 2... Camera, 3... Buffer memory, 4... Memory for video information, 5... Keyboard, 6... Central processing unit, 7... For search information Memory, 8...Light pen 9...Monitor TV 10.
...Video combining device, 11...Video information memory, 12
... Randomizing device, 13'... Memory for reference information.

Claims (1)

[Scope of utility model registration request] A first means for holding a kanji document as video information, and a second means for holding a group of kanji notations corresponding to kana characters as video information.
a third means for associating the kanji notation group with a code; a fourth means for holding search information including a kanji correspondence code corresponding to a kana name and an address of the first means; It has a kana character input means for inputting a kana name, and a button and an image display means, and the image display means reads a kanji notation group corresponding to the kana name inputted from the input means from the second means. , -display, obtain a code corresponding to the designated kanji notation group from the displayed kanji display group from the third means, read out the search information from the fourth means based on this code, and perform this search. A kanji document retrieval device characterized in that a kanji document corresponding to a kana name input from the address of the first means that information has is read from the first means and displayed on the video display means.