JPH1125256A - Method for preventing input error in data base system using image-recognized result as input data - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a method for preventing an input error in a data base system in which an image-recognized result is used as input data. SOLUTION: This method comprises an image fetching device 1, an image processor 2, a data base device 3, and a sheet on which a mark is put on any one part or plural parts of selection items. The sheet is set in the image fetching device 1, and the specific area of the fetched image is analyzed for recognizing the presence or absence of the mark in a given shape so than data can be extracted, and the extracted result can be used as input data. In this case, when the sheet is not set, the image fetching device 1 is operated, and an incorrect recognition factor attached to the image fetching device 1 is preliminary detected by the image processing, and when it is judged that the position of the detected incorrect recognition factor is a position where a recognition error is likely to be generated, warning is issued.

Description

DETAILED DESCRIPTION OF THE INVENTION

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for preventing an input error in a database system using image recognition results as input data.

[0002]

2. Description of the Related Art Conventionally, as a database system which extracts characters, figures, and marks from image data by image processing means and uses them as input data, there are the following systems.

One of them is a mark sheet system which is generally used for recognizing a mark at a position of a sheet. This is to illuminate a predetermined position on the sheet with light, thereby recognizing the presence or absence of a mark on the sheet.

In this method, in order to adjust the light irradiation position to a target position, it is necessary to mechanically align the direction of the sheet and not to allow the sheet to be displaced. Therefore, the scale of the system must be increased, and the cost of introducing the equipment is high. In addition, in the case that the mark sheet is filled out, there is a drawback in that since the shape of the mark differs depending on the person, a sample example must be displayed to show a good example and a bad example.

[0005] As a second method, an OCR (Optical Character Reorder) system using a fax has been partially put into practical use. In this method, a slip generated in an inventory management business and a product ordering business is sent by facsimile, and the terminal that receives it recognizes the content by OCR and registers it in a database. This system has a drawback in that if dust or the like is attached to a sheet when a fax is transmitted from the fax machine, the receiving terminal cannot distinguish the dust from characters, numerals, and marks. In addition, even if there is no garbage, the check frame is often represented by a square figure, and thus the entered mark has various shapes, and the mark may not be recognized. In many cases, a mark is written out of the frame.

[0006]

In the above-mentioned mark sheet system, the equipment becomes large as described above, and the entry example of the mark must be displayed. Therefore, the OCR system using a fax cannot distinguish between garbage and data. Sometimes, it was difficult to recognize the mark.

SUMMARY OF THE INVENTION The present invention has been made to solve the above-mentioned problems, and an object of the present invention is to provide an image recognition result capable of recognizing correct data as written on a sheet. An object of the present invention is to provide a method for preventing an input error in a database system as input data.

[0008]

In order to solve the above-mentioned problems, according to the present invention, an image capturing device, an image processing device, a database device, and one or more selected items are marked. It consists of paper with
In a database system, the paper is set in an image capturing device, data is extracted by means for analyzing a specific area of the captured image and recognizing the presence or absence of a mark of a predetermined shape, and the extracted result is used as input data. When the image capture device is not set, the image recognition device is operated to detect in advance the erroneous recognition factors attached to the image capture device by image processing, and the position of the erroneous recognition factor detected by the determination means is determined to be a position where a recognition error occurs. Warn if done.

According to the second aspect of the present invention, when an erroneous recognition factor is recognized, its shape and size are automatically registered in a database, and the contents of the database are stored in the form of a mark recognized from a newly set sheet. It is compared with the candidate, and if it is determined that the cause of the misrecognition, a warning is issued.

[0010] According to a third aspect of the present invention, the image capturing device, the image processing device, the database device, and a sheet for marking any one or a plurality of selected items are set, and the sheet is set in the image capturing device. Then, data is extracted by means for analyzing a specific area of the captured image and recognizing the presence / absence of a mark having a predetermined shape. In a database system in which the extracted result is used as input data, the result recognized by the image processing device is If it is determined that there is a possibility that a part of the mark is out of the recognition area, it is moved in a direction in which there is a possibility that the part is out of the recognition area so that the entire mark can be recognized.

According to a fourth aspect of the present invention, there is provided an image capturing device, an image processing device, a database device, and a sheet for marking any one or a plurality of selected items, and the sheet is set in the image capturing device. Then, data is extracted by means for analyzing a specific area of the captured image and recognizing the presence / absence of a mark having a predetermined shape. If the mark is not recognized, it is determined that the paper is in the opposite direction or the paper is out of the image capturing area, and a warning is issued.

[0012] According to the fifth aspect of the present invention, the image capturing device, the image processing device, the database device, and a sheet for marking one or a plurality of selected items are set, and the sheet is set in the image capturing device. Then, data is extracted by means for analyzing a specific area of the captured image and recognizing the presence / absence of a mark of a predetermined shape, and in a database system using the extracted result as input data, two or more different locations on paper are used. Marking two or more marks in a specific position in advance, and detecting the position shift of at least one mark at the time of image capture and determining that it is not possible to recognize the contents of the paper, the entire captured image Is automatically moved or rotated to make a correction to a recognizable position.

According to the present invention, the frame for checking the sheet is a frame graphic having a predetermined shape.

[0014]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS One embodiment of the present invention will be described with reference to FIGS. 1 is a configuration diagram of a hardware device, FIG. 2 is a diagram showing a flow for preventing an input error, FIG. 3 is an explanatory diagram of a method of changing a reading area, and FIG. 4 is an outline of a transcription work support system for hospitals. FIG. 5, FIG. 5 shows a meal card, FIG. 6 shows a warning screen 1 at the time of identification of an erroneous recognition factor, FIG. 7 shows a warning screen at the time of a paper reading error, and FIG. It is a figure showing warning screen 2 of.

An input error prevention method in a database system using image recognition results as input data has the following configuration.

According to the first aspect, an erroneous recognition factor detecting means for detecting an erroneous recognition factor attached to the image capturing device 1 when a sheet is not set, and a recognition error occurs in the position of the detected erroneous recognition factor. There is provided a determination unit for determining whether the position is at the position, and a warning unit for generating a warning when it is determined that the position is at a position where an error occurs.

According to a second aspect of the present invention, a means for registering a detected misrecognition factor, a mark candidate recognized from a newly set sheet, a collating means for collating the registered content, and a factor for the misrecognition. Means for determining, and a warning means for issuing a warning when it is determined to be a cause of erroneous recognition.

According to claim 3, means for judging whether or not the mark is out of the area to be recognized,
Means for changing the recognition area when it is determined that the mark is out of the area to be recognized.

According to a fourth aspect of the present invention, there is provided means for judging whether or not the sheet arrangement is normal based on whether or not the sheet has a predetermined mark and whether or not the mark can be recognized.

According to the fifth aspect, the paper is at least 2
Image correction means having two marks, and rotating and moving the image so that the mark on the captured image is at a predetermined position.

According to a sixth aspect of the present invention, there is provided a frame graphic having a predetermined shape which allows a recognition means to recognize a frame for checking a sheet and enables a writer to recognize an entry method from the frame graphic.

It consists of: Hereinafter, description will be given with reference to the flowchart of FIG. 2 showing a flow for preventing an input error.
Note that the numbered lines in FIG. 2 indicate each step or state,
And a database.

In accordance with the first aspect of the present invention, in step (4), the image capturing device 1 determines whether or not a sheet is set, and when no sheet is set (state 5). Then, a blank reading is performed in the image capturing device 1 (step 6). Then, the captured image is subjected to a labeling process (step 7) to extract an erroneous recognition factor (garbage). Everything labeled in the labeling process is a cause of misrecognition.

Next, the center of gravity of the misrecognition factor recognized in (Step 9) is calculated. Further, of the images captured by the image capturing device 1, a small area for operating the image for checking the presence or absence of a mark by the image processing device 2 is registered in the reading region database 8 in advance, and the center of gravity of the erroneous recognition factor is set. Is compared with the reading area, and if the position of the center of gravity is within the area, a warning is issued because there is a possibility of erroneous recognition in (Step 10) (Step 11).

According to the second aspect of the present invention, in step (9), the position of the center of gravity of the erroneous recognition factor is compared with the read area. Of the shape and size of the object
2) is automatically registered (step 13). Specifically, the coordinates of each pixel of the recognized erroneous recognition factor are converted into relative coordinates from the pixel closest to the center of gravity, and all the converted coordinates are registered. The size is an area, and is the number of all pixels. When the marked sheet is set in the image capturing device 1 and the captured image is recognized as a candidate for the mark (when the image passes from the state 14 to the state 15), the paper is registered in the erroneous recognition factor database (12). The content is always collated with the mark candidate (step 15). Pattern matching is used as a matching method. Database (1
When the misrecognition factor extracted from 2) is moved and rotated little by little, and completely matches the mark candidate, the mark candidate is determined to be a misrecognition factor and a warning is issued (step 16). .

According to the third aspect of the present invention, as a result of performing the labeling process on the read area registered in the read area database 8 (step 18), the center of gravity 19 of the label shown in FIG. If it is located in the end area 21 outside the range 20, the center of the reading area is translated in the direction in which the center of gravity is present (step 22), and recognition is performed again. When moving,
The center of the frame moves in one of the directions of arrow 23 when the frame is square and in the direction of arrow 24 when the frame is V-shaped.

According to the fourth aspect of the present invention, a mark for recognizing the direction of the sheet is previously marked at a specific position on the sheet, and the mark set in the reading area database 8 may exist. An area is operated and recognized by a labeling process (step 25). The area of the expected mark is registered in the reading area database 8 and if the area of the expected mark matches the area of the recognized label,
It is determined that the orientation of the paper is correct, and other processing is continued (step 15). If there is no label as a result of the recognition, or if the area of the expected label and the area of the recognized label are different, it is determined that the paper direction is reversed or the paper is largely outside the image reading area ( Step 2
7) and a warning is issued (step 2).
8).

According to the fifth aspect of the present invention, two or more marks are previously marked at different specific positions on the paper, and the coordinates of the center of gravity of the marks recognized by the labeling process and the reading area database 8 are used. The center of gravity coordinates of each mark when the paper is set ideally with no set displacement are compared (step 29). If at least one of the center of gravity coordinates is different, the entire captured image is automatically converted. Correction to a recognizable position is performed by moving and rotating (step 30).

Corresponding to the invention of claim 6, the frame for checking the paper is made into a frame figure of a predetermined shape, so that it is urged to input a predetermined mark at the time of entry.

[0030]

An embodiment of the present invention will be described with reference to FIGS. FIG. 4 is a diagram showing an outline of a transcription work support system for hospitals. In this embodiment, the task of a nurse transferring a work card 31 issued from each medical office to a medical chart is automated. This system includes a personal computer 32, an image scanner 33, a printer 34, and a database 35. The work card 31 shown in FIG.
It has a function of analyzing the image with OCR software, converting the description into characters, numbers, and marks from the image, and registering it in the database 35. Further, the chart 35 can be easily created by editing the data in the database 35. Transcription errors can be reduced compared to handwritten transcription work, and transcription work and data search work can be significantly improved.

Hereinafter, as an example, the contents of a card (see FIG. 5) for changing the contents of a meal, which is called a meal sheet at the time of hospitalization, after surgery, or after an examination, will be transcribed to a chart. FIG.
Is marked at a predetermined position (in the V-shaped frame) on the sheet of the meal sheet shown in FIG. Also, black circles 37 and 38 are attached to the paper in two places. The upper mark 37 serves both as a mark for specifying the direction of the sheet and a mark for correcting the sheet displacement. The lower mark 38 is a second mark for correcting the sheet displacement.

First, before scanning the paper, the image scanner 33 is read once without placing the paper in the image scanner 33 in advance, and if there is any dust and the position may be erroneously recognized, Then, a warning screen as shown in FIG. 6 is displayed. In this case, the image scanner 33 is to be cleaned. All recognized garbage is registered in the misrecognition factor database 12.

When inputting data, a sheet is set on the image scanner 33 as shown in FIG. 4, and a GUI (graphic user interface) button 39 is pressed on the screen of the personal computer 32 to take in an image of the sheet. The means for determining whether or not the paper orientation is correct displays a warning screen as shown in FIG. 7 if the paper is upside down or the position on the image scanner 33 is greatly shifted. In this case, it is necessary to get the paper at the correct position.

When the paper image is taken in the correct orientation,
This time, the displacement of the paper is corrected. When placing paper on the image scanner 33, it is not always possible to place the paper at an ideal position. Normally, the paper is inclined or the lid 4 of the image scanner 33 is closed.
When 0 is closed, it moves in parallel. In order to correct this shift, it is necessary to match the read coordinates of the read area database 8 with the coordinates to be read of the captured image.

When a person is to have a mark, there are various kinds of quirks depending on the person, and the position where the mark is marked on the frame naturally differs. If a mark is placed near the edge of the frame, the entire mark will not fit in the reading area.

At present, when the area of one label in the reading area exceeds a certain threshold value, the label is recognized as a mark. In this case, the label is recognized as a mark. Can not.

In order to compensate for this error, if the barycenter position of the label in the range recognized in the reading area is located at the edge of the frame, the reading area is changed to change a part of the mark. Can be recognized even when is outside the area (frame).

The mark recognized through the above process is compared with the erroneous recognition factor database 8. For this, pattern matching is used. If it is determined by the pattern matching that there is dust, a warning as shown in FIG. 8 is generated.

The last recognized mark is registered in the patient management database 35.

[0040]

According to the first aspect of the present invention, an erroneous recognition that dust is erroneously recognized as a mark at the time of image recognition is performed by removing an erroneous recognition factor attached to the image capturing device in advance. This has the effect of preventing recognition.

According to the second aspect of the present invention, it is possible to prevent an erroneous recognition that dust is erroneously recognized as a mark during image recognition by removing an erroneous recognition factor attached to a sheet. It has the effect of being able to.

According to the third aspect of the present invention, even if a mark is put on the end of the predetermined frame, the mark can be recognized and erroneous recognition can be prevented.

According to the fourth aspect of the invention, it is possible to prevent the image processing apparatus from recognizing erroneous information (mark) when the paper is read in the reverse direction or when the paper is largely displaced. This has the effect.

According to the fifth aspect of the invention, by correcting the position of the sheet placed on the image capturing device, it is possible to prevent the image processing device from recognizing erroneous information (presence or absence of a mark). This has the effect that it can be performed.

According to the sixth aspect of the present invention, a mark of a specific shape can be made conscious of a mark to be marked, and a mark of the specific shape can be filled, thereby preventing erroneous recognition. . In addition, there is no need to show an entry example, and the entry space can be saved.

[Brief description of the drawings]

FIG. 1 is a hardware configuration diagram of an embodiment of the present invention.

FIG. 2 is a diagram showing a flow for preventing an input error in the above.

FIG. 3 is an explanatory diagram of a method of changing a reading area according to the first embodiment.

FIG. 4 is a diagram showing an outline of a transcription work support system for a hospital according to the first embodiment.

FIG. 5 is a view showing a meal card of the above.

FIG. 6 is a diagram showing a warning screen at the time of identifying an erroneous recognition factor according to the first embodiment;

FIG. 7 is a diagram showing a warning screen when a paper reading error occurs in the embodiment.

FIG. 8 is a diagram showing a warning screen at the time of identifying an erroneous recognition factor.

[Explanation of symbols]

 Reference Signs List 1 image capturing device 2 image processing device 3 database device 8 reading area database 12 misrecognition factor database 31 paper 32 personal computer 33 image scanner 34 printer 35 database 36 chart 37 mark 38 mark 39 GUI button 40 lid of image scanner

 ──────────────────────────────────────────────────続 き Continuing on the front page (72) Inventor Brendan-Joseph Mulan 1048 Kadoma Kadoma, Kadoma-shi, Osaka Matsushita Electric Works Co., Ltd. In company

Claims (6)

[Claims] 1. An image capturing device, an image processing device, a database device, and a sheet for marking any one or a plurality of selected items. The sheet is set in the image capturing device, and the captured image is In a database system in which a specific area is analyzed and data is extracted by means for recognizing the presence or absence of a mark having a predetermined shape, and the extracted result is used as input data, an image capturing device is operated when paper is not set and an image is captured. Image recognition characterized in that an erroneous recognition factor attached to a capture device is detected in advance by image processing, and a warning is issued when the position of the erroneous recognition factor detected by the determination means is determined to be a position where a recognition error occurs. An input error prevention method in a database system in which results are used as input data. 2. The input error prevention method according to claim 1, wherein when an erroneous recognition factor is recognized, its shape and size are automatically registered in a database, and the contents of said database are read from newly set paper. An input error prevention method in a database system using image recognition results as input data, characterized in that the image recognition results are compared with recognized mark candidates, and a warning is issued when it is determined as a cause of erroneous recognition. 3. An image capturing device, an image processing device, a database device, and a sheet for marking any one or a plurality of selected items. The sheet is set in the image capturing device, and the captured image is In a database system in which a specific area is analyzed and data is extracted by means for recognizing the presence or absence of a mark having a predetermined shape, and the extracted result is used as input data, a part of the mark is recognized as a result of recognition by the image processing apparatus. If it is determined that there is a possibility that the image is out of the database, the image recognition result is moved to a direction in which the image may be out of the recognition area so that the entire mark can be recognized. How to prevent input errors in the system. 4. An image capturing device, an image processing device, a database device, and a sheet for marking any one or a plurality of selected items, the sheet is set in the image capturing device, and the Data is extracted by means for analyzing a specific area and recognizing the presence / absence of a mark of a predetermined shape. In a database system using the extracted result as input data, a predetermined position on a sheet is marked in advance, and the mark is Is not recognized, it is determined that the paper orientation is reverse or the paper is out of the image capturing area, and a warning is issued. Input error prevention method. 5. An image capturing device, an image processing device, a database device, and a sheet for marking any one or a plurality of selected items. The sheet is set in the image capturing device, and the captured image is read. In a database system in which data is extracted by means of analyzing a specific area and recognizing the presence or absence of a mark having a predetermined shape, and the extracted result is used as input data, two data points are previously stored at two or more different specific positions on a sheet. With the above markings on, if the position of at least one mark is detected during image capture and it is determined that the contents of the paper cannot be recognized, the entire captured image is automatically moved or moved. A method for preventing input errors in a database system using image recognition results as input data, characterized by performing correction to a recognizable position by rotating. Law. 6. The input error preventing method according to claim 1, wherein the check frame of the sheet is a frame graphic having a predetermined shape, wherein the image recognition result is input data. Prevention method.