JP2003016384A - Optical character, recognition device and display method for recognition result picture - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an optical character recognition device, a display method for a recognition result correction picture, and a display method for a recognition result confirmation picture, which make it possible to speedily and easily do correcting operation, confirming operation, etc., for a recognition result with high reliability, suppress increases in processing load and storage capacity as much as possible, and can speed up processing with economically advantageous constitution. SOLUTION: When a rejected character is present in a recognition object field, a recognition processing part 13 stores its rejected character pattern in a recognition result information storage part 14 in the form of a multi-gradation image and informs a recognition result correction part 15 of the rejection. The recognition result correction part 15 once informed displays, at a recognition result output part 16, a recognition result picture for rejection correction having an original image display area where the rejected character pattern is displayed as the multi-gradation image instead of the normal recognition result picture.

Description

Detailed Description of the Invention

[0001]

TECHNICAL FIELD The present invention relates to a gray scale,
The present invention relates to an optical character recognition device that performs character recognition by inputting a multi-gradation image such as a color image, and a method of displaying a recognition result screen in the optical character recognition device.

[0002]

2. Description of the Related Art Conventionally, in an optical character recognition device, there is JP-A-4-34671 (pattern recognition device) as a technique for confirming and correcting the recognition result of pattern recognition. This device controls the screen display for the pattern and the recognition result when performing the confirmation and correction for the recognition result obtained by performing the pattern recognition processing on the input image to ensure the confirmation and correction work and improve the efficiency.

In the recognition device according to the above-mentioned prior art, since the correction screen of the recognition result is displayed as a binary image, a partial loss of information is likely to occur, and the visual recognition of the operator is made. However, there is a high degree of occurrence of character patterns that cannot be read due to faintness, crushing, etc., and character patterns that are misread, and there is a problem in terms of workability and reliability.

Therefore, in order to improve the visibility of the operator, a device has been developed which displays a correction screen in color, grayscale, etc. as an input image in color, grayscale, etc. Since the image is displayed in gray scale, color, etc., a huge storage capacity is required to store the image data, and therefore a high-speed, large-capacity storage device is required, so the system configuration is enlarged. In addition, the cost is high, the amount of data to be handled is significantly increased, and accordingly, the processing load of image processing and display processing is significantly increased, so that speedup cannot be expected. .

[0005]

As described above, since the conventional character recognition device is configured to display the correction screen of the recognition result as a binary image, partial information loss is likely to occur, and Even in the visual recognition of the operator,
There is a high degree of occurrence of character patterns that cannot be read due to crushing or that are misread, which is a problem in terms of workability and reliability. In addition, a device that displays a correction screen in color, grayscale, etc., displays the entire image in grayscale, color, etc., so a high-speed, large-capacity storage device is required, and the system configuration becomes large and expensive. In addition, the amount of data to be handled increases remarkably, and accordingly the processing load of image processing and display processing also remarkably increases, so that there is a variety of problems in terms of practicability such as high speed cannot be expected.

The present invention has been made in view of the above circumstances,
It is possible to perform recognition result correction work and confirmation work quickly and easily with high reliability, suppress the increase in processing load and storage capacity as much as possible, and speed up processing with an economically advantageous configuration. An object of the present invention is to provide an optical character recognition device and a method of displaying a recognition result screen.

[0007]

According to the present invention, when the information of the recognition result is saved, only the character pattern of the reject character is saved as the information of the multi-gradation image, and the other information is saved as the binary image. Then, on the correction screen of the recognition result, only the character pattern of the rejected characters is displayed as a multi-gradation image, and the other information is displayed as a binary image. As a result, it is possible to easily and speedily perform the operation of correcting the reject character with an economically advantageous configuration that suppresses an increase in processing load and storage capacity as much as possible.

That is, the present invention is, in an optical character recognition device, an image acquisition means for acquiring a multi-gradation image, a recognition processing means for performing character recognition of the image acquired by the image acquisition means, and the recognition. A recognition result information storage unit that stores the rejected character pattern of the result recognized by the processing unit as a multi-gradation image, and stores the character pattern of the entire recognition result as a binary image,
A recognition result screen using a character pattern stored in the recognition result information storage means, and a graphics user interface means for displaying the same, and the reject character is multi-gradated on the recognition result screen for reject correction. And a normal recognition character is displayed as a binary image.

Further, the present invention is, in an optical character recognition apparatus, an image acquisition means for acquiring a multi-gradation image, a recognition processing means for character recognition of the image acquired by the image acquisition means, and the recognition. Of the character patterns recognized by the processing means,
A recognition result information storage unit that stores a character pattern of a predetermined specific area in a multi-gradation image, and stores at least another character pattern in a binary image, and a character pattern stored in the recognition result information storage unit. And a graphic user interface means for creating and displaying a recognition result screen using the recognition result screen, the entire recognition result is displayed as a binary image, and the character pattern of the specific region is displayed in multiple gradations. It is characterized by doing.

Further, the present invention is a method of displaying a recognition result screen in an optical character recognition device, wherein character recognition is performed on a read multi-gradation image, and the character pattern of the entire recognition result is binarized. Stored as an image, the reject character is stored as a multi-gradation image, and the character pattern of the entire stored recognition result is displayed as a binary image on the screen displaying the recognition result,
The stored rejected character pattern is displayed as a multi-tone image.

Further, the present invention is a method of displaying a recognition result screen in an optical character recognition device, wherein character recognition is performed on a read multi-gradation image, and the character pattern of the entire recognition result is binarized. It is stored as an image, and the character pattern of a specific area specified by format control etc. is stored as a multi-gradation image, and the character pattern of the entire stored recognition result is displayed as a binary image on the screen for checking the recognition result. And displaying the character pattern of the specific area in a multi-gradation image.

By applying the above-described configuration or method of the present invention, it is possible to perform the recognition result correction work, confirmation work, and the like with high reliability, quickly and easily, and increase the processing load and storage capacity as much as possible. In addition, the processing can be speeded up with an economically advantageous configuration.

[0013]

BEST MODE FOR CARRYING OUT THE INVENTION Embodiments of the present invention will be described below with reference to the drawings.

FIG. 1 is a block diagram showing the configuration of a character recognition device according to an embodiment of the present invention.
1, an image storage unit 12, a recognition processing unit 13, a recognition result information storage unit 14, a recognition result correction unit 15, a recognition result output unit 16, and the like.

The image acquisition unit 11 acquires a multi-tone image such as a grayscale or color image obtained by reading and scanning the form and stores it in the image storage unit 12. Image storage unit 12
Holds the multi-tone image of the form image acquired by the image acquisition unit 11 until the recognition processing is completed.

The recognition processing unit 13 recognizes the character in the reading field, which is the target of the character recognition process, from the multi-gradation image held in the image storage unit 12, and the recognition result including the recognition result and the position coordinate with respect to the image. Is stored in the recognition result information storage unit 14. At this time, the recognition result information storage unit 14 stores the read characters of the entire form as binary image information. On the other hand, for a character rejected according to the recognition result, the rejected character pattern is stored in the recognition result information storage unit 14 as multi-gradation image information.

Each time the recognition result correction unit 15 obtains a recognition result by the above-described recognition processing, the recognition result correction unit 15 uses the GUI as shown in FIG. 3 or 4 based on the information stored in the recognition result information storage unit 14.
A screen (here, a recognition result confirmation screen or a recognition result screen for reject correction) is created, and the recognition result output unit 16
And the correction process is executed on the screen. Here, if there is no reject as a result of recognition, the normal recognition result screen shown in FIG. 3 is displayed, and only when there is a reject, the reject correction recognition result screen shown in FIG. 4 is displayed.

FIG. 2 is a flow chart showing a processing procedure of the character recognition device in the embodiment of the present invention. Here, a processing example of executing the correction processing of the recognition result whenever there is a rejection in the recognition processing is shown. However, it is not limited to this.

FIG. 3 and FIG. 4 are views showing examples of the screen structure of the recognition result screen in the embodiment of the present invention, respectively.
According to the recognition result in the recognition processing unit 13, in the correction process, the normal recognition result screen shown in FIG. 3 or the recognition result screen for reject correction shown in FIG.
Is displayed in. FIG. 3 shows a screen configuration of a normal recognition result screen displayed when there is no reject character in the recognition result of the recognition processing unit 13, and FIG.
7 shows a screen configuration of a reject correction recognition result screen displayed when a recognition character in FIG. In these recognition result screens, here, a display area A1 for displaying the entire form including a field to be recognized as a binary image, and the entire field to be corrected displayed in the area A1 as a binary image. The display area A2 to be enlarged and displayed and the character pattern to be corrected in the field displayed in the area A2 are normally displayed as a binary image (when there is no reject character) (see FIG. 3), and the reject character is If there is, it is configured to have a multi-gradation and a correction area A3 which is displayed by three kinds of binary images with different threshold values (see FIG. 4).

Further, when there is no reject character in the correction area A3, the character pattern is displayed as a binary image corresponding to the character in the field display area (see FIG. 3), while there is a reject character. In this case, the reject character pattern is displayed in multiple gradations (gray pattern) (see FIG. 4).
An original image display area 21 is provided. Further, when there is a rejected character, the density (threshold level value of binarization) of the character pattern of the original image displayed in multiple gradations in the original image display area 21 is changed stepwise (here, three steps). Further, there is provided a density-specific binary image display area 22 for displaying the binary images of respective levels of the reject character pattern in order of density (see FIG. 4). Further, a density changing slider 2 for variably controlling the density (threshold value) of each character pattern displayed in the density-specific binary image display area 22.
3 is provided.

FIG. 5 is a diagram showing a multi-gradation reject character pattern (gray pattern) displayed in the original image display area 21 in FIG. 4, and here, the pattern density is represented by diagonal lines (hatching). ing.

FIG. 6 shows an example of displaying the multi-gradation reject character pattern (gray pattern) shown in FIG. 4 as a binary image applied by a conventional apparatus. This is an example of a character pattern that cannot be determined.

FIG. 7 is a table of three types of binary images in which the densities (threshold values) of the rejected character patterns displayed in the density-specific binary image display area 22 shown in FIG. 4 are changed stepwise. It is a figure which expands and shows an example.

Now, the operation of the embodiment of the present invention will be described with reference to the drawings.

The image acquisition unit 11 acquires a multi-tone image such as a grayscale or color image, and the image storage unit 12
Remember. When the image data (multi-tone image) for each reading form that is the target of the recognition process is stored in the image storage unit 12, the recognition processing unit 13 starts the character recognition process (steps S11 and S12 in FIG. 2).

The recognition processing unit 13 reads the multi-tone image of the form image to be recognized from the image storage unit 12, and executes the recognition process of the reading field to be recognized from the multi-tone image ( 2 (step S13 in FIG. 2), the recognition result and position information are stored in the recognition result information storage unit 14 as recognition result information, and the recognition result is notified to the recognition result correction unit 15 (step S14 in FIG. 2).

Further, the recognition processing unit 13 binarizes the multi-tone image of the form reading image including the recognition target field, stores the binary image information of the entire form in the recognition result information storage unit 14, and rejects it. When there is a character, the character pattern of the rejected character is stored in the recognition result information storage unit 14 as it is in the multi-tone image information.

As a result of the above recognition, if there is no reject character (NO in step S15 in FIG. 2), the recognition result correction unit 15
Receives a notification of normal reading from the recognition processing unit 13, uses the information stored in the recognition result information storage unit 14 to create a normal recognition result screen shown in FIG. 3, and displays the screen on the recognition result output unit. It is displayed on the screen 16 to confirm the recognition, and if necessary, the operator can correct the character (steps S16a and S17 in FIG. 2).

When there is a reject character as a result of the recognition (YES in step S15 in FIG. 2), the recognition result correction unit 15 receives a notification of the reject character from the recognition processing unit 13 and stores the recognition result information. Using the information stored in the unit 14, instead of the normal recognition result screen shown in FIG. 3, a recognition result screen for reject correction shown in FIG. 4 is created, and the screen is displayed on the recognition result output unit 16. (FIG. 2 Step S
16b), it is possible to correct the reject character (step S17 in FIG. 2).

On the recognition result screen for this reject correction, as shown in FIG. 4, a multi-gradation reject character pattern (gray pattern) is displayed in the original image display area 21 and corresponds to the reject character pattern. Three patterns of binary images with different densities (threshold values) are displayed in order of densities in the density-specific binary image display area 22, and a density-changing slider that changes the pattern density of the density-specific binary image display area 22. 23 is displayed.
The operator visually corrects the reject character pattern on the recognition result screen for reject correction.

In the operation of correcting the reject character pattern by the operator using the recognition result screen for reject correction shown in FIG. 4, the binary image of the entire recognition target field displayed in the display area A2 is used. The reject character and its digit position in the field can be easily recognized.

Further, a multi-gradation (gray pattern) reject character pattern displayed in the original image display area 21,
Corresponding to the reject character pattern, the operator can visually recognize and make correction while simultaneously comparing the binary images of the three patterns arranged and displayed in the binary image display area by density 22 in the order of density. It is possible to obtain a correction result that is extremely unlikely to be mistakenly recognized, and it is possible to perform correction with high reliability.

A specific correction example in this case will be described with reference to FIGS. 5 and 6. In the conventional device, the reject character pattern is displayed as a binary image as shown in FIG. 6, and therefore, when the operator corrects the binary image of this character pattern by visual judgment, "7" and "9" are displayed. Is difficult to determine, and is erroneously determined ("9" is changed to "7", or "7"
Is judged to be “9”).

On the other hand, in the embodiment of the present invention, the reject character pattern is displayed as a multi-gradation (gray pattern) image as shown in FIG. 6, so that the loop portion in "9" is blurred. It can be determined that the characters are lightened and only a part remains like noise, so that it is possible to correct the correct number "9" as the answer. Further, as shown in FIG. 7, the density of each pattern in the density-specific binary image display area 22 changes in conjunction with the density operation of the density change slider 23, so that the shape of the written character pattern is confirmed. be able to.

In the above-described embodiment, the reject character pattern is stored in the recognition result information storage unit 14 only when a reject character is generated. The present invention can be applied to the case where it is desired to check the character pattern of a high field by collating it with the original image.

In this case, for example, the "recognition result screen" shown in FIG. 4 is used as a "recognition result confirmation screen", and the field to be confirmed is, for example, FC (format control).
Is defined in advance (the defined field is referred to as a confirmation designation field here).

Regarding the confirmation processing of the recognition result using the recognition result confirmation screen, in the flowchart shown in FIG. 2, it is checked in step S14 whether or not the confirmation designation field exists, and the confirmation designation field is checked. If it exists, the character pattern of the confirmation designation field is stored in the recognition result information storage unit 14 as a multi-gradation image, and step S15 is performed.
When it is checked whether or not the confirmation designation field exists, the character pattern of the confirmation designation field is displayed in the original image display area 21 of the recognition result confirmation screen in the multi-gradation image. (Gray pattern)
And the recognition result of the confirmation designation field is displayed as a binary image in the display area A2 as a comparison target of the multi-tone image, so that a field such as an amount column is not allowed to be erroneously recognized or is highly important. By confirming and specifying the field in advance, the recognition result of the field can be visually checked easily and accurately.

In the above embodiment, the correction process is performed every time a reject is made in the recognition process, but the present invention is not limited to this and, for example, becomes a processing target regardless of the presence or absence of a reject. The recognition processing of all the forms is executed first, and if there is a reject at that time, only the rejected character pattern is set as a multi-gradation image, and the whole is stored as a binary image in the recognition result information storage unit 14 for processing. It is also possible to adopt a functional configuration in which after performing the recognition process for all the target forms, the correction process is performed only for the rejected forms.

[0039]

As described above in detail, according to the present invention, it is possible to perform the recognition result correction work, the confirmation work and the like with high reliability quickly and easily, and to increase the processing load and the storage capacity as much as possible. Therefore, the processing can be speeded up with an economically advantageous configuration.

[Brief description of drawings]

FIG. 1 is a block diagram showing a configuration of a character recognition device according to an embodiment of the present invention.

FIG. 2 is a flowchart showing a processing procedure of the character recognition device in the embodiment of the present invention.

FIG. 3 is a diagram showing a configuration example of a normal recognition result screen in the embodiment of the present invention.

FIG. 4 is a diagram showing a configuration example of a reject correction recognition result screen in the embodiment of the present invention.

FIG. 5 is a diagram showing a density pattern of rejected characters displayed in the original image display area in the above embodiment.

FIG. 6 is a diagram showing an example in which the character pattern shown in FIG. 5 is displayed as a binary image.

FIG. 7 is an enlarged view showing a character pattern displayed in a density-specific binary image display area in the embodiment.

[Explanation of symbols]

11 ... Image acquisition unit 12 ... Image storage unit 13 ... Recognition processing unit 14 ... Recognition result information storage unit 15 ... Recognition result correction unit 16 ... Recognition result output unit 21 ... Original image display area 22 ... Binary image display area by density 23 ... Density change slider A1 ... Display area A2 for displaying entire form in binary image Display area A3 for enlarged display of correction target field in binary image ... Correction area

Claims (8)

[Claims] 1. An image acquisition unit for acquiring a multi-gradation image, a recognition processing unit for recognizing characters of the image acquired by the image acquisition unit, and a rejected character pattern as a result of recognition by the recognition processing unit. Recognition result information storing means for storing a character pattern of the entire recognition result as a binary image, and a graphic for creating and displaying a recognition result screen using the character pattern stored in the recognition result information storing means. Optical character recognition, which comprises a user interface means, displays rejected characters in multiple gradations on the recognition result screen for reject correction, and displays normally recognized characters in a binary image. apparatus. 2. The graphic user interface unit simultaneously displays a multi-gradation character pattern and a binary image of the character pattern on the recognition result screen. Optical character recognition device. 3. The graphic user interface means simultaneously displays a multi-gradation-displayed character pattern and a binary image of the character pattern on the recognition result screen, and displays the character pattern of the binary image. 2. The optical character recognition device according to claim 1, wherein the display is variably displayed at an arbitrary binarization level. 4. The graphic user interface means, on the recognition result screen, a multi-gradation displayed character pattern and a plurality of character patterns obtained by binarizing the character and having different binarization levels. Is displayed at the same time, the optical character recognition device according to claim 1. 5. The graphic user interface means associates a character pattern displayed in multi-gradation on the recognition result screen with a character digit position in a field in which the character exists or in an entire image and simultaneously. The optical character recognition device according to claim 1, which is displayed. 6. An image acquisition unit for acquiring a multi-gradation image, a recognition processing unit for performing character recognition of an image acquired by the image acquisition unit, and a character pattern recognized by the recognition processing unit, which are predetermined. A recognition result information storage unit that stores a character pattern of a specific area in a multi-gradation image and stores at least another character pattern in a binary image, and recognizes using the character pattern stored in the recognition result information storage unit. And a graphic user interface unit for creating and displaying a result screen, wherein the entire recognition result is displayed as a binary image on the recognition result screen, and the character pattern of the specific region is displayed in multiple gradations. Optical character recognition device. 7. The character recognition of the read multi-tone image is performed, the character pattern of the entire recognition result is stored as a binary image, and the reject character is stored as the multi-tone image, and the recognition result is displayed. A method for displaying a recognition result screen, characterized in that the character pattern of the entire stored recognition result is displayed on the screen as a binary image, and the stored rejected character pattern is displayed as a multi-gradation image. 8. A character recognition of a read multi-tone image is performed, a character pattern of the entire recognition result is stored as a binary image, and a character pattern of a specific area designated by format control or the like is applied to the multi-tone image. The character pattern of the entire stored recognition result is displayed as a binary image on the screen for confirming the recognition result, and the character pattern of the specific region is displayed as a multi-gradation image. How to display the recognition result screen.