JPH07160811A - Character recognizing device - Google Patents

Abstract

PURPOSE:To the attribute or size of the precision by making a recognizing device possible to correction the size and attribute of an area by before a recognition processing by changing the attribute or size of the area by an area correction means. CONSTITUTION:The binary image data of a document to be a recognition object is fetched by a scanner 4 and areas with attributes of characters, tables and graphics, etc., are extracted for the fetched image data based on the image data of resolution 100 DPI for area extraction. For the extracted area, a frame showning the range of each area and image data are overlapped and it is displayed on a CRT 7. When the state displayed at present is image data and an area frame, the attributes of the area frame and the area are displayed. When the attributes of the area frame and the area are displayed, the display is switched to image data and the area frame. When an area correction is selected by a keyboard 5 or a mouse 6, a recognition processing according to the attribute of the area is performed for the selected area. As for the image area, the compression of image data is performed and data quantity is reduced.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to character data, graphic data, and the like from binary image data captured by using an optical means such as a scanner for making a database of printed documents and reusing documents.
The present invention relates to a character recognition device that extracts a region for each attribute such as a table, an image, and a ruled line, and performs recognition processing according to each attribute.

[0002]

2. Description of the Related Art A flow of recognition processing in a conventional character recognition device will be described below. (1) Capture binary image data of a document to be recognized by a scanner. (2) Characters, figures,
Areas having attributes of tables, images (photos), and ruled lines are extracted, and a frame indicating the range of the extracted areas and the attributes of the areas are displayed. (3) Select an area in which the displayed area result is recognized by a pointing device such as a mouse. (4) The recognition processing is performed on the selected area according to the area attribute.

When a character area is selected, character recognition is performed, when a graphic area is selected, graphic vectorization processing is performed, and when a table area is selected, a table structure is recognized, and then a cell is recognized. Character recognition inside. Further, the image area is compressed to reduce the data amount. (5) The recognition result is displayed for the character recognition result, and for the recognition result with low recognition accuracy, a plurality of candidate characters are selected, or character editing such as insertion / deletion of characters is performed to make a desired result Achieve data acquisition.

[0004]

However, in the above-mentioned conventional character recognition device, there is a case where what is actually an image is recognized as a character, and the character recognition is applied to the image, and the character recognition result is distorted. Was sometimes. In such a case, the operator has to correct the random recognition result, which causes a problem that the operator's burden becomes enormous.

Therefore, an object of the present invention is to provide a character recognition device which can reduce the burden on the operator.

[0006]

According to the present invention, a character, a table, a figure,
A method for extracting an area with attributes of images and ruled lines, a means for displaying the extracted area and the binary image data in an overlapping manner, a means for displaying the area frame and the area attribute in an overlapping manner, and a size of the area Means for scaling, means for changing the attributes of a region,
A means is provided for performing recognition processing on the area according to the area attribute.

[0007]

According to the present invention, the attribute or size of the area can be changed by the area correcting means, and the operator can change the attribute size to the correct attribute size prior to the recognition processing, thereby improving the accuracy of subsequent recognition. As a result, it is possible to reduce the operator's troublesome correction.

[0008]

DESCRIPTION OF THE PREFERRED EMBODIMENTS A character recognition device according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram of a character recognition device according to an embodiment of the present invention. A central processing unit (hereinafter abbreviated as CPU) 1 executes a character recognition program stored in a read only memory (hereinafter abbreviated as ROM) 2. A random access memory (hereinafter abbreviated as RAM) 3 stores image data obtained by capturing a recognition target document from the scanner 4 and data used in a character recognition program. Keyboard 5, mouse 6
Is a device from which an operator gives a confirmation / correction command to the recognition result from the outside. The CRT 7 is a display device that displays the recognition result executed by the CPU 1.

FIG. 2 is a flow chart of the character recognition apparatus in one embodiment of the present invention, and FIG. 3 is a display example in which a frame representing an area and binary image data are displayed in an overlapping manner in the character recognition apparatus in one embodiment of the present invention. FIG. 4 and FIG. 4 are display example diagrams in which areas and attributes are overlapped and displayed in the character recognition device according to the embodiment of the present invention. Hereinafter, the character recognition processing of the character recognition device according to the embodiment of the present invention will be described with reference to FIGS.

The scanner 4 takes in binary image data of a document to be recognized (S1). In this embodiment, the binary image data fetched for performing the region extraction processing at high speed is OR-reduced to a resolution of about 100 DPI and stored in the RAM 3.

In order to confirm whether or not the recognition processing is performed on the captured image data, the captured image data is C
It is displayed on RT7 (S2). The image data displayed here is data that has been reduced so that the entire image data can be displayed.

Next, it is determined by the keyboard 5 or the mouse 6 whether or not to continue the processing for the captured image data (S3). When the processing end is selected, the process jumps to end.

In the case of continuing processing, the resolution 1 for area extraction
Based on the image data of 00DPI, areas with attributes of characters, tables, figures, images (photos), and ruled lines are extracted (S4).

As shown in FIG. 3, a frame representing the range of each area and image data are superimposed on the area extracted in S4,
Displayed on the CRT 7 (S5), and external command (input) from the operator using the keyboard 5 or mouse 6
Wait for (S6).

When there is an input from the keyboard 5 or the mouse 6, it is determined whether or not an area for recognition processing is designated (S7). If the area to be recognized is selected, the process jumps to S16. If it is not the area selection, the process proceeds to S8.

In S8, it is determined whether or not the enlarged display is selected. If the enlarged display is selected, the process jumps to the enlarged display process of S13, and if not, the process proceeds to S9.

In S9, it is determined whether or not the reduced display is selected. If the reduced display is selected, the process jumps to the reduced display process of S14, and if not, the process proceeds to S10.

In step S10, it is determined whether the area correction has been selected. If the area correction is selected, the processing jumps to the area correction processing of S15, and if not, the processing proceeds to S11.

In S11, it is determined whether or not the screen switching of the area extraction result display is selected. When the screen switching is selected, the process proceeds to S12, and when the input is other than that, S6 is performed.
To return to the input waiting state.

Step S12 is a process when screen switching is selected by the keyboard 5 or the mouse 6. When the currently displayed state is the image data and the area frame as shown in FIG. 3, the area frame and the area attribute as shown in FIG. 4 are displayed. Further, when the currently displayed state is the area frame and the area attribute are displayed as shown in FIG. 4, the display is switched to the display of the image data and the area frame as shown in FIG. That is, the screen switching is toggled between the states shown in FIGS. The state in FIG. 3 is shown in FIG.
The state of is provided to confirm that the attributes of the region are detected correctly. The area attributes may be distinguished by the pattern of the frame line or by the color, but in this case, the black and white CRT 7 cannot make the determination, so this display switching method is realized. After the area display is switched, the process returns to the input waiting state of S6.

Step S13 is a process when the enlarged display is selected by the keyboard 5 or the mouse 6.
In the processing here, the magnification of the image data currently displayed is increased by one step and redisplayed. By providing this function, the detected area can be corrected with high accuracy. After the enlarged display process is completed, the process returns to the input waiting state of S6.

S14 is a process when the reduced display is selected by the keyboard 5 or the mouse 6.
In this processing, the magnification of the image data currently displayed is reduced by one step and the image data is displayed again. Select this function when you want to see the image of the entire document of the enlarged image data.
After the reduction display process is completed, the process returns to the input waiting state of S6.

Step S15 is a process when area correction is selected by the keyboard 5 or the mouse 6.
This function is selected when the detected area is corrected, such as when the area extracted result is incorrect.
The correction items include changing the coordinate value of the area, changing the attribute, deleting the area, adding, dividing, and integrating, and it is possible to obtain a desired area result. After the area correction process is completed, the process returns to the input waiting state of S6.

In S16, the process is performed when the area correction is selected by the keyboard 5 or the mouse 6. A recognition process is performed on the selected area according to the area attribute. When the character area is selected, the character recognition is performed, when the figure area is selected, the vectorization process of the figure is performed, and when the table area is selected, the table structure is recognized. To recognize. Further, the image area is compressed to reduce the data amount.
For the result of character recognition, display the recognition result,
For the recognition result with low recognition accuracy, a plurality of candidate characters are selected, or character editing such as insertion / deletion of characters is performed to obtain desired data.

[0025]

Since the present invention is configured as described above, the size and attributes of the area can be corrected prior to the recognition processing, and the recognition accuracy can be improved, so that the burden on the operator can be reduced.

[Brief description of drawings]

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

FIG. 2 is a flowchart of a character recognition device according to an embodiment of the present invention.

FIG. 3 is a display example diagram in which a frame representing an area and binary image data are displayed in an overlapping manner in the character recognition device according to the embodiment of the present invention.

FIG. 4 is a diagram showing a display example in which an area and an attribute are overlapped and displayed on the character recognition device in the embodiment of the present invention.

[Explanation of symbols]

 1 Central Processing Unit (CPU) 2 Read Only Memory (ROM) 3 Random Access Memory (RAM) 4 Scanner 5 Keyboard 6 Mouse 7 CRT

Claims (1)

[Claims] 1. A means for extracting an area with attributes of characters, tables, figures, images and ruled lines from binary image data obtained by binarizing a character recognition target document, and superposing the extracted area and the binary image data. Display means, area frame and area attribute overlapping display, area size scaling means, area attribute changing means, and area recognition processing according to area attributes A character recognition device comprising means for performing.