JPH01166184A - Document processor - Google Patents

Abstract

PURPOSE:To improve the recognition efficiency of a character at a surface part by recognizing the character included within an area divided by a ruled line as one belonging to the surface part. CONSTITUTION:A line segment data memory part 4 stores each data concerning with the start point and the end point of each line segment in document information inputted through an image input part 1. A ruled line recognition part 15 recognizes each line segment which continues in an identical direction as one ruled line based on the stored data of the memory part 4 and simultaneously stores each data concerning with the start point and the end point of each ruled line in a ruled line data memory part 5. Then, a character within a ruled line output part 16 sends the character part included within the area divided by the ruled line to a character recognition part 17 based on the stored data of the memory part 5 and a video signal from the image input part 1 and makes the character recognition part 17 to recognize the character.

Description

[Detailed description of the invention] [Industrial application field]

This invention specifies the table part from document information that is roughly divided into text parts, table parts, and graphic parts, recognizes the ruled lines that make up this table part, and recognizes the characters within the ruled lines by a character recognition unit. The present invention relates to a document processing device.

[Conventional technology]

Conventionally, the main target of the recognition process for document information is the text within the document, which is first divided into rows or columns of characters, then separated (cut) into individual characters, and then processed based on an optical reading method. This was done using the procedure of character recognition.

[Problems to be solved by the invention]

In general, document information is roughly divided into a text part, a table part, and a graphic part, and the table part consists of ruled lines and character parts (more precisely, letters, numbers, symbols, etc.) within the ruled lines. Text portions can be recognized by character recognition means mainly based on the optical reading method as described above, and graphic portions can still be recognized at a certain level of recognition rate by means of figure recognition means based on a similar method. However, due to the special conditions that the characters belonging to the back part are written within the ruled lines, ■ it is difficult to recognize the characters because the line spacing and character spacing is different, ■ it takes a lot of time even if it is possible to recognize them, ■ Separate sentences separated by ruled lines may be mistakenly recognized as one continuous sentence. - Vertical and horizontal writing may be confused and misrecognized. For example, in the back part shown in FIG. 10(a), there is one line of characters "abc" in the area to the left of the center of the ruled line,
Two lines of letters rdefgj and rhijk in the central area.
There is. After removing the ruled lines as shown in FIG. 2B, when examining the characters horizontally, the one-line portion branches into two-line portions, making it difficult to make a decision. Also, as shown in Figure 11(a), on the back side, "qrst" and ruv
wxyz" is written in the adjacent ruled line area, if you remove the ruled line as shown in FIG.
There is a risk that the character group rqrStuVWX)'ZJ may be mistakenly recognized as a continuous line of characters. In other words, with the conventional technology, individual characters can be recognized accurately to a fairly high level, but recognition of characters included in the back part is insufficient, the recognition rate is low, and the recognition time is long. There is a problem. SUMMARY OF THE INVENTION It is an object of the present invention to provide a document processing device that can solve the above-mentioned problems of the prior art and can accurately and quickly recognize the hidden parts of document information.

[Means to solve the problem]

In order to achieve the above object, the document processing device according to the present invention specifies the back part from document information that is roughly divided into a text part, a front part, and a graphic part, and recognizes the ruled lines that make up this back part. At the same time, the device recognizes the character portion within the ruled line by a character recognition unit, the line segment storing each data regarding the start point and end point of each line segment in the document information input via the image input unit. a data storage unit; a ruled line recognition unit that recognizes each line segment continued in the same direction as one ruled line based on the data stored in the line segment data storage unit; starting point,
a ruled line data storage unit that stores each data regarding the end point; and the character portion included in the area partitioned by each of the ruled lines based on the data stored in the ruled line data storage unit and the video signal from the image input unit. and a ruled line character output section for sending the character to the character recognition section.

[For use]

The line segment data storage unit stores each data regarding the starting point and ending point of each line segment in the document information input through the image input unit. Based on the data stored in the line segment data storage unit, the ruled line recognition unit recognizes that each line segment continues in the same direction as a ruled line, and also stores each data regarding the starting point and end point of each ruled line in the ruled line data storage unit. Store. Based on the data stored in the ruled line data storage section and the video signal from the image input section, the ruled line character output section sends the character section included in the area divided by the ruled lines to the character recognition section, and performs character recognition. Character recognition is performed by the section.

【Example】

A document processing device showing an embodiment of the present invention will be described below with reference to the drawings. FIG. 1 is a block diagram showing the configuration of this document processing device.
This document processing device 20 is roughly divided into an image input section 1, various memories, and various processing sections. The various memories include an original image memory 2 that stores the document video signal data from the image input section I, a thinning image memory 3 that stores the thinned data processed by a thinning processing section 13 (to be described later), and a thinning image memory 3 that stores ruled lines. A line segment data memory 4 that stores constituent line segment data, a ruled line data memory 5 that stores ruled line data composed of the line segment data, and a graphic memory 6 that stores graphic part data as document information. It is. Note that the graphic memory 6 is not particularly handled in the document processing device 20 according to the present invention. The various processing units include the line thinning processing unit 13 mentioned above, the line segment coordinate determining unit 14 that determines the coordinates of the starting point and ending point of each line segment forming a ruled line based on the data in the line thinning image memory 3, and the line segment coordinate determining unit 14 in the line segment data memory 4. A ruled line coordinate determining unit 15 that recognizes each ruled line based on data and determines the coordinates of its starting point and ending point.
, a ruled line character output section 16 that specifies and outputs the characters included in the ruled lines, and a character recognition section 17 that receives and recognizes the character output. The operation of this document processing device 20 will be described below, with reference basically to the flowchart shown in FIG. 2, and supplementary to the block diagram shown in FIG. 1, and the like. In FIG. 2, in step S1, the image input section 1 (first
The video signal of the document information from (see figure) is memorized and stored in the original image memory 2. "Thinning process" in step S2
This is a type of image shaping process that is performed by the thinning processing section 13 based on the video signal data stored in the original image memory 2, and converts the image into thin lines of a predetermined width. The result is stored in the thinning image memory 3. "Determination of line segment coordinates" is performed in step S3. This "determination of line segment coordinates" is performed by the line segment coordinate determination unit 14 based on the data stored in the thinning image memory 3, and each line segment constituting the ruled line is starting point. This is the process of determining each coordinate of the end point. Here, the line segment is
It is a straight line section divided by free ends or intersections, and its start and end points correspond to the free ends or intersections. Determination of line segment coordinates will be specifically explained with further reference to FIGS. 3 and 4. Note that FIG. 3 is an illustrative diagram of a table portion in document information, and FIG. 4 is a diagram showing each line segment forming the ruled lines of this table portion. Alphabets within the ruled lines indicate characters to be written. As shown in Figure 4, ruled lines are constructed as a collection of line segments between their intersections, and each line segment has a number assigned to each intersection.
It is represented by or ■. For example, line segment L1 is ■-
■, line segment L2 is ■-■, and so on, line segment L21 is 0
−■, is. The coordinates of the starting point and end point of each line segment correspond to the coordinates of the intersection marked with ■ or [phase], and the coordinates of each intersection point can be easily obtained using well-known image processing technology, so each line segment "Determination of ruled line coordinates" is performed in step S4 in which the coordinates of the start point and end point are determined. This is done by the ruled line recognition unit 15 based on the data from the line segment data memory 4, and each ruled line forming the table part This process determines the coordinates of the starting point and ending point. A ruled line is defined as a set of a plurality of line segments having the same direction, and is indicated by K1 to K7 in FIG. 5, which shows ruled lines constituting a table portion, for example. In Figure 5, the number given to each intersection point is
to ■ are the same as in FIG. If a ruled line is represented by numbers for its starting point and ending point, each ruled line has a number from 1 to 7.
is as shown in the correspondence diagram shown in FIG. Obtaining the corresponding diagram in FIG. 6 corresponds to determining the ruled line coordinates. By the way, the ruled lines of the table part have a frame shape with the outside closed and vertical and horizontal ruled lines arranged inside, as shown in FIG. 3 already explained. There are various forms such as ), (a) in the same figure, and (c) in the same figure. That is,
Compared to the ruled lines in Figure 3, in Figure 7 (a) there are no vertical ruled lines on both the left and right sides, in Figure 7 (b) there are no internal vertical ruled lines, and in Figure 7 (C) there are no vertical ruled lines on both sides. There is no such thing. In step S5, "determination of divided area by ruled lines" is performed, which is a process of determining an area divided by two to four ruled lines, and the ruled line recognition unit 15 performs the "determination of ruled line coordinates". It is carried out in connection with. Next, explanation will be given with reference to FIGS. 8 and 9. In addition, FIG. 8 is a diagram showing an area divided by ruled lines, and FIG.
10 is an overall view of the ruled lines, (b) is a view of the maximum area by the ruled lines, (c) is a view of the intermediate area, and (d) is a view of the final area. First, as shown in Figure 8, the four parts that make up the outer frame are
2. 7. The area Mo is determined by 8. Next, as shown in FIG. 8(C): 2. 7. with 8 on , and 3 on two horizontal lines
．． 5, the area MO is subdivided into areas M1.
M2. M3 is determined. Next, as shown in FIG. 8(d), 4. The area is subdivided by 6 to determine a new area, but two steps are taken: first by using 4 for the ruled line, and then by using 6 for the next ruled line. That is, due to the ruled line 4, the area M1 is changed to the area M11. M12, region M2 is region M21°Mg2 (broken line frame), region M3 is region M31. Each is subdivided into M32 (dashed line frame). Next, by the ruled line 6, the area M22 is changed to the area M23.
M24, area M32 and area M33. Each is subdivided into M34. FIG. 9 is a diagram showing a hierarchical structure of each region divided by ruled lines, and is a diagram in which the above-mentioned regions are arranged in a determined order. That is, the code attached to the circle is the area code mentioned above, and for example, area Mo is area Ml, M2 . M3
Then, area M1 is subdivided into areas Mll, M12
The same applies hereafter. Now, returning to FIG. 2, in step S6, only the characters within the area demarcated by the ruled lines are outputted by the ruled line character output section 16 based on the data from the ruled line data memory 5 and the data from the original image memory 2. Sent to the character recognition unit 17,
Character recognition is performed here (step S7). Although not directly relevant to the present invention, data of graphic information within ruled lines is stored in the graphic memory 6 based on the data in the original image memory 2 and the ruled line recognition unit 15. and,
Based on the data in the graphic memory 6, a graphic within the ruled line is recognized by a graphic within the ruled line output section (not shown) and a graphic recognizing section. This processing is "output of figures within ruled lines" in step S8 and "figure recognition" in step S9.

【Effect of the invention】

As explained above, in the present invention, the line segment data storage section stores data regarding the starting point and end point of each line segment in the document information input via the image input section; Based on the data stored in the line segment data storage unit, each line segment that continues in the same direction is recognized as one ruled line, and each data regarding the starting point and end point of each ruled line is stored in the ruled line data storage unit; Characters in the ruled line The output section sends the character part included in the area divided by the ruled lines to the character recognition section based on the data stored in the ruled line data storage section and the video signal from the image input section, and the character recognition section converts the characters. Recognition takes place. Therefore, the present invention has the following superior effects compared to the conventional technology. (1) The ruled lines are recognized by the ruled line recognition section of the device of the present invention, and the characters included in the area demarcated by the ruled lines are recognized by the character recognition section as belonging to the front section without being confused with the text section. As a result, the recognition rate of the front part can be improved. (2) Since characters belonging to the same type of ruled line are considered to be related, the recognition processing speed can be improved, so that the overall time required for recognizing the front part can be shortened.

[Brief explanation of the drawing]

FIG. 1 is a block diagram showing the configuration of an embodiment according to the present invention, FIG. 2 is a flowchart showing the operation of this embodiment, and FIG.
The figure is an example of a table part in document information, Figure 4 is a diagram showing each line segment making up the ruled lines of this table part, Figure 5 is a diagram showing the ruled lines making up this table part, and Figure 6 is a diagram showing each line segment making up the ruled lines of this table part. Figure 7 shows the correspondence between ruled lines and their starting and ending points. Figure 7 (a) is another ruled line, Figure 7 (b) is yet another ruled line, and Figure 7 (C) is yet another ruled line. Figure 8
The figure shows an area divided by ruled lines.
) is an overall view of the ruled line, (b) is a view of the maximum area divided by the ruled line, (C) is a view of the intermediate area, and (d) is the same (a view of the final area, Figure 9 is a diagram showing the hierarchical structure of areas partitioned by ruled lines, Figure 10 is an explanatory diagram of an example of a case where errors may occur in conventional recognition of characters within ruled lines, and (a) in the same figure shows the hierarchical structure of areas partitioned by ruled lines. In this case, Fig. 11 is an explanatory diagram of another example, Fig. 11 is an explanatory diagram of another example, Fig. 11 is an explanatory diagram of the case in which the ruled lines are removed, Fig. 11 is an explanatory diagram of the case in which the ruled lines are removed, and Fig. 11 (b) is in the case in which the ruled lines are removed. 2: original image memory, 3: thinning image memory, 4: line segment data memory, 5: ruled line data memory, 6: figure memory, 13: thinning processing unit, 14. : line segment coordinate determination unit, 15: ruled line recognition unit, 16: ruled line character output unit, 17:
Character recognition unit, 20: document processing device. 1st group %2V %61 ni7: @-■ K8: ■-■ Akira 11 ff
i

Claims (1)

[Claims] The table portion is specified from document information roughly divided into text portions, table portions, and graphic portions, and the ruled lines that make up this table portion are recognized, and the character portions within the ruled lines are recognized as characters. a line segment data storage section that stores data regarding the starting point and end point of each line segment in the document information input through the image input section; a ruled line recognition unit that recognizes each line segment continued in the same direction as one ruled line based on stored data; and a ruled line data storage that stores data regarding the starting point and end point of each ruled line recognized by the ruled line recognition unit. and; based on the data stored in the ruled line data storage section and the video signal from the image input section, the character section included in the area partitioned by each ruled line is sent to the character recognition section. A document processing device comprising: a character output unit;