JPH0773271A - Area dividing method - Google Patents

Abstract

PURPOSE:To automatically set the threshold value for integration of a character rectangle and also to separate the text area from the footers, the headers, etc., when a document area is divided. CONSTITUTION:A circumscribed rectangle detecting part 8 detects a circumscribed rectangle of the black picture element coupling component of an image obtained by applying the OR compression to an input document image through an image compressing part 7. A rectangle sorting part 9 sorts the rectangles into the character rectangles and other rectangles. A character area integrating part 10 automatically sets the threshold value for integration of characters based on the vertical/horizontal ratio of the character rectangle or a space distribution and integrates the lines of the character rectangle. However the integration of the text lines and the footers or headers is inhibited between the character rectangles. Furthermore the threshold value is increased by punctuation marks so that the failure of integration of punctuation marks can be prevented. Then the character rectangles of footers and headers are integrated with the character area different from the text area.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an area dividing technique for dividing an information area having different attributes such as a character area and a figure area from image data such as a document.

[0002]

2. Description of the Related Art When recognizing characters in a document in which character areas and chart areas are mixed, or when an image of such a document is filed on a storage medium such as an optical disk or transmitted by facsimile. For example, it is necessary to divide the information areas having different attributes and switch to the optimum processing method for each area.

Regarding such information area division, a method of dividing into a character area and other areas by using a peripheral distribution (Japanese Patent Laid-Open No. 62-71379) and a circumscribed rectangle of a black pixel group by utilizing the peripheral distribution are extracted. Then, a method of forming a character area by integrating the rectangles and dividing the character area from other areas (Japanese Patent Laid-Open No. Hei 3 (1999) -1999)
-126181) is known.

However, the former method has a problem that it is difficult to deal with the skew of the image, and the parameter must be set for each target document. The latter method has a problem that skew correction processing is required depending on the degree of image skew due to the use of peripheral distribution for rectangle extraction. There is a problem that it is difficult to deal with a complicated document having a structure including an article area.

The applicant's 1992 2673
The specification and drawings attached to the No. 13 patent application disclose an improved region segmentation method that utilizes rectangle integration to extract character regions. According to this method, skew correction can be dispensed with, and a document having a complicated structure and documents of various character sizes can be dealt with. However, if the setting of the integration parameter (threshold distance between rectangles) in the line direction of the character rectangle is inappropriate, adjacent areas that should not be combined will be combined, or a character string in a certain line direction will be created. There is a risk that the text area to be displayed and a character string in a different direction close to the text area (such as a title for the text) are mistakenly integrated.

[0006]

SUMMARY OF THE INVENTION Therefore, an object of the present invention is to improve the above-mentioned problems or problems relating to area division, and to improve the accuracy of area division for documents having complicated structures or documents having various character sizes. It is to improve stability.

[0007]

According to the present invention, a rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like at a predetermined magnification, and the extracted rectangle is defined as a rectangle of characters and other rectangles. Provided is an improved method of dividing an area, which classifies a rectangle of characters and integrates a rectangle of characters in the row direction.

That is, according to the first aspect of the invention, the degree of fusion between characters due to image compression is estimated by taking the ratio of the height and width of the rectangle of the character, and the degree of fusion of the characters is estimated according to this degree of fusion. The threshold value of the inter-rectangle distance for integrating the rectangles in the row direction is set.

According to the second aspect of the present invention, the distribution of the distances between the rectangles of the characters adjacent to each other in the line direction is measured, and based on the distribution, the rectangles of the characters are integrated in the line direction. Set the threshold of the distance between rectangles.

According to the third aspect of the present invention, the degree of fusion between characters due to image compression is estimated by taking the ratio of the height and width of the character rectangle, and the character rectangle is integrated in the line direction. As a method of setting the threshold value of the inter-rectangle distance for the above, a method of setting by the fusion degree or a method of setting by the distribution of the distance between the rectangles of the characters that are close to each other in the row direction is used depending on the degree of fusion degree. To select.

According to the fourth aspect of the present invention, during the process of integrating the character rectangles in the line direction, it is determined whether or not the focused character rectangle is a punctuation mark or a similar symbol rectangle. However, when it is determined that the rectangle is a punctuation mark or a specific symbol similar to it, the threshold value of the distance between rectangles regarding the integration of the rectangle of interest and the rectangle of the next character is set to a rectangle of a character other than the punctuation mark or a similar specific symbol. Increase more than the case of integration.

According to the fifth aspect of the invention, during the process of integrating the character rectangles in the line direction, one of the two rectangles to be integrated and the other rectangle are in different directions. When it is determined whether or not it is a rectangle of characters in a line, and when it is determined to be a rectangle of characters in a different direction, the integration of the two rectangles is prohibited.

Further, according to the invention of claim 6,
The method according to claim 5, wherein, when a line having a short length in the line direction is present in the lines obtained by integrating the rectangles of characters in the line direction, the line is moved from the short line to the line direction. On the other hand, the process of searching for short lines in the vertical direction and integrating them is performed.

According to an eighth aspect of the present invention, a rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like with a predetermined magnification, and the extracted rectangle is classified into a character rectangle and other rectangles. , The character rectangle is integrated in the body character line direction, and the line obtained by combining the rectangles is combined to cut out the character part area, but the line obtained by combining the character rectangle in the body character line direction is extracted. By classifying into body text lines and lines other than body text lines, and giving lines other than body text lines a unique attribute different from body text lines, the text is integrated when the lines are integrated into the character area. Avoid merging non-line lines with body text lines.

According to a ninth aspect of the invention, in the method of the eighth aspect of the invention, lines other than the text line of the text existing inside or around the rectangle classified into the rectangles other than the characters and the ruled lines by the classification of the rectangles. Among them, it is characterized by classifying it into a line in which a line that can be integrated in the direction perpendicular to the text line direction is found and a line in which it cannot be found, and giving each line a unique attribute. .

The features of the invention according to claim 10 are as follows.
In the method of the invention described above, regarding lines other than the text character line, lines having the same attribute are integrated.

The eleventh aspect of the invention is characterized in that, in the method of the ninth aspect, lines having an attribute other than the text line are deleted.

According to the twelfth aspect of the present invention, in the method of the eighth aspect, the rectangles classified into rectangles other than characters and ruled lines by the rectangle classification are subdivided into a figure rectangle and a photograph rectangle. , The lines other than the text line in the photo rectangle are erased, or the photo attribute is given.

[0019]

The difference in the degree of fusion between the characters on the compressed image of the image of the document or the like depends on whether the character interval of the document or the like is wide or narrow. The invention according to claim 1 or 3 estimates the fusion degree between the characters from the aspect ratio of the character rectangle, thereby adapting to the character interval of the document to be processed or the like, and integrating the character rectangle in the line direction. It is possible to optimize the threshold value of the distance between the rectangles. Further, the invention according to claim 2 or 3 optimizes the threshold value with high accuracy based on the distribution of the distance between the character rectangles even in the case of a character or the like in which the accuracy of the threshold value is easily deteriorated depending on the fusion degree between the characters. Can be converted.
Therefore, according to the inventions of claims 1 to 3, it is possible to deal with various documents having different character intervals.

Further, according to the invention described in claim 1, it may be possible to expect a high-speed determination process of the threshold value, while according to the invention described in claim 2, it is possible to expect a more accurate optimization of the threshold value. . According to the third aspect of the invention, such two threshold value determining methods can be selectively used according to the width of the character spacing of the processed document.

In a word processing document or the like, an interval between a punctuation mark or a specific symbol and a succeeding character may be considerably wider than an interval between ordinary characters. Therefore, for example, the method according to the present invention according to claims 1 to 3 is adaptive. If the thresholds for character rectangle integration that have been set to or that are set manually are slightly deviated, punctuation marks and specific symbols that should be integrated on the same line as normal characters may be separated without being integrated. is there. Claim 4
According to the described invention, the punctuation mark or the like can be prevented from being separated by increasing the threshold value for integrating the punctuation mark or the like and the next character.

In a vertically written document, a horizontally written header or footer character string may be close to the end of a character line. According to the invention described in claim 5, it is possible to prevent such erroneous integration of the characters of the header or footer and the characters of the body line. Further, according to the invention of claim 6, the character strings of the header and the footer can be integrated as a character area (line) different from the body character area.

In many cases, the accuracy of the rectangle classification process is not high enough to classify a halftone original such as a photograph or a block of figures composed of fine line figures into completely non-characters by the rectangle classification. This is because a rectangle of about the character size contains many characters that cannot be recognized as a character unless character recognition is performed. If there is a high rate of failure in rectangle classification, unnecessary small areas will occur as character areas due to noise, halftone, and the like. However, strictly classifying the rectangles based on the result of character recognition is not good in terms of processing speed.

According to the eighth to twelfth aspects of the present invention, even if it is difficult to determine whether it is a character rectangle or a figure or other rectangle from the size of the rectangle, it is possible to make a determination from the information of the surrounding area. However, by classifying the lines in which the character rectangles are integrated, it is possible to prevent other areas from being integrated into the text area of the body, and to prevent generation of unnecessary small areas such as a halftone photo area.

According to the invention of claim 8, after the character rectangle is integrated into a line, the line is classified into a body character line and other lines. In the invention of FIG. 9, the lines other than the text line are further classified by paying attention to the positional relationship with the rectangles other than the characters and the ruled lines.
Such a classification is determined from information such as the size of a line, the distance between lines, the distance from other character rectangles, etc.
The process itself can be performed at high speed, and the rate at which character rectangles that are difficult to determine by a rectangle alone can be identified increases, and the rate at which character rectangles in different line directions can also be identified increases. It becomes possible to extract a region. Further, according to the invention of claim 10 or claim 11, the lines classified into lines other than the text character line are integrated or erased for each of the same attributes, so that an extra small area can be prevented from occurring. . According to the twelfth aspect of the invention, even if a character string exists inside the photo area, the character area does not cause the photo area to fall apart.

[0026]

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS FIG. 1 shows a schematic configuration of an example system in which the present invention is implemented. In FIG. 1, the data of the document image read by the scanner 1 is stored in the image memory 3 by the image input unit 2. The document image data in the image memory 3 is read by the image compression section 7 of the area identification section 6, where OR compression is performed at a predetermined compression rate, and then the document image data is returned to the image memory 3. The document image data in the image memory 3 can be displayed on the display 5 by the image display unit 4. It is also possible to directly input the document image data from the image input unit 7 to the image compression unit 7 and store the image data compressed here in the image memory 3.

The area identification unit 6 includes the image compression unit 7 and
Circumscribing rectangle detection unit 8, rectangle classification unit 9, character region integration unit 10
Consists of.

The circumscribing rectangle detecting section 8 is a section for examining the connection of black pixels from the compressed document image data in the image memory 3 and extracting a rectangle circumscribing the black pixel connected component. The rectangles extracted here include rectangles of various attributes such as ruled lines, noises, tables, graphs, figures, in addition to rectangles of characters (a character string of one or more characters). The rectangle classification unit 9 is a unit that classifies such various rectangles from the shape information of the rectangles and the internal data. There are various possible methods for processing this rectangular classification, and an example thereof will be described in detail later.

The character area integration unit 10 integrates the rectangles classified by the rectangle classification unit 9 into the character rectangles and integrates them into the character area. This is a part directly related to the gist of the present invention. .

In the present embodiment, the result of the area identification by the area identification unit 6 is used for the character recognition processing by the character recognition unit 11. That is, the area identification unit 6 is used as a preprocessing unit for optical character recognition. But this is just one example.

FIG. 2 shows an example of the internal structure of the character area integrating unit 10. In FIG. 2, the parameter setting unit 10A is a unit that sets a parameter serving as a threshold for integration between character rectangles, and the processing content thereof will be described later with reference to FIGS.
Will be described with reference to. The inter-character integration unit 10B is a part that performs a process of integrating character rectangles in the row direction, and the content of the process will be described later with reference to FIGS. 6 to 8.
The region integration unit 10C is a unit that integrates the lines integrated by the process of the inter-character integration unit 10B in the direction perpendicular to the lines, and the processing content thereof will be described later with reference to FIG.

Example 1 of Parameter Setting Process of Parameter Setting Unit 10A Since a general document has characters clogged in the line direction, OR
On the compressed document image, adjacent characters are fused to form a long rectangle in the line direction. On the other hand, in word processing documents, the characters are often separated from each other, and the characters are not easily fused even after the OR compression.

As described above, the degree of fusion between characters depends on the distance between characters on the original document image, that is, OR.
The inter-character distance can be estimated from the degree of fusion of characters in the line direction due to compression. Therefore, the optimum intercharacter integration parameter (threshold value) corresponding to the interfusion degree of characters is obtained in advance by experiment, and this is set as a table format or a general formula, and based on the measured interfusion degree of characters. Optimal parameters can be set automatically. This is the processing principle in this example.

The processing contents of this example will be described in detail with reference to the schematic flow chart shown in FIG. Step 3
In 0A, in the case of a horizontal writing document, a rectangle height (rectangle size in the direction perpendicular to the line direction) histogram is taken for the rectangles classified into the character rectangles in the rectangles extracted from the compressed image, and the frequency is the highest. The size of the rectangle with a large is the standard character size. Next, the rectangular width (rectangular size in the row direction) of the character rectangle whose height is equal to or smaller than the standard character size is checked. Then, calculate the ratio (aspect ratio of the character rectangle) of the total number of character rectangles with a rectangle height less than the standard character size and the total number of rectangles with a rectangle width less than the standard character size (rectangle height less than the standard character size). Ask.

This ratio shows the approximate degree of fusion between characters. As described above, in the case of a document in which characters are closely packed, such as a newspaper, this ratio has a fairly small value. However, in a document such as a word processing document in which characters are not sufficiently spaced, this ratio is close to 1. Become.

In the next step 30B, the optimum parameter is determined by applying the ratio obtained in the previous step to the above-mentioned table of optimum parameters (character-to-character integrated threshold) or a general formula.

Even if the fixed value is used as the intercharacter integration parameter, there is no inconvenience when the document to be processed is specialized. However, when processing various documents, it was necessary to manually set the optimum parameters according to the documents. According to this example and each example described later, it is possible to eliminate such a troublesome manual setting, set optimum parameters for various documents, and perform appropriate character-to-character integration.

Example 2 of Parameter Setting Process of Parameter Setting Section 10A The process contents of this example will be described with reference to the schematic flow chart shown in FIG. Here, it is assumed that the line direction on the document image is horizontal, as in the above description.

In step 40A, regarding the rectangles classified into characters in the rectangles extracted from the compressed image, two character rectangles adjacent to each other in the row direction (horizontal direction) are integrated except for the horizontal distance (character spacing) condition. Check if there is any inconvenience. For example, when the difference between the rectangle heights of one character rectangle and the other character rectangle exceeds a predetermined limit (for example, in a vertical writing document, one character rectangle is a character on the body line and the other is the body text). (If it is a title character for), it is determined that both character rectangles should not be integrated, and if they do not exceed the limit, it is determined that they can be integrated. Do this for all adjacent character rectangles.

In step 40B, the previous step 40
The horizontal distance between adjacent character rectangles determined to be integrated in A is measured, and a histogram of the horizontal distance is obtained.

In step 40C, the horizontal distance histogram obtained in the previous step 40B is used as a frequency in the table or the general formula representing the relation between the character spacing (horizontal distance) and the optimum intercharacter integrated parameter prepared in advance. Determines the optimal intercharacter integration parameter by fitting the highest horizontal distance value.

The method of this example is excellent in the parameter optimization capability for a special case where the character setting process example 1 is difficult to optimize and the character spacing is wide.

Example 3 of Parameter Setting Process of Parameter Setting Unit 10A The process contents of this example will be described with reference to the schematic flow chart shown in FIG. Note that a horizontal document is assumed here.

Step 50A corresponds to step 30A in FIG.
This is a step for performing the same processing as. Step 5
0B selects the next processing method according to the ratio (degree of inter-character fusion) obtained in the previous step 50A. For example, when the ratio is close to 1 (when the character spacing is wide and the processing method shown in FIG. 4 is more suitable than the processing method shown in FIG. 3), step 50D is selected as the next processing, and if not, step 50D is selected. Select 50C. Step 50C is the same processing step as step 30B in FIG. 3, and steps 50D to 50F are the same processing steps as steps 40A to 40C in FIG.

When the standard character size is used for the processing for character area integration or area type discrimination (for example, in the case of the device according to the above-mentioned Japanese Patent Application No. 4-267313), description will be given with reference to FIG. The processing method may reduce the processing time. That is, in this example, if there is no problem in terms of accuracy in the processing method of FIG. 3, the processing of step 50C is selected so that the benefit of shortening the processing time can be expected, and the character spacing is wide and the accuracy of FIG. If the above processing method is preferable, the processing is switched to the processing after step 50D.

Example 1 of character integration processing by the character integration unit 10B: To determine whether or not to integrate a character rectangle and another character rectangle into the same line, the distance between the character rectangles in the line direction and the line It is general to consider the character rectangle position in the vertical direction with respect to (virtual line). Here, not only when the threshold value (integrated parameter) related to the distance between the character rectangles in the line direction is not optimized with a fixed value, but also when it is automatically optimized as described above, some wraparound occurs. When there is a punctuation mark that should be integrated into a single line in a word processing document,
There is a risk of failing to integrate certain similar symbols. This is because, in a word processing document or the like, a space between a punctuation mark or a specific symbol (generally referred to as a punctuation mark) similar to the punctuation mark and the next character may be too wide as compared with an ordinary character space.

According to this example, it is possible to avoid the failure of integration of such punctuation marks. The processing contents of this example will be described below with reference to the flowchart of FIG.

In step 60A, it is determined whether the character rectangle to be integrated or focused is a punctuation mark or the like. Information such as the shape, size, and positional relationship on the virtual line of the focused character rectangle can be used for this determination. For example, the focused character rectangle is extremely smaller than other character rectangles,
Alternatively, when it is elongated, it is determined to be a punctuation mark or the like.

When it is determined in step 60A that the focused character rectangle is not a punctuation mark or the like, the parameter setting unit 1
The parameters set in 0A are used as they are in step 60D.
Is used as a threshold for the distance between the character rectangles in the line direction.
However, if it is determined to be a punctuation mark or the like, in step 60C, the parameter set in the parameter setting unit 10A is increased by, for example, the character size, and this is increased in step 60D to determine the character rectangle in the line direction related to the target rectangle. It is used as a threshold value for the distance.

In step 60D, the integration condition of the character rectangle of interest and the character rectangle immediately adjacent to it (assuming processing from the left end to the right end of the line) is checked. If the integration condition is not satisfied, the focused character rectangle and the adjacent character rectangle are not integrated (step 60F). When the integration condition is satisfied, the focused character rectangle and the adjacent next character rectangle are integrated (step 60E).

As described above, when the focused character rectangle is a punctuation mark or the like, the threshold value of the distance between the character rectangles in the line direction is increased. Therefore, the distance between the punctuation mark and the next character is a normal character in a word processing document or the like. If the interval is too wide, it can be merged in the same line. In addition, the parameter setting unit 1
The benefit of being able to tolerate some of the parameters set by 0A can also be expected.

Example 2 of character integration processing by the character integration unit 10B For example, in the case of a vertically written document, the end of the line of the text (text line), the footer (page number etc.) of horizontal writing, and the header (title etc.) Are often close to each other. According to the present example, it is possible to prevent such an erroneous integration of the body line characters that are close to each other and the characters of the footer and the header. Hereinafter, the processing content will be described with reference to the flowchart shown in FIG. 7.

In step 70A, it is checked whether the integration condition regarding the distance in the line direction (text line direction) of the two focused character rectangles is satisfied (whether the distance is less than or equal to the threshold value set by the parameter setting unit 10A). If the integration condition of the inter-character distance is not satisfied, the process branches to step 70F to proceed to the process of the next character rectangle without integrating the character rectangle.

If the integration condition is satisfied, step 7
The size of the character rectangle in 0B (for example, the height of the rectangle)
Check whether the difference between is less than a predetermined value. If the size difference is large, the process branches to step 70F and the process proceeds to the next character rectangle processing without integration. Therefore, for example, as shown in FIG. 8A, when there is a large character rectangle a in which a character string having the number of horizontally written pages is integrated and is close to the character rectangle B of the text line in the vertically written document. In addition, the wrong integration can be prevented.

If it is determined in step 70B that the rectangle size difference is small, in step 70C, one of the two character rectangles of interest has a character rectangle adjacent to the line (text line) in the direction perpendicular to the line. Check if it exists. If it exists, the process branches to step 70F. As a result of such a check, as shown in FIG. 8B, for example, there is a character rectangle b of a footer (the number of pages, etc.) near the text line character rectangle B, and the footer characters are above and below this character rectangle b. When rectangles a and c exist, it is possible to prevent erroneous integration of the character rectangle B and the character rectangle b.

If it is determined in step 70C that there is no vertically adjacent character rectangle, it is checked in step 70D whether the difference between the aspect ratio of one character rectangle and the normal character rectangle is small. When it is determined that the aspect ratio is normal, the character rectangles are integrated in step 70E, and the process proceeds to the next character rectangle.

However, if the difference in the aspect ratio is large, the process branches to step 70F. Therefore, for example, in FIG.
As shown in (c), when a narrow character rectangle a in which a character string with a small footer is integrated exists close to the body line character rectangle B, erroneous integration of the character rectangle a and the body line character rectangle B is prevented. it can.

Processing of the area integrating unit 10C In most documents, a footer or a character string of a header having a line direction different from that of the text line is integrated into the text line because the distance between the text line and the characters of the text line is large. Absent. Further, even when the distance between the characters in the body line and the characters such as the footer is small, the integration is prevented by the inter-character integration processing method shown in FIG. Therefore, a character string such as a footer or a header having a line direction different from that of the body line remains as a short line without being integrated into the body line.

For example, in FIG. 9, a to k are lines obtained by character integration, in which a to i are text lines,
j and k are short lines such as footers.

The area integration unit 10 integrates text lines such as a to i in FIG. 9 into a character area (broken line frame 90 in FIG. 9) by a method similar to the conventional method. Look up,
If a short line has occurred, a line shorter than that line in the direction perpendicular to the main text line is searched for, so that j,
A short line such as k is integrated into another character region (broken line frame 91 in FIG. 9) whose line direction is different from the body region.

Processing of Rectangle Classification Unit 9 An example of rectangle classification will be described. First, the types of rectangles are roughly classified according to the size and shape of the rectangles. For example, as shown in FIG. 10A, the height of the rectangle (height
t) is classified into class 1 to class 3 according to the size relation with the standard character size (size). Also, the width of the rectangle is divided into class 1 to class 3 according to the size relationship with the standard character size. Standard font size is
If you know in advance, you can use that value,
If it is not known in advance, a method of adaptively determining from the distribution of the heights of the rectangles extracted from the document image may be adopted.

The rectangle height class and rectangle width class thus obtained are applied to, for example, the classification table shown in FIG. 11B to determine the rectangle classification. If the rectangle width is class 1 or class 2 and the rectangle height is class 1 or class 2, the rectangle is classified as a character candidate or a capital letter candidate. Rectangle width is class 1 and rectangle height is class 3
If so, it is classified as a vertical ruled line candidate.

For the rectangles classified as the table area candidates,
Scan a rectangular area to extract only black runs longer than a specified horizontal value, and extract the circumscribed rectangle surrounding the connected ones as horizontal fictitious ruled lines. Similarly, extract only black runs that are long in the vertical direction. Then, the circumscribing rectangle surrounding the connected one is extracted as an imaginary ruled line in the vertical direction. Then, paying attention to the number, position and length of the imaginary ruled lines,
Enclosed frame> or <Figure, graph, etc.>.

As for the rectangles that are candidates for horizontal ruled lines, the rectangular area is scanned to extract only black runs that are long in the horizontal direction, and a circumscribed rectangle that surrounds the connected long black runs is extracted in the horizontal direction. It is extracted as an imaginary ruled line, and the rectangle is classified into a <character rectangle> or a <horizontal separator> that vertically divides a character region, focusing on the number, position, and length of the imaginary ruled lines.

For rectangles classified as vertical ruled line candidates, the rectangular area is scanned to extract only black runs that are long in the vertical direction, and a circumscribing rectangle that encloses the connected long black runs is fictitious in the vertical direction. Extracted as ruled lines, focusing on the number, position, and length of the fictitious ruled lines, the rectangles are classified into <character rectangles>, <vertical separators> that divide the character region into left and right, or <figure, graph, etc.>.

Next, another embodiment of the present invention will be described. The system configuration for this embodiment may be the same as the system configuration of the above embodiment shown in FIG. The processing content may be the same as that of the above embodiment except the processing of the character area integrating unit 10. Hereinafter, the processing contents of the character area integration unit 10 in this embodiment will be described. A schematic flow of the processing is shown in FIG.

With respect to the rectangles classified into characters by the rectangle classification unit 9, those having a small distance in the direction of the text line are integrated to extract the line (step 101). The direction of the lines to be integrated may be designated in advance or may be automatically determined.
Further, the threshold value of the inter-rectangle distance at the time of this integration may be designated in advance, or may be adaptively set by the method described in the above embodiment. The rectangles extracted as rows here are only those with a small distance between rectangles, and no special processing is required. Rather, there is a risk that adverse effects will occur if rows with a large distance are integrated into a row.

The lines extracted by this character rectangle integration may have a small width. Such a line may be a single character, a character in a different line direction (for example, a character rectangle that is vertically adjacent to the line direction of interest), or a non-character figure or photograph. Part (for example, captions, characters such as graph coordinate values, etc.)
Is likely to be.

Therefore, in the present embodiment, lines having a small width among the lines obtained by the integration of the character rectangles are classified into "lines other than the body character line", and other large lines, that is, "body lines". It has an attribute different from "character line" (step 10
2). To describe this classification in more general terms, the size of the line obtained by integrating the character rectangles in the text line direction, the line distance to the next line, the distance to the rectangle classified as other than the character, etc. It means that the information is classified into "body text line" and "line other than text line".

Except for the "text line" classified in the previous step, around or inside (that is, overlapped with) a rectangle (referred to as "figure or other rectangle") classified into a rectangle that is neither a character nor a ruled line by the rectangle classification unit 9. Is present (step 103).

Regarding the "lines other than the text line" existing around or inside the "figure or other rectangle", a line that can be integrated in the direction perpendicular to the text line direction is searched for around the line, and such a line is searched. If a line that can be integrated is found, the "line other than the text line" is finally classified into "vertical line" (step 104), integrated with the searched line, and "vertical lines" having a closer distance are integrated ( Step 105).

Even if there are "lines other than the text line" located around or inside the "figure or other rectangle" and no line that can be integrated in the vertical direction is found, they are finally classified into "figure lines" (step 104), or merge with a close "row of the figure" or delete (step 105).

In step 102, "lines other than text line"
Of the lines classified into, those that do not exist around or inside the "figure or other rectangle" are finally classified into "lines other than the text line" (step 104), and are merged between lines that have the same attribute and are close to each other. (Step 105).

Next, with respect to the "text line", those close to each other are integrated to generate a text area (step 1).
06).

According to a modification of the present embodiment, the rectangle classification unit 9 further subdivides the "drawing rectangle" into the "drawing rectangle" and the "photographic rectangle". Then, in step 105, the "line other than the text line" in the "photograph rectangle" is erased or changed to the photo attribute. Since halftone photo areas are often classified into small rectangular blocks compared to line figures, this modification is effective in preventing halftone photo areas from falling apart. Is. Further, the character string in the line figure area can be extracted as a character string without being erased.

[0076]

As is apparent from the above description, according to the inventions of claims 1 to 7, it is possible to perform area division of various documents having different character intervals without the need for manpower. It is possible to prevent line integration of punctuation marks and similar symbols in word processing documents, and erroneous integration of character strings such as headers and footers of vertical writing documents with character strings of text lines.
Further, a character string such as a header or a footer can be integrated as a character area different from the body area, which enables highly accurate area division for various documents. Further, according to the inventions of claims 8 to 12, the lines in which the character rectangles are integrated in the body character line direction are classified, and the body character lines and other lines are classified, and the lines other than the body character lines are further classified. By doing so, it is possible to extract the body text area more accurately, and even if there is a high rate of failure in rectangle classification for each rectangle, the rectangles can be reclassified strictly based on the result of character recognition. It is possible to substantially improve the classification accuracy of rectangles without performing inefficient processing, prevent the generation of unnecessary character areas, and prevent unnecessary division of the photo area.

[Brief description of drawings]

FIG. 1 shows a schematic configuration of a system according to an embodiment of the present invention.

FIG. 2 shows a configuration example of a character area integration unit.

FIG. 3 shows a flow of a first example of parameter setting processing.

FIG. 4 shows a flow of a second example of parameter setting processing.

FIG. 5 shows a flow of a third example of parameter setting processing.

FIG. 6 shows a flow of an example 1 of inter-character integration processing.

FIG. 7 shows a flow of an example 2 of inter-character integration processing.

FIG. 8 shows an example of a character rectangle excluded from the integration target with the character rectangle of the text line.

FIG. 9 is an explanatory diagram of region integration processing.

FIG. 10A is an explanatory diagram of classification of height and width of a rectangle. (B) It is explanatory drawing of the table of the attribute classification by the height and width of a rectangle.

FIG. 11 shows a processing flow of a character area unifying unit in another embodiment of the present invention.

[Explanation of symbols]

 1 Scanner 2 Image Input Section 3 Image Memory 4 Image Display Section 5 Display 6 Area Identification Section 7 Image Compression Section 8 Enclosing Rectangle Detection Section 9 Rectangular Classification Section 10 Character Area Integration Section 10A Parameter Setting Section 10B Character Integration Section 10C Area Integration Section 11 Character recognition part

Claims (12)

[Claims] 1. A rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like with a predetermined magnification, the extracted rectangle is classified into a character rectangle and a rectangle other than the character rectangle, and the character rectangle is displayed. In the area division method that integrates in the direction, the degree of fusion between characters due to image compression is estimated by taking the ratio of the height and width of the rectangle of the character, and according to this degree of fusion, in the direction of the line of the rectangle of the character. An area dividing method characterized by setting a threshold value of a distance between rectangles for the integration of the. 2. A rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like at a predetermined magnification, the extracted rectangle is classified into a rectangle of a character and a rectangle other than that, and a rectangle of the character is displayed. In the area dividing method for integrating in the direction, a distribution of distances between rectangles of characters adjacent to each other in the row direction is measured, and based on the distribution, a distance between rectangles for integrating rectangles of the characters in the row direction is calculated. A region dividing method characterized by setting a threshold value. 3. A rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like with a predetermined magnification, the extracted rectangle is classified into a character rectangle and other rectangles, and a character rectangle is displayed. In the area segmentation method that integrates in the direction, the degree of fusion between characters due to image compression is estimated by taking the ratio of the height and width of the character rectangle, and the space between rectangles for integrating the character rectangle in the row direction is estimated. As a method of setting the distance threshold, a method of setting by the fusion degree or a method of setting by the distribution of the distance between the rectangles of the characters that are close to each other in the row direction is selected according to the magnitude of the fusion degree. Characteristic area division method. 4. A rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like with a predetermined magnification, the extracted rectangle is classified into a character rectangle and other rectangles, and a character rectangle is displayed. In the area dividing method of integrating in the direction, during the process of integrating the character rectangle in the line direction, it is determined whether the rectangle of the character of interest is a punctuation mark or a rectangle of a specific symbol similar thereto, and When it is determined that the rectangle is a similar specific symbol, the threshold value of the inter-rectangular distance related to the integration of the focused rectangle and the rectangle of the next character is increased as compared with the case of the integration of the rectangles of the characters other than the punctuation mark or the similar specific symbol. An area dividing method characterized by: 5. A rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like with a predetermined magnification, the extracted rectangle is classified into a rectangle of characters and a rectangle other than that, and a rectangle of characters is displayed. In the area dividing method for integrating in one direction, during integration processing of a character rectangle in the line direction, one of the two rectangles to be integrated and the other rectangle are rectangles of characters in different directions. A region dividing method, characterized in that, when it is determined whether or not it is a rectangle of characters in lines in different directions, integration of the two rectangles is prohibited. 6. If a line having a shorter length in the row direction is present in the lines obtained by integrating the rectangles of the characters in the line direction, the line is more perpendicular to the line direction than the short line. 6. The area dividing method according to claim 5, further comprising the step of searching for short rows in a direction and integrating them. 7. The area dividing method according to claim 5 or 6, wherein a threshold value of an inter-rectangular distance for integrating character rectangles in the row direction is set in the area dividing method according to claim 1, 2 or 3. An area dividing method characterized by determining a threshold value of a distance between rectangles. 8. A rectangle circumscribing a black pixel connected component is extracted from an image obtained by compressing an image of a document or the like with a predetermined magnification, and the extracted rectangle is classified into a character rectangle and other rectangles, and the character rectangle is the text. In the area division method that integrates in the character line direction and integrates the lines obtained by combining the rectangles to cut out the character part area, the line obtained by combining the character rectangles in the body character line direction is the body character line. And a line other than the text line, and by giving a line other than the text line a unique attribute different from the text line, the line other than the text line can be merged when the line is integrated into the character area. A region segmentation method characterized by avoiding the integration of lines and body text lines. 9. The area dividing method according to claim 8,
Lines other than text lines existing inside or around the rectangle classified into rectangles other than characters and ruled lines by the classification of rectangles, but lines that can be integrated around the text line in the direction perpendicular to the text line direction. A region segmentation method characterized by classifying it into those that can be found and those that cannot be found, and having unique attributes for each. 10. The region dividing method according to claim 9, wherein lines having the same attribute are integrated with respect to lines other than the text character line. 11. The area dividing method according to claim 9, wherein a line having an attribute other than a text character line is erased. 12. The area dividing method according to claim 8, wherein a rectangle classified into a rectangle other than a character or a ruled line by the rectangle classification is classified into a figure rectangle and a photograph rectangle, and the inside of the photograph rectangle is divided. A region segmentation method characterized by erasing lines other than a certain text line or adding a photo attribute.