KR100310306B1 - Method of analyzing document image on computer and recording media recording it - Google Patents

Abstract

The present invention relates to a method of analyzing a form document image on a computer and a recording medium recording the same. For automatic processing of a form included in a document image, the line segment of the form is extracted and the form item is extracted to form the form. By processing, the purpose of the present invention is to provide a method of analyzing a document image, which can effectively process printed documents containing various structures, hand-drawn tables, and complex form documents, and a computer-readable recording medium recording the same. have.

A format document image analysis method for automatically processing a form from an image including a form document includes: a first step of extracting a form segment by analyzing an image including the form document by mathematical morphology; A second step of filtering the line segment extracted in the first step and extracting the formatting line by grouping the filtered line segments; And the vertical extracted in the second step. And a third step of extracting items of a proper form from a horizontal line.

Description

Image Analysis Method of Form Document on Computer and Recording Media Recording the Same

The present invention relates to a method for analyzing a form document image on a computer and a recording medium recording the same. More particularly, the line segment of a form is extracted and an item of a form is extracted for automatic processing of a form included in the document image. The present invention relates to a format document image analysis method and a recording medium capable of effectively processing a printed document, a hand-drawn table, a complex form document, and the like, by processing a form.

The format document is in the same format, even if the position and size of each item are strictly defined, and if the position and size of the item are deformed to the extent that they maintain their topology and content. Can be divided into logical formats.

When processing a form document in a physical format, only a portion of the form may be removed from the scanned document image, and then the corresponding portion may be processed using OCR. However, this method requires that prior knowledge of all forms to be interpreted is registered as a form model. However, since most documents used in an office are modified by each user to be re-edited or resized according to a predetermined format, in order to use the conventional method, it was necessary to register several form models for one form. For this reason, the user has a lot of difficulties in automatically processing the scanned form document image.

Accordingly, the present invention is to solve the problems of the prior art as described above, an object of the present invention is to extract the line of the form and to extract the items of the form for automatic processing of the form (form) contained in the document image The present invention provides a method of analyzing an image of a format document, which analyzes the structure of a form, separates and recognizes extracted form items into a file format that can be reused, and a computer-readable recording medium recording the same.

1 is a flowchart illustrating a method of analyzing a form document according to the present invention.

2 is a detailed flowchart of the segment extraction step of the form according to the present invention.

3 and 4 show examples processed by the present invention.

The method of the present invention for achieving the above object is a form document image analysis method for automatically processing a form from an image including a form document, the form segment is extracted by analyzing the image containing the form document by mathematical morphology First step; A second step of filtering the line segment extracted in the first step and extracting the formatting line by grouping the filtered line segments; Vertical extracted in the second step. Extracting items of a proper format from a horizontal line; And a fourth step of recognizing the extracted format item by separating individual characters and then converting the file.

In addition, the recording medium according to the present invention, in a computer-readable recording medium for extracting the form through the structural analysis from the image containing the form document, by analyzing the image containing the form document through a morphological closing operation Then, extracting a connecting element that becomes a horizontal segment or a vertical segment through a morphological opening operation; Filtering each of the extracted line segments with respect to a line segment having a threshold value T and extracting a format line by grouping all the filtered horizontal line segments; Extracting form items from the extracted horizontal and vertical form lines; And a program for executing the step of converting a file after recognizing the extracted form item separately by character.

This method does not process only the matching of images with the conventional method, but analyzes the structure of the format document and uses its physical coordinates and logical structure together. When the logical structure is used to match the format document, it is possible to appropriately cope with a change in the position and size of the format document, thereby solving the problem of having to repeatedly register the same form, such as when classifying the form into a physical form.

Further, according to the present invention, even when the line of the form which was the cause of the item extraction error (error rate of about 11.3% to 16.7%) is partially broken in the conventional method, and when the written character and the line of the form are in contact with the item, It is also possible to process complex forms, such as when a rectangle is enclosed within another rectangle. Therefore, the error rate of item extraction can be significantly lowered compared with the conventional method.

Hereinafter, exemplary embodiments of the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a view for explaining a method for analyzing a form document according to the present invention, and FIG. 2 is a detailed representation of the line segment extraction step 11 of the form of FIG. 1.

The analysis method of the present invention comprises the steps of extracting the form line by mathematical morphology (mathematical morphology), the step of extracting the form line through the filtering and grouping (grouping) of the extracted line segment, and the extracted vertical. It consists of extracting items of proper form from the horizontal line.

Similarly, the computer-readable recording medium may extract the formatting line by mathematical morphology, filter the extracted line segment, and group the filtered line segment to form the formatting line. A program for executing the step of extracting, extracting items of a proper format from the extracted vertical and horizontal lines, and separating and recognizing the extracted items individually are recorded.

Looking at the configuration of the present invention in more detail with reference to Figures 1 and 2 as follows.

First, a process of extracting line segments of a form (11 of FIG. 1) will be described.

For the scanned form document image, first, the broken portions of the form may be connected through morphological closing operations (21 and 22) as shown in FIG. 2, and then the morphological opening operation is performed. (23,24) (connected component being a horizontal segment or a vertical segment through morphological opening operations C _h or C _v Is extracted as in Equation 1.

Here, the morphological closing operation refers to performing the erosion operation 22 after performing the Boolean operation 21, and the morphological opening operation refers to performing the Boolean operation 24 after performing the erosion operation 23.

[Equation 1]

here, B _h Wow B _v Is [1 ... 1] of width L, B ′ _h Wow B ′ _v Is the [1 ... 1] width L 'when the widest gap is L', Θ is the erosion, Represents a division.

The method of extracting the segment of the form by such mathematical morphology is irrelevant to the contact between the segment of the character and the form document. In addition, although the allowable value of the inclination degree of the form differs according to the thickness of the line segment width of the form, in a general form document having a line thickness of about 0.4 mm, a form inclined about 12 ° can be processed.

Next, look at the process of filtering and grouping the extracted line segment (12 of Figure 1) through the above process.

Among the extracted line segments, horizontal lines with large heights and slopes and vertical lines with large and small slopes are filtered, and when all slopes exceed the threshold T for all the line segments. Where threshold T is T = tg ^-1 (1/50), T = for dotted line format tg ^-1 (1/24).

After performing a filtration process on all line segments, horizontal (or vertical) segments in close proximity to one another (or vertical) are grouped into a single line to find the lines of the form. Here, the grouping process aligns all horizontal line segments from top to bottom, and initializes the current number of formatting lines to a value of '0'. It then determines if the current line is close to the current line, if it is close, includes the current line in the current line, if not, creates a new line to increase the number of lines by 1, and then adds the current line to the current line. Include it in the line. This process is performed for all horizontal lines.

On the other hand, in determining whether the current line segment is close to the current line, if two segments i and j are on the same side as the standard line, d _ij Is | d _i -d _j |, if two segments i and j are on the other side of the standard line, the distance d _ij Is d _i + d _j | Here, the standard line segment has the same slope as the format line, and the line segment passes the longest line segment of the form line, and d _i and d _j refer to the standard line segment and the distance from the center of each line segment.

Next, look at the step of extracting the item of the form from the line extracted through the process (13 of Fig. 1).

In order to determine whether the horizontal or vertical line of the extracted form is a valid edge of the rectangle surrounding the form's items, the following conditions are used.

First, if the sum of the rectangles covers an entire candidate rectangle, the candidate rectangle is excluded from consideration.

If the edges of the other rectangles cover the entire edge of one of the candidate rectangles, the candidate edges are valid, and thus the candidate rectangles are extracted.

In addition, since the edges of the other rectangles cover a portion of one edge of one candidate rectangle, and the edge of the candidate is valid when the length of the uncovered portion is long enough, the candidate rectangle is similarly extracted.

Each item extracted through the above process is processed through a separate character separation and recognition process (14), and then goes through a file conversion process (15). In the embodiment of the present invention, the file conversion format is processed in the HTML format and represented as an HTML file.

3 and 4 is a view showing the results processed by the present invention, Figure 3 is a view showing an example of the execution of the analysis of the structure of the complex form in contact with the line of the form and the characters written in the form and items, 4 is a diagram showing an example of analyzing the structure of a hand-drawn style in which numbers are written on the items, and expressing the results in HTML.

As described above, the present invention is recorded on a computer-readable recording medium and processed by a computer.

According to the present invention as described above, in order to analyze the structure of any form document filled with contents, including the structural analysis of the logical form which is regarded as the same form even if there is a deformation in the range maintaining the topology, the size and position It is not sensitive to the variation of, allowing for the classification of forms with minimal registration. In addition, if the lines of the form which were difficult to process by the conventional methods are broken, and in the case of the dotted line form, when the handwritten characters and the line of the form are in contact with each other, the form is inclined, and the result is good. Could get That is, as a result of performing various experiments using the present invention, the extraction rate of the item was about 97% or more, which was much improved compared to the conventional art.

Claims (9)

In the form document image analysis method for automatically processing the form from the image containing the form document, A first step of extracting the format line segment by analyzing the image including the format document by mathematical morphology; A second step of filtering the line segment extracted in the first step and extracting the formatting line by grouping the filtered line segments; Vertical extracted in the second step. Extracting items of a proper format from a horizontal line; And And a fourth step of recognizing the extracted format item by separating individual characters and converting the file. The method of claim 1, The first step, A first sub-step of analyzing the input image through a morphological closing operation so that broken portions of the form can be connected; After performing the first sub-step, through the morphological opening operation connecting elements that become horizontal segments or vertical segments ( C _h or C _v Format document image analysis method comprising a second sub-step of extracting. The method of claim 2, Connecting element C _h And connecting elements that become vertical segments C _v Is a format document image analysis method, characterized in that calculated by the following equation. here, B _h Wow B _v Is [1 ... 1] with a width of about 11/6, B ′ _h Wow B ′ _v Is [1 ... 1] about 1/6 wide, Θ is erosion, Each represents a division. The method of claim 1, The process of filtering the line segment of the second step, A format document image analysis method, characterized in that for each of the extracted line segments, the horizontal line segment having a large height and slope, the vertical line segment having a large width and a small slope, and the line segment whose slope exceeds a threshold value T for all segments. The method of claim 4, wherein The threshold T is T = tg ^-1 (1/50), T = for dotted line format tg ^-1 Format document image analysis method, characterized in that (1/24). The method of claim 1, Grouping the filtered line segment of the second step, A first sub-step of aligning all horizontal line segments and initializing the number of current formatting lines to a value of '0'; A second sub step of determining whether the current line segment is close to the current formatting line after alignment and initialization; A third sub step of including the current line segment in the current format line if the current line segment is close to the current format line; A fourth sub step of generating a new formatting line by increasing the number of formatting lines by 1 if the current line segment is not close to the current formatting line, and then including the current line segment in the new formatting line; And And a fifth sub step of repeating the grouping according to the proximity of the line segment and the format line to the last horizontal line. The method of claim 6, In the sub-step, the process of determining whether the current line segment is close to the current formatting line, When the current line segment i and the current format line j are located on the same side of the standard line segment, the distance between the current line segment and the current format line ( d _ij ) Is obtained by | d _i -d _j | If the current line segment (i) and the current format line (j) are located on the other side of the standard line segment, the distance between the current line segment and the current format line ( d _ij ) Is obtained by | d _i -d _j | The distance between the current line segment and the current formatting line ( d _ij ) Is compared with the threshold value to determine whether the proximity document image document analysis. Here, the standard line segment has the same slope as the format line and passes through the longest line segment of the format line, and d _i and d _j refer to the distance between the standard line segment and the current line line to the current line. The method of claim 1, The third step, A first sub step of excluding the candidate rectangle if the sum of the rectangles covers the entire candidate rectangle; A second sub step of extracting the candidate rectangle as a format item when edges of other rectangles cover the entire edge of one candidate rectangle; And And a third sub-step of extracting the candidate rectangle as a form item when the edges of the other rectangles cover a portion of one edge of a candidate rectangle and the length of the uncovered part is long enough. Way. In the computer-readable recording medium for extracting the form through the structural analysis from the image containing the form document, Analyzing the image including the form document through a morphological closing operation, and extracting a connecting element that becomes a horizontal segment or a vertical segment through a morphological opening operation; Filtering each of the extracted line segments with respect to a line segment having a threshold value T and extracting a format line by grouping all the filtered horizontal line segments; Extracting form items from the extracted horizontal and vertical form lines; And A computer-readable recording medium having recorded thereon a program for executing the step of converting a file after recognizing the extracted form item separately.