JPH08202819A - Underline extraction method - Google Patents

Abstract

PURPOSE: To precisely extract an underline filled in a document picture even if an inclination and blurring exist or it is brought into contact with a character. CONSTITUTION: A document is sampled and the values of respective pixels are quantized to the binary of '1' for black and '0' for white in a document input stage 1. In a pixel substitution stage 2, the black pixels are substituted for white pixels if more than one black pixels are included at every prescribed number of continuous pixels. In a vertical black run extension stage 3, the picture where the black pixels continue in a vertical direction is executed. In a horizontal black run extraction stage 4, a horizontal black run having more than prescribed length is extracted for the picture obtained in the vertical black run extraction stage 3. In a vertical black run degradation stage 5, a degradation processing is executed for the number of times similar to that of an extension processing executed in the vertical black run extension stage 3. In an underline candidate extraction stage 6, an underline candidate is extracted. In a lower contour extraction stage 7, the lower contour of the underline candidate is extracted. In an underline extraction stage 8, the lower contour is scanned from a left end and the underline is extracted.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for extracting underlines printed or handwritten below a character string in a document.

[0002]

2. Description of the Related Art Conventionally, when a keyword or the like written in a document is recognized by OCR, the position of the character to be recognized is marked by O.
As a method for instructing the CR, an underline is drawn below the keyword to be recognized.
As a method for extracting this underline, FIG.
As shown in FIG. 3, a method of extracting a group of figures in which the run length of black pixels is a certain length or more and extracting the figure as an underline can be considered. In FIG. 17, among the black runs in the horizontal direction, those whose continuous amount exceeds a certain value are collectively made into one figure and extracted as an underline. As a result, the rectangular area surrounded by the broken line is extracted as the underline area.

[0003]

This method was effective for underlines output by using a word processor or the like, and underlines horizontally written by using a ruler, but it is possible to write by handwriting freely. However, there is a drawback in that it is difficult to extract an underline that is subject to fluctuation due to camera shake and an underline that is obliquely input even if it is a straight line. Further, there is a drawback that extraction cannot be performed if there is a white pixel such as a blur in the middle of the black pixel run. As a method of compensating for this drawback, there is a method of tracing an outline and extracting an underline having an elongated circumscribed rectangle of an outline point as shown in FIG. In the method of FIG. 18 as well as in the method of FIG. 17, the portion surrounded by the broken line is extracted as an underline, but in this method as well, the characters written above the underline become the underline as shown in FIG. When touched, the contour tracing extends to the character portion, which makes it difficult to extract the underline.

An object of the present invention is to accurately extract an underline entered in a document image even if there is an inclination or a hand shake, or even if it is in contact with a character, and for a blur or a broken line at the time of input. It is to provide an underline extraction method that can deal with the problem to some extent.

[0005]

An underline extraction method according to the present invention comprises a document input step of inputting a document and converting the input document into a document image represented by a binary number by sampling and quantization. , The document image obtained by the document input step is scanned, and for every fixed number of consecutive pixels continuous in the horizontal direction, if the consecutive pixels include a predetermined number or more of black pixels, all white pixels are For the pixel replacement step of performing the process of replacing with black pixels and the document image obtained in the pixel replacement step,
A vertical black run decompression step in which a black run in which black pixels are continuous in the vertical direction is performed once or more, and a horizontal black run in a certain range length is calculated from a document image obtained in the vertical black run decompression step. The same as the decompression process of the black run in the vertical black run decompression stage in which the black run in which the black pixels are continuous in the vertical direction is applied to the document image obtained in the horizontal black run extraction stage and the horizontal black run extraction stage. By taking the logical product of the vertical black run degeneracy stage performed only the number of times, the document image obtained in the vertical black run degeneracy stage, and the pixel replacement image obtained in the pixel replacement stage,
An underline candidate extraction step of extracting an underline candidate image, and a lower contour extraction step of extracting a lower contour of the underline by performing a logical operation on the underline candidate image obtained in the underline candidate extraction step, Among the lower contour points obtained in the lower contour extraction step, there is an underline extraction step of extracting, as an underline, a contour point sequence that is continuous for a certain length or longer.

In the underline lower contour extraction processing, the vertical black pixel group of the underline candidate image is shifted by one pixel in the upward direction, and the exclusive OR of the image before shift and the image after shift is calculated. This is performed by obtaining the taken image and obtaining the logical product of the image and the inverted image of the shifted image.

According to another underline extraction method of the present invention, by searching for a lower contour point of another underline existing within a certain distance above the underline extracted in the underline extraction step, the document is extracted. The method further includes a double underline extraction step of extracting the double underline written or printed below the character string of the image.

[0008]

When a predetermined number or more of black pixels are included in every predetermined number of continuous pixels in the original document expressed in binary, all white pixels are replaced with black pixels and continuous. A pixel replacement image in which black pixels are replaced is created. As a result, the underline can be stably extracted even if a broken line is used as the underline or is faint. Next, in order to stably extract a horizontal black pixel run of a fixed length even if the underline is tilted, the vertical black pixel run is expanded and then fixed length The horizontal black pixel run is extracted, and the vertical black pixel run is degenerated. Next, the candidate area of the underline is extracted by taking the logical product of the image subjected to the degeneracy processing and the pixel replacement image, and the lower contour point of the underline candidate area is extracted by the combination of the shift of the document image and the logical operation. After that, the lower contour points are traced to extract the underline.

Therefore, it is possible to accurately extract the underline even if the underline entered in the document image has an inclination or a hand shake or is in contact with a character.

According to the present invention, not only handwritten underlines written in a document but also underlines and ruled lines printed in the document can be extracted by the same method. Further, not only horizontal ruled lines but also vertical ruled lines can be extracted by the same method.

[0011]

Embodiments of the present invention will now be described with reference to the drawings.

FIG. 1 is a flowchart showing an underline extraction method according to an embodiment of the present invention.

The underline extraction method of this embodiment includes a document input step 1, a pixel replacement step 2, a vertical black run decompression step 3, a horizontal black run extraction step 4, a vertical black run degeneration step 5, and an underline candidate extraction step 6. It is composed of a lower contour extraction stage 7 and an underline extraction stage 8.

In the document input step 1, a document is sampled by using a facsimile, a scanner or the like, and the value of each pixel is quantized into a binary value of 1 for black and 0 for white to create an original image. In the pixel replacement step 2, for every fixed number of pixels continuous in the horizontal direction of the created original image, if at least one black pixel is included therein, all white pixels are replaced with black pixels.
In the vertical black run extension stage 3, in order to stably extract the horizontal black run even if the underline is inclined,
The input image (black pixel) is thickened (extended) by one pixel vertically downward. In the horizontal black run extraction step 4, a horizontal black run having a length equal to or more than a certain length is extracted from the image obtained in the vertical black run extension step 3. In the vertical black run degeneracy stage 5, the decompression process is performed in the same number as the decompression process performed in the vertical black run decompression stage 3 in order to restore the decompressed vertical black run to the original state. The degeneration process is realized by shifting the entire image by one pixel in the direction opposite to the decompression process performed in the vertical black run decompression stage 3, that is, in the upward direction, and taking the logical product with the image before the shift. When the degeneracy process is further performed, this process may be repeated. In the underline candidate extraction step 6, the underline candidates are extracted by taking the logical product of the image obtained in the vertical black run degeneracy step 5 and the image obtained in the pixel replacement step 2. The logical product is calculated because the underline comes into contact with the ruled line in the vertical black run decompression stage 5 when the underline exists close to the ruled line, and the underline remains even after the vertical black run degeneracy stage 5. If it is close to the ruled line,
This is because the ruled line and the underline may not be separated. In the lower contour extraction step 7, the lower contour of the underline candidates obtained in the underline candidate extraction step 6 is extracted. In the underline extraction step 8, the pixel array obtained in the lower contour extraction step 7 is scanned from the left end, and those having a certain length or more are extracted, and the entire figure including the entire lower contour points is extracted as an underline. .

Next, the specific operation of the underline extraction method of this embodiment will be described with reference to FIGS.

FIG. 2 shows an example of a document image binarized in the document input stage 1. The underline is input with a tilt, and includes white pixels due to blurring.
On the other hand, in the pixel replacement step 2, if at least one of the four consecutive horizontal pixels with Δ in FIG. 2 as the left end includes a black pixel, all white pixels in this pixel group are replaced with black pixels. As a result, the original image shown in FIG. 2 becomes as shown in FIG. As a result, the underline can be stably extracted even if a broken line is used as the underline or the line is faint. In the next vertical black run decompression stage 3, a black run decompression process is performed in which each black pixel group in the vertical direction is expanded vertically by one pixel. For example, a vertical black pixel group indicated by an arrow in FIG.
(A)) is shifted down by one pixel (FIG. 4 (b)), and is logically ORed with the unshifted image (FIG. 4 (a)) (FIG. 4).
By performing (c)), the vertical black run extension processing is realized.
FIG. 5 shows an image after the vertical black run expansion processing. When further decompression processing is performed, this processing may be repeated.
In the horizontal black run extraction step 4, a horizontal black run having a length of 10 or more is extracted from the image obtained by the vertical black run expansion processing (FIG. 6). In FIG. 6, the horizontal black run shown in FIG. 3 is indicated by a wave underline. On the other hand, the image shown in FIG.
FIG. 7 shows the extracted black runs of 0 or more. It can be seen that horizontal black runs are extracted more stably than in FIG. In the example shown here, the expansion process is performed only on the original image in the downward direction, but the expansion process may be performed in the upward direction or both directions. Further, although the expansion processing for one pixel is performed a plurality of times, the expansion processing for two or more pixels may be performed once. In the vertical black run degeneracy step 5, the bottom black pixel of each black pixel group in the vertical direction of the image obtained in the horizontal black run extraction step 4 is changed to a white pixel. FIG. 8 shows how the vertical black run degeneracy step 5 is performed. That is, the vertical pixel group (FIG. 8A) is shifted upward by one pixel (FIG. 8B), and the logical product of this and the original pixel group (FIG. 8A) is obtained ( FIG. 8 (c)). FIG. 9 shows the result of performing the vertical black run degeneracy process twice on the horizontal black run (FIG. 7) extracted in the horizontal black run extraction step 4. By performing the decompression / degeneration process, it can be seen that the horizontal black run having a length less than 10 in FIG. 3 is also extracted. In the underline candidate extraction step 6, the logical product of the image obtained in the vertical black run degeneracy step 5 and the image obtained in the pixel replacement step 2 is calculated. This is because when the underline is close to a figure such as a ruled line, the underline may not be separated from the figure such as a ruled line even by the degeneracy process. This will be specifically described with reference to the drawings. FIG. 10 shows an example in which a ruled line exists near the underline shown in FIG. FIG. 11 is a diagram in which the vertical black run is extended to this underline, and FIG. 12 is a diagram in which the degeneracy process is performed. In this way, the underline and the ruled line come into contact with each other, resulting in one figure. Even in such a case, the logical product of the image of FIG. 10 and the image of FIG. 12 is taken to stably extract the ruled line by the run length, and the portion contacted by the expansion / reduction process is shown in FIG.
It will be possible to restore like. Next, in the lower contour extraction step 7, the lower contours of the underline candidates obtained in the underline candidate extraction step 6 are extracted. The principle of lower contour extraction is shown in FIG. FIG. 13A shows an image in the vertical direction of Δ in FIG. First, an image (FIG. 13 (b)) obtained by shifting the image of FIG. 13 (a) upward by one pixel is created, and an image obtained by exclusive ORing with the original image (FIG. 13 (a)) ( FIG. 13C) is obtained. By this logical operation, an image in which the third and sixth pixels from the bottom of the image (FIG. 13A) are black pixels is obtained. This image (Fig. 13 (c)),
Image (Fig. 13 (d)) that is the reverse of the image (Fig. 13 (b))
The original image (Fig.
An image (FIG. 13E) in which only the third pixel from the bottom of (a) is a black pixel can be obtained. In this way, the lower contour point of the underline can be obtained from the original image only by a logical operation between the images. As described above, by applying the logical operation in combination to the image obtained in the underline candidate extraction step 6, the lower contour points of the underline indicated by a double circle in FIG. 14 are extracted. In the final underline extraction step 8, as shown in FIG. 15, first, the left edge (A) of the contour point obtained in the lower contour point extraction step 7 is detected. Next, the lower contour point is traced in the right direction (arrow direction). If there is no continuous lower contour point in the right direction, the lower contour point is traced by changing the step. In FIG. 15, the tracking step is moved up by one to track the lower contour point. Depending on the image, the lower contour point may jump two or more steps. In this case, a constant value may be provided, and lower contour points having steps within the range of the value may be continuously traced. Here, if a too large value is set, another underline that is originally different will be extracted as the same underline. When the tracking is completed, the number of continuous pixels of the lower contour point is counted, and for the objects having a certain length or more, the entire figure including the entire lower contour point is extracted as an underline.

It should be noted that when a document includes a linear figure, such as a ruled line, and this is distinguished from an underline, the underline may be doubled. FIG. 16 shows a double-lined underline. If there is another lower contour point ◆ of another underline within a certain distance above the lower contour point ⊚ extracted in FIG. 16, this is extracted together as a double underline so that a normal ruled line in the document is obtained. It is possible to distinguish In addition, as shown in FIG. 16, when the upper line of the underlined doubly drawn is wider than the width of the lower line, the portion protruding from the lower line,
That is, the lower contour pixel may be searched toward both ends as indicated by the arrow in FIG.

The present invention has been described above by taking the extraction method of an underline written by handwriting under a character or the like in a binary document image input from a scanner, a facsimile or the like as an example. It is obvious that the underlines and ruled lines printed on the document are not limited to the above, and can be extracted in exactly the same manner.
Further, it is obvious that not only the underline written or printed below the character string, but also a straight line such as a horizontal ruled line or a vertical ruled line can be extracted by the same method.

[0019]

As described above, the present invention has the following effects. (1) According to the first aspect of the invention, the underline can be accurately extracted even if the underline entered in the document image has an inclination or camera shake or is in contact with a character. In addition, it is possible to deal with blur and broken lines at the time of input to some extent. (2) The invention of claim 3 can extract double underlines.

[Brief description of drawings]

FIG. 1 is a flowchart showing an underline extraction method according to an embodiment of the present invention.

FIG. 2 is a diagram showing an example of a binarized document image.

FIG. 3 is a diagram showing a document image after pixel replacement.

FIG. 4 is a diagram showing vertical black run extension processing.

FIG. 5 is a diagram showing an image after vertical black run expansion processing.

FIG. 6 is a diagram showing an example in which a black run having a length of 10 or more is extracted.

FIG. 7 is a diagram showing an example in which a black run having a length of 10 or more is extracted by performing vertical black run extension processing twice.

FIG. 8 is a diagram showing vertical black line expansion processing.

FIG. 9 is a diagram showing horizontal black runs extracted by two vertical black run expansion / reduction processes.

FIG. 10 is a diagram showing an example of underlines and ruled lines that are close to each other.

FIG. 11 is a diagram showing underlines and ruled lines that have undergone decompression processing.

FIG. 12 is a diagram showing contacted underlines and ruled lines after the shrinking process.

FIG. 13 is a diagram showing a lower contour extraction process.

FIG. 14 is a diagram showing an example of extracting underline candidates.

FIG. 15 is a diagram showing an underline candidate extraction process.

FIG. 16 is a diagram showing an example of extracting a double underline.

FIG. 17 is a diagram showing underline extraction using run length.

FIG. 18 is a diagram showing underline extraction using a circumscribing rectangle.

FIG. 19 is a diagram showing an example in which underline extraction using a circumscribed rectangle is difficult.

[Explanation of symbols]

 1 Document Input Stage 2 Pixel Replacement Stage 3 Vertical Black Run Expansion Stage 4 Horizontal Black Run Extraction Stage 5 Vertical Black Run Degeneration Stage 6 Underline Candidate Extraction Stage 7 Lower Contour Extraction Stage 8 Underline Extraction Stage

Claims (3)

[Claims] 1. An underline extraction method for extracting an underline printed or handwritten below a character string in a document, the method comprising inputting a document, and sampling and quantizing the input document.
A document input step of converting into a document image represented by a numerical value, and scanning the document image obtained by the document input step,
A pixel replacement step of performing a process of replacing all white pixels with black pixels when a predetermined number or more of black pixels are included in each of a fixed number of consecutive pixels continuous in the horizontal direction; From the document image obtained in the vertical black run decompression stage, a vertical black run decompression stage in which a black run in which black pixels are continuous in the vertical direction is decompressed once or more for the document image obtained in the pixel replacement stage; A horizontal black run extraction step of extracting a horizontal black run of a certain range of length, and degeneracy of a black run in which vertical black pixels are continuous with respect to the document image obtained in the horizontal black run extraction step. A vertical black run degeneracy step in which processing is performed the same number of times as the decompression processing performed in the vertical black run decompression step; a document image obtained in the vertical black run degeneracy step; and a pixel replacement image obtained in the pixel replacement step. By taking the logical product of An underline candidate extraction step of extracting an underline candidate image; a lower contour extraction step of extracting a lower contour of the underline by performing a logical operation on the underline candidate image obtained in the underline candidate extraction step; An underline extraction step of extracting, as an underline, a series of contour points having a predetermined length or more among the lower contour points obtained in the lower contour extraction step. 2. The underline lower contour extraction processing shifts the vertical black pixel group of the candidate image of the underline in the upward direction by one pixel, and the exclusive logic of the image before the shift and the image after the shift is used. 2. The underline extraction method according to claim 1, which is performed by obtaining a summed image and obtaining a logical product of the image and an image obtained by inverting the shifted image. 3. Searching for a lower contour point of another underline existing within a certain distance above the underline extracted in the underline extraction step, thereby writing under the character string of the document image or The underline extraction method according to claim 1, further comprising a double underline extraction step of extracting the printed double underline.