JP5679284B2 - Pixel number conversion method, program for executing the same, and pixel number conversion device - Google Patents

Description

  The present invention relates to a pixel number conversion method for converting the number of pixels of a binary line drawing image such as a facsimile image or a newspaper page image to perform resolution conversion and image enlargement / reduction.

  In recent years, with the diversification of media devices and the improvement of functions of printing equipment such as CTP and rotary press, the resolution of binary line drawing images such as text images and newspaper images has been increased and reduced (number of pixels conversion). Needs are growing.

  The first prior art of pixel number conversion processing is a conversion technology suitable for handling text images, and the number of pixels is converted using interpolation processing for each pixel by simple thinning processing, logical sum method, and nearest neighbor method. There is a way. In these conversion processes, when the conversion magnification is not an integer multiple, the converted pixel value (density value) is converted into an integer by rounding down or rounding up the calculation result after the decimal point (see, for example, Non-Patent Document 1). .

  The second prior art is a conversion technique suitable for handling a binary pseudo gray image, and there is a method for converting the number of pixels using an interpolation process such as a nine-division method, a projection method, or an inverse distance method. (For example, refer nonpatent literature 1).

  Furthermore, as a third prior art, in Patent Document 1, an interpolation line image table is created by obtaining interpolation bits (bits to be interpolated at the time of conversion) and thinning bits (bits to be reduced at the time of conversion) based on the conversion magnification. Using this, after performing the conversion process from the original image (input binary image) to the converted image (output image), the line widths of the vertical and horizontal lines in the converted image are narrowed or thickened. A pixel number conversion method for correcting the line width in a unified manner is described. Hereinafter, this method will be briefly described with an example of increasing the resolution.

  First, as shown in FIG. 25, the leftmost end of the image in the main scanning direction is set as the 0th bit, and after resolution conversion corresponding to the bit position 112 of the pixel 111 before resolution conversion based on the conversion magnification of the image. The bit position 114 of the pixel 113 is obtained. Then, a bit (interpolation bit) 115 that requires interpolation for resolution conversion is obtained, and the obtained interpolation bit 115 and a bit 116 that does not require interpolation are prepared in advance as a line image buffer (number of pixels in the main scanning direction after resolution conversion). And the same bit buffer) 117.

  Next, the pixel value of the interpolation bit 115 is set to the pixel value of the pixel located on the left side (for example, if the pixel value of the pixel on the left side of the interpolation bit 115 is 1 (black pixel), the pixel of the interpolation bit 115 Set the value to 1) and perform pixel interpolation. The above processing is sequentially performed line by line to replace the entire original image with a converted image, thereby increasing the resolution.

  Thereafter, regarding the line width of the black line in the main scanning direction (corresponding to the thickness of the vertical line), the thin line width and the thick line width are obtained based on the conversion magnification of the image, and the line in the main scanning direction is obtained. Create a width correction table.

  The creation of this line width correction table is made by narrowing down the result of the calculation of x × k rounded down, assuming that the line width in the main scanning direction of the image before correction is x (x is an integer) and the conversion magnification is k. The line width is set to m0, and the result obtained by rounding up the decimal point of the x × k calculation result is set to a thick line width of m1. Therefore, for example, in the case of increasing the resolution from 454 dpi to 602 dpi (conversion magnification k = 1.32 times), when the line width of the image before correction is “2”, the thin line width m0 = 2 Thus, the thick line width m1 = 3. This is obtained sequentially for all the line widths x = 1 to x = n (n is the maximum number of pixels of the line width in the main scanning direction existing in the image before conversion) before correction.

  When the creation of the line width correction table in the main scanning direction is completed, the same processing is performed for the line width of the black line in the sub scanning direction (corresponding to the thickness of the horizontal line), and the sub scanning direction of the image before correction is performed. A thin line width size p0 and a thick line width size p1 with respect to the line width y (y is an integer) are obtained to create a line width correction table in the sub-scanning direction.

  Then, the image after resolution conversion using the line image buffer 117 is targeted, and the line width correction processing is performed using the above-described line width correction table. At this time, if the correction to the thin line width is specified, the line widths x and y of the image after the resolution conversion are converted to m0 and p0, and conversely, if the correction to the thick line width is specified. The line widths x and y are converted into m1 and p1. Thereby, the thickness of the vertical line and the horizontal line is made uniform, and the variation in the line width and the collapse of the outline are eliminated.

JP 2007-265055 A

Hiroshi Masashima and 4 others, "Performance evaluation and processing speed improvement methods for various enlargement / reduction methods of binary images", IPSJ Transactions Sep.1985 Vol.26 No.5

  The logical sum method and nearest neighbor method mentioned as the first conventional technique can maintain a smooth outline of a line drawing, suppress the difference between the density of the original image and the density of the image after pixel number conversion, and reduce the original image. A converted image having a density close to that of the image can be obtained.

  However, in this method, as shown in FIG. 26 (a), when attention is paid to two horizontal lines 91 and 92 (or 93 and 94) located at different locations on the original image, these are indicated on the original image. In spite of the same line width (thickness), as shown in FIG. 26B, the converted image may be output with different line width (thickness) (the horizontal line 91 ′ is a horizontal line). The horizontal line 93 'is thinner than the horizontal line 94', and the horizontal line 93 'is thicker than the horizontal line 94'. This is because the pixels constituting the horizontal lines 91 and 92 or the adjacent pixels are affected by the value of the surrounding pixels. If the pixel positions are different, the influence from the surroundings is different. As a result, the value may or may not be reversed depending on the location.

  Further, in the 9-division method, the projection method, the inverse distance method, and the like cited as the second conventional technique, the difference between the density of the original image and the density of the image after the pixel number conversion is suppressed to be smaller than that of the first conventional technique. However, on the other hand, there is a problem that noise (a pixel whose pixel value is unintentionally inverted) occurs in the outline portion of the line drawing. This is also caused by the influence of the peripheral pixel values as in the case of the first prior art, but in the case of the second prior art, the peripheral pixel values are compared with those of the first prior art. In the same line, a portion where the value is inverted and a portion where the value is not inverted appear in the same line, and the contour of the line drawing appears uneven. As a result, the contour characteristics (smoothing) are destroyed, and the shape of the characters is lost, which may lead to degradation of image quality.

  On the other hand, in the pixel number conversion method described in Patent Document 1 cited as the third conventional technique, the line width is unified after the resolution conversion, so the line width variation problem that occurs in the first conventional technique. In addition, it is possible to eliminate the noise problem of the contour portion that occurs in the second prior art.

  However, in this method, since the correlation between the main scanning direction and the sub-scanning direction is completely independent without considering the correlation between the horizontal line and the vertical line based on the shape of the line drawing, as shown in FIG. There is a problem that irregularities occur in the portion where the end of the line is in contact with the horizontal line, or the portion where the end of the horizontal line is in contact with the vertical line, so that the collapse of the shape of the character cannot be completely eliminated. Also, in this method, since the output is unified to either a narrow line width with a thin line width or a thick line width with a thick line width, the density of the original image and the density of the image after conversion of the number of pixels are set. There is also a problem that the difference tends to increase.

  Accordingly, the present invention has been made in view of the above-described problems in the prior art, and is capable of suppressing the disorder of the shape of the character due to the pixel number conversion and obtaining a high-quality converted image. An object is to provide a number conversion method and the like.

In order to achieve the above object , the present invention provides a pixel number conversion method for generating a binary converted image by converting the number of pixels of a binary original image, wherein the line of the binary original image is based on a conversion magnification. A first step of creating a line width conversion table for designating a conversion line width on the binary conversion image corresponding to the line width of the image; and a main scanning direction or a sub scanning direction for the binary original image And detecting the coordinates of the leading pixel of the line image extending to one side of the line image, calculating the conversion coordinates on the conversion image corresponding to the detected coordinates, and the second step of detecting the length of the line image And a third step of acquiring a conversion line width corresponding to the length of the detected line image using the line width conversion table, and using the calculated coordinates as a starting point and either the main scanning direction or the sub-scanning direction. And having the length of the acquired conversion line width A fourth step of creating an image on the converted image, and the converted image created in the fourth step is scanned in the other of the main scanning direction or the sub-scanning direction, and extends in the other of the main scanning direction or the sub-scanning direction. And a fifth step of detecting a line image in which black pixels and white pixels are mixed, and a line image in which black pixels and white pixels detected in the fifth step are mixed, and the number of black pixels and white pixels And a sixth step of converting the value of the pixel belonging to the smaller side to the value of the larger side, and the fifth step includes the scanned pixel of interest and a peripheral adjacent to the pixel of interest A pixel pattern formed by pixels, a predetermined end pattern indicating that the target pixel is a starting point of a line image in which the black pixels and white pixels are mixed, and / or a portion other than the starting point of the line image Indicates that Comparing the predetermined continuous portion pattern, that you detect a line image in which the black pixels and white pixels are mixed depending on whether said pixel pattern and said end pattern or the continuous part pattern matches Features.

  And according to this invention, while calculating the coordinate on the conversion image of the front end pixel of a line image, a conversion line width is acquired using a line width conversion table, Then, the acquired conversion line width and front end pixel are acquired. Therefore, the conversion line width corresponding to the line width on the original image can be uniquely determined at any position on the image. For this reason, it is possible to avoid variations in line width due to conversion, and it is possible to perform conversion processing while maintaining the character shape.

  In addition, after performing one conversion process in the main scanning direction or the sub-scanning direction, a line image in which black pixels and white pixels are mixed is detected for the other in the main scanning direction or the sub-scanning direction. In order to convert the value of the pixel belonging to 1 to the value on the larger side, the unevenness on the other side in the main scanning direction or the sub-scanning direction caused by one conversion process in the main scanning direction or the sub-scanning direction can be eliminated. It becomes possible to correct the disturbance.

In the pixel number conversion method, for the converted image created in the sixth step, the coordinates of the tip pixel of the second line image extending in the other of the main scanning direction or the sub-scanning direction are detected, A seventh step of detecting the length of the second line image; calculating a conversion coordinate on the conversion image corresponding to the coordinate of the tip pixel of the detected second line image; and the line width conversion table. And an eighth step of acquiring a conversion line width corresponding to the detected length of the second line image, and the coordinates calculated in the eighth step as a starting point, in the main scanning direction or the sub-scanning direction A ninth step of creating on the converted image a line image having the length of the converted line width acquired in the eighth step and the converted image generated in the ninth step in the main scanning direction or Scan in one direction in the sub-scanning direction A tenth step of detecting a line image in which black pixels and white pixels are mixed while extending in one of the scanning direction and the sub-scanning direction, and a line image in which black pixels and white pixels detected in the tenth step are mixed And an eleventh step of comparing the number of black pixels with the number of white pixels and converting the value of pixels belonging to the smaller side to the value of the larger side , the tenth step being the scanned A pixel pattern formed by a pixel of interest and peripheral pixels adjacent to the pixel of interest is compared with a predetermined end pattern and / or continuous pattern, and the pixel pattern and the end pattern or continuous pattern are line image in which the black pixels and white pixels are mixed depending on whether matching can you to detect.

  According to the above configuration, it is possible to correct the variation in line width while ensuring the correlation between the vertical line and the horizontal line. Thereby, it is possible to eliminate the unevenness problem that occurs at the intersection of the vertical line and the horizontal line, and to suppress the distortion of the character shape due to the conversion of the number of pixels.

Furthermore, the present invention is 2 NeHara a order of the program to execute the pixel number conversion method for generating a conversion to binary converted image the number of pixels of the image to the computer system, the number of pixels described above either A conversion method is executed by a computer device .

Furthermore, the present invention is a pixel number conversion device that generates a binary converted image by converting the number of pixels of a binary original image, and based on the conversion magnification, the line width of the line image of the binary original image is set. A line width conversion table creating unit for creating a line width conversion table for designating a conversion line width on the corresponding binary converted image, and a line image existing in the binary original image using the line width conversion table A conversion processing unit that changes the line width of the image and generates the binary conversion image; and a character shape correction unit that performs a character shape correction process on the image generated by the conversion processing unit, the conversion processing unit including the binary Detecting the coordinates of the leading pixel of the line image extending in one of the main scanning direction or the sub-scanning direction for the original image, and detecting the length of the line image, and detecting by the detecting unit Conversion destination to calculate the conversion coordinates on the conversion image corresponding to the coordinates A coordinate calculation unit, a line width acquisition unit that acquires a conversion line width corresponding to the length of the line image detected by the detection unit using the line width conversion table, and the coordinates calculated by the tip coordinate calculation unit And a line unit conversion unit that extends in one of the main scanning direction or the sub-scanning direction and that creates a line image having a length of the conversion line width acquired by the line width acquisition unit on the converted image, the character shape correction The line scans the converted image created by the conversion processing unit in the other of the main scanning direction or the sub-scanning direction, extends to the other of the main scanning direction or the sub-scanning direction, and is a line image in which black pixels and white pixels are mixed. For the line image detected by the concavo-convex line image detection unit and the line image detected by the concavo-convex line image detection unit, the number of black pixels and the number of white pixels are compared, and the value of the pixel belonging to the smaller side is set to the value of the larger side. and a correcting unit for converting the concave The line drawing detection unit indicates a pixel pattern formed by the scanned pixel of interest and peripheral pixels adjacent to the pixel of interest, and the pixel of interest is a starting point of a line image in which the black pixels and white pixels are mixed. A predetermined end pattern and / or a predetermined continuous pattern indicating a part other than the start point of the line image are compared, and the pixel pattern matches the end pattern or the continuous pattern. the black pixels and white pixels, characterized that you detect line images mixed depending on whether.

  As described above, according to the present invention, it is possible to suppress the disorder of the character shape caused by the pixel number conversion and obtain a high-quality converted image.

1 is a block diagram showing an embodiment of a pixel number conversion apparatus according to the present invention. (A) shows the number of pixels in the main scanning and sub-scanning directions of the original image, and (b) shows the number of pixels in the main scanning and sub-scanning directions of the converted image. It is a figure which shows the line | wire width conversion table of the main scanning direction. It is a figure which shows the pixel number conversion process of the main scanning direction. It is a figure which shows the pixel number conversion process of a subscanning direction. It is a figure which shows the collation pattern for detecting the upper end part of the uneven | corrugated vertical line at the time of the character correction of a subscanning direction. It is a figure which shows the collation pattern for detecting parts other than the upper end part of the uneven | corrugated vertical line at the time of the character correction of a subscanning direction. It is a figure which shows the character shape correction | amendment process of a subscanning direction. It is a figure which shows the collation pattern for detecting the uneven | corrugated horizontal line at the time of character shape correction | amendment of the main scanning direction. It is a figure which shows the shape correction process of the main scanning direction. 3 is a main flowchart showing an overall operation of the pixel number conversion apparatus of FIG. 1. It is a sub-flowchart which shows the pixel number conversion process of the main scanning direction. It is a figure which shows the character shape correction | amendment process of a subscanning direction. It is a sub flowchart which shows the character shape correction | amendment process of a subscanning direction. It is a figure for demonstrating the character shape correction process of a subscanning direction. It is a figure for demonstrating the collation process of step S23 of FIG. It is a figure for demonstrating the collation process of step S28, S30 of FIG. It is a figure for demonstrating the collation process of step S32 of FIG. It is a figure for demonstrating the character shape correction process of a subscanning direction. It is a figure for demonstrating the character shape correction process of a subscanning direction. It is a sub flowchart which shows the pixel number conversion process of a subscanning direction. It is a sub-flowchart which shows the shape correction process of the main scanning direction. It is a figure which shows an example of the image processed by the pixel number converter of FIG. It is a figure which shows an example of the state which eliminated the dispersion | variation in line | wire width. It is a figure which shows an example of the conventional pixel number conversion method. It is a figure which shows an example of the state from which two lines which have the same thickness on an original image become different thickness on a conversion image, and are output. It is a figure which shows an example of the unevenness | corrugation which arises in the intersection of a vertical line and a horizontal line.

  Next, modes for carrying out the invention will be described in detail with reference to the drawings. In the following description, a text image in which the character color is black and the background color is white will be described as an example.

  FIG. 1 is a block diagram showing an embodiment of a pixel number conversion apparatus (hereinafter referred to as “conversion apparatus”) according to the present invention. This conversion apparatus 1 is roughly divided into a memory 2 and a line width conversion table. A line width conversion table creation unit 3 for creating a pixel number, a pixel number conversion processing unit 4 for performing a pixel number conversion process on the input binary image (original image), and a character shape correction for the converted image after the pixel number conversion It is comprised from the character shape correction | amendment process part 5 which performs a process.

  The various processing units 3 to 5 are not necessarily configured by hardware, and a part or all of the processing units 3 to 5 may be configured by software (program).

  The memory 2 is a storage medium for storing image data to be processed by the conversion apparatus 1 and storing processing values and image data processed by the processing units 3 to 5. In the memory 2, as image data to be processed by the conversion apparatus 1, the data value of each pixel is a binary gradation value (for example, gradation value = 1 for a black pixel, gradation value = for a white pixel = Binary image data represented by 0) is input.

  The line width conversion table creation unit 3 uses the conversion magnification calculation unit 3a for obtaining the conversion magnification corresponding to the designated conversion resolution or the enlargement / reduction degree of the image, and the conversion magnification obtained by the conversion magnification calculation unit 3a, And a table creation unit 3b for creating the conversion table 3c.

  The conversion magnification calculation unit 3a includes the number of pixels in the main scanning direction and the sub-scanning direction of the input binary image (original image), and the main scanning direction and the sub-scanning of the output binary image (image after pixel number conversion). Based on the number of pixels in the direction, the respective conversion magnifications in the main scanning direction and the sub-scanning direction are obtained.

  The conversion magnification rx in the main scanning direction is, as shown in FIG. 2, when the number of pixels in the main scanning direction of the original image 30 is L and the number of pixels in the main scanning direction of the image 31 after the pixel number conversion is L ′. The conversion magnification ry in the sub-scanning direction is obtained by the following formula 1, and the number of pixels in the sub-scanning direction of the original image 30 is H, and the number of pixels in the sub-scanning direction of the image 31 after the pixel number conversion is H ′. In this case, it is obtained by the following formula 2.

  The table creation unit 3b creates a line width conversion table 3c for each of the main scanning direction and the sub-scanning direction using the conversion magnifications rx and ry obtained by the conversion magnification calculation unit 3a. Here, a method of creating the line width conversion table 3c in the main scanning direction will be described first.

  Assuming that the line width in the main scanning direction of the black line of the original image is wx, the pure conversion line width Wx when this is converted according to the conversion magnification rx in the main scanning direction is obtained by the following equation 3 and shown in FIG. It becomes street. FIG. 3 shows a case where the original resolution of 600 dpi is increased to a resolution of 909 dpi (conversion magnification rx = 1.515).

  However, since the pure conversion line width Wx is a numerical value including a decimal, normalization (integerization) is performed while dividing the mode into the following three modes.

"Mode 1: thick processing"
In this mode, the fractional part of the pure conversion line width Wx is uniformly rounded up to an integer, which is set to a thick line width Whx. As shown in FIG. 3, the thick line width Whx is, for example, 2 when the black line width wx = 1 in the main scanning direction of the original image, and from the pure conversion line width Wx (= 1.51). Is also a large value. For this reason, when the line width conversion is performed in mode 1, the ratio of black pixels in the entire image is increased, and the image subjected to the line width conversion in mode 1 is an image having a higher density than the original image.

"Mode 2: Thinning process"
In this mode, the fractional part of the pure conversion line width Wx is uniformly rounded down to an integer, which is used as a narrow line width Wtx. If the line width wx = 1 is also focused on, the thin line width Wtx is 1 and is smaller than the pure conversion line width Wx (= 1.51). Therefore, when line width conversion is performed in mode 2, the ratio of black pixels in the entire image is reduced, and the image subjected to line width conversion in mode 2 is an image having a lower density than the original image.

“Mode 3: Optimal line width processing”
In this mode, with respect to the numerical value after the decimal point of the pure conversion line width Wx, rounding up is performed with a predetermined threshold t as a boundary, and rounding down is distributed, thereby converting the pure conversion line width Wx into an integer and obtaining the intermediate line width Wmx. . If the decimal point of the pure conversion line width Wx is smaller than the threshold value t, the intermediate line width Wmx is an integer obtained by rounding down the numerical value after the decimal point of the pure conversion line width Wx. Is larger than the threshold value t, the value after the decimal point of the pure conversion line width Wx is rounded up to an integer.

  As shown in FIG. 3, for example, focusing on the range of the line width wx = 1 to 5 of the original image in the main scanning direction, the intermediate line width Wmx is when the line width wx = 1, 3, and 5 of the original image. , The same value as the thick line width Whx, and the same value as the thin line width Wtx when the line width wx = 2 and 4 of the original image (however, the “intermediate line width Wmx” in FIG. The case where t is 0.5 is shown).

  That is, the intermediate line width Wmx is a value obtained by mixing the thick line width Whx and the thin line width Wtx, and is a conversion line width having a characteristic of mutually canceling errors from the pure conversion line width Wx (for example, the original line width Wmx). An error between the pure conversion line width Wx and the intermediate line width Wmx when the intermediate line width Wmx = 2 (the same value as the thick line width Whx) with respect to the image line width wx = 1 (2-1.51 = +0.49) is a pure conversion line width Wx and an intermediate line width Wmx when the intermediate line width Wmx = 3 (the same value as the thin line width Wtx) with respect to the line width wx = 2 of the original image. (Cancelled with an error of 3−3.03 = −0.3).

  For this reason, the image subjected to line width conversion in mode 3 is an image having a density close to the density of the original image as compared to the image subjected to line width conversion in modes 1 and 2. As a result of repeated experiments by the present inventor, when the threshold value t = 0.5, the density difference between the image after pixel number conversion and the original image is minimized, and the density closest to the density of the original image It has been found that a converted image can be obtained.

  Then, the normalization processing consisting of the above three modes is performed from the line width wx = 1 of the original image (the minimum number of black line width pixels in the main scanning direction existing in the original image) to i (i is the original image). The process is performed up to the maximum number of pixels (600 in the case of FIG. 3) of the black line width in the main scanning direction existing therein to complete the line width conversion table 3c in the main scanning direction.

  Next, a method for creating a line width conversion table in the sub-scanning direction will be described. In creating the line width conversion table in the sub-scanning direction, the sub-scanning direction pure conversion line width Wy is obtained by the following equation 4 using the sub-scanning direction conversion magnification ry obtained by the above equation 2.

  Then, as in the creation of the line width conversion table 3c in the main scanning direction, the pure conversion line width Wy is normalized (integerized) while being divided into the following three modes.

"Mode 1: thick processing"
The pure conversion line width Wy is uniformly rounded up to an integer, and this is set as a thick line width Why.

"Mode 2: Thinning process"
The fractional part of the pure conversion line width Wy is uniformly rounded down to an integer, and this is used as the thin line width Wty.

“Mode 3: Optimal line width processing”
With respect to the numerical value after the decimal point of the pure conversion line width Wy, the value is rounded up with a predetermined threshold value t as a boundary, and rounded down. As a result, the pure conversion line width Wy is converted to an integer to obtain an intermediate line width Wmy. Even in the sub-scanning direction, the optimum value is the threshold t = 0.5.

  Normalization processing consisting of the above three modes is performed by changing the original image line width wy = 1 (minimum number of line width pixels in the sub-scanning direction existing in the original image) to j (j is the maximum existing in the original image). To the line width conversion table 3c in the sub-scanning direction is completed.

  In FIG. 1, for convenience of explanation, the line width conversion table 3 c is described in the line width conversion table creation unit 3, but the line width conversion table 3 c is not necessarily included in the line width conversion table creation unit 3. It is not necessary to hold, and you may make it hold | maintain using the free space of the memory 2. FIG.

  The pixel number conversion processing unit 4 in FIG. 1 includes a main scanning conversion unit 11 that performs pixel number conversion in the main scanning direction and a sub-scanning conversion unit 12 that performs pixel number conversion in the sub-scanning direction.

  The main scanning conversion unit 11 includes a horizontal line detection unit 11a that detects a horizontal line (a black line image extending in the main scanning direction) included in the original image, and a front end for obtaining the leading edge coordinate Ax of the horizontal line detected by the horizontal line detection unit 11a. The coordinate detection unit 11b, the black line length detection unit 11c for obtaining the length Dx of the horizontal line detected by the horizontal line detection unit 11a, and the tip coordinate Ax ′ on the converted image corresponding to the tip coordinate Ax obtained by the tip coordinate detection unit 11b A line width acquisition unit 11e that acquires a conversion line width Dx ′ corresponding to the length Dx using the conversion tip coordinate calculation unit 11d that calculates the line width conversion table 3c in the main scanning direction generated by the table generation unit 3b, A line unit conversion unit 11f that converts the number of pixels in units of lines using the tip coordinates Ax ′ obtained by the conversion tip coordinate calculation unit 11d and the conversion line width Dx ′ obtained by the line width acquisition unit 11e is provided.

  As shown in FIG. 4A, the horizontal line detection unit 11a scans the original image 40 in the main scanning direction and detects a horizontal line 41 existing in the original image 40. The tip coordinate detection unit 11b obtains the coordinates Ax of the pixel 42 (the pixel that is the tip of the horizontal line 41) located at the left end of the horizontal line 41 detected by the horizontal line detection unit 11a.

  The black line length detection unit 11c obtains the length Dx in the main scanning direction of the horizontal line 41 detected by the horizontal line detection unit 11a. The converted leading edge coordinate calculation unit 11d uses the following equation 5 to calculate the coordinates Ax ′ on the converted image (the image obtained by converting the number of pixels in the main scanning direction) 43 corresponding to the coordinates Ax on the original image 40 with respect to the leading edge pixel 42. Is obtained (see FIG. 4B).

  The line width acquisition unit 11e determines the length Dx of the horizontal line 41 detected by the horizontal line detection unit 11a as “the line width wx of the black line of the original image in the main scanning direction” in the line width conversion table 3c in the main scanning direction shown in FIG. To obtain the corresponding conversion line width Dx ′ (thick line width Whx, thin line width Wtx, or intermediate line width Wmx). Which of the thick line width Whx, the thin line width Wtx, and the intermediate line width Wmx is acquired as the conversion line width Dx ′ depends on the mode selected at that time.

  As shown in FIG. 4B, the line unit conversion unit 11f sets the length of the conversion line width Dx ′ acquired by the line width acquisition unit 11e from the coordinate Ax ′ obtained by the conversion tip coordinate calculation unit 11d. A horizontal line 44 is created on the converted image 43.

  In the main scanning conversion unit 11, the above processing by the horizontal line detection unit 11 a to the line unit conversion unit 11 f is performed on all the horizontal lines existing on the original image 40, and the pixel number conversion processing in the main scanning direction is performed on the original image 40. Apply to the whole. If there are a plurality of short horizontal lines 45 and 46 within one line in the main scanning direction of the original image 40 as in the areas A and B of FIG. Considering the horizontal line extending in the direction, the tip coordinate Ax and the length Dx of each horizontal line 45, 46 are obtained and conversion processing is performed.

  Returning to FIG. 1, the sub-scanning conversion unit 12 includes a vertical line (sub-line) included in an image that has undergone pixel number conversion in the main scanning direction (strictly, an image that has been subjected to character correction processing in the sub-scanning direction described later). (A black line image extending in the scanning direction) for detecting a vertical line detection unit 12a for detecting a vertical line tip coordinate Ay detected by the vertical line detection unit 12a, and a vertical line detection unit 12a for detection. A black line length detection unit 12c that determines the length Dy of the vertical line, a conversion tip coordinate calculation unit 12d that calculates the tip coordinate Ay ′ on the converted image corresponding to the tip coordinate Ay obtained by the tip coordinate detection unit 12b, and a table Using the line width conversion table 3c in the sub-scanning direction created by the creation unit 3b, the line width acquisition unit 12e that acquires the conversion line width Dy ′ corresponding to the length Dy, and the tip coordinates obtained by the conversion tip coordinate calculation unit 12d Ay 'and the variable acquired by the line width acquisition unit 12e Using the line width Dy ', and the line unit conversion unit 12f for performing pixel number conversion is provided by line basis.

  Also in the conversion process in the sub-scanning direction, as in the case of the conversion process in the main scanning direction, as shown in FIGS. 5A and 5B, “detection of the vertical line 47 included in the converted image 43 in the main scanning direction”. "Processing", "Detection processing of the coordinate Ay of the pixel 48 that becomes the tip (upper end) of the detected vertical line 47, the length Dy of the vertical line 47", "the tip coordinate Ay 'on the converted image 49 in the sub-scanning direction" "Calculation process", "Process for obtaining the conversion line width Dy 'in the sub-scanning direction by querying the length Dy with the line width conversion table 3c in the sub-scanning direction", and "Conversion line width Dy starting from the tip coordinate Ay'. Processing for creating a vertical line 50 having a length of 'on the converted image 49 "is performed. Note that the front end coordinate Ay ′ in the sub-scanning direction can be obtained by the following Equation 6.

  In FIG. 1, for convenience of explanation, the conversion processing unit for the main scanning direction (main scanning conversion unit 11) and the conversion processing unit for the sub scanning direction (sub scanning conversion unit 12) are completely separated. However, since the functions of both are common except for the difference in scanning direction, for example, the tip coordinate detection unit, black line length detection unit, conversion tip coordinate calculation unit, line width acquisition unit, and line unit conversion unit are unified. The main scanning conversion unit 11 and the sub-scanning conversion unit 12 can share the same.

  1 includes a sub-scanning correction unit 21 that performs character correction processing in the sub-scanning direction, and a main-scanning correction unit 22 that performs character correction processing in the main-scanning direction.

  The sub-scanning correction unit 21 targets the converted image 43 (see FIG. 4B) generated by the main scanning conversion unit 11 of the pixel number conversion processing unit 4, and includes uneven vertical lines (a mixture of black pixels and white pixels). Vertical line detection unit 21a that detects the vertical line), counter 21b that counts the number of white and black pixels included in the vertical line detected by the vertical line detection unit 21a, and the output of the counter 21b And a linear correction unit 21c that corrects the vertical line detected by the vertical line detection unit 21a.

  The concavo-convex vertical line detection unit 21a scans the converted image 43 that has been subjected to the pixel number conversion in the main scanning direction in units of one pixel in the sub-scanning direction, and shows the scanned pixel (target pixel) and peripheral pixels in FIG. In contrast to the “end pattern A” and “end pattern B” shown in FIG. 7 and the “continuous pattern A” and “continuous pattern B” shown in FIG. 7, the vertical lines in which black pixels and white pixels shown in FIG. 52 is detected.

  Here, the “edge pattern A” shown in FIG. 6A includes a white pixel of interest pixel 51a and a white pixel placed above and adjacent to the pixel of interest 51a. During main scanning conversion by the main scanning conversion unit 11 for the white pixel 51b adjacent to the side opposite to the advancing direction of the sub-scanning scan by the unit 21a and the black pixel located to the left of the target pixel 51a (the target pixel 51a). This is a matching pattern consisting of three pixels with a black pixel 51c adjacent on the opposite side of the scan advance direction), and the pixel of interest was a white pixel when detecting the upper end (starting point) of the vertical line. It is what is sometimes used.

  In addition, the “edge pattern B” illustrated in FIG. 6B includes a black pixel of interest 51d and a white pixel (a pixel of interest 51d that is adjacent to the pixel of interest 51d). Scan at the time of main scanning conversion by the main scanning conversion unit 11 with respect to the white pixel (white pixel adjacent to the opposite side of the advance direction of the sub-scanning scan) 51e and the white pixel (rightward pixel 51d) of the target pixel 51d. (White pixel adjacent in the advancing direction) 51f) and a pattern for collation consisting of three pixels, which are used when the pixel of interest is a black pixel in detecting the upper end of the vertical line.

  Furthermore, the “continuous part pattern A” shown in FIG. 7A is a pattern for matching consisting of two pixels, a white pixel of interest 51g and a black pixel 51h located to the left of the pixel of interest 51g. When detecting a portion other than the upper end of the vertical line, it is used when the target pixel is a white pixel.

  In addition, the “continuous portion pattern B” illustrated in FIG. 7B is a matching pattern including two pixels, a black pixel of interest 51i and a white pixel 51j located to the right of the pixel of interest 51i. When detecting a portion other than the upper end portion of the vertical line, it is used when the target pixel is a black pixel.

  A specific method for detecting a vertical line in which black pixels and white pixels are mixed will be described in detail later.

  Returning to FIG. 1, the counter 21 b counts the number of black pixels and the number of white pixels in the vertical line 52 based on the output of the uneven vertical line detection unit 21 a. Although not shown, the counter 21b is provided with a counter for black pixels and a counter for white pixels.

  Based on the output of the counter 21b, the linear correction unit 21c compares the number of black pixels and the number of white pixels in the vertical line 52, and converts the value of the pixel belonging to the smaller pixel type to the value of the larger pixel type. Output. For example, if the number of black pixels is larger as shown by the vertical line 52 in FIG. 8A, all white pixels included in the vertical line 52 are converted to black pixels as shown in FIG. 8B. Then, the entire vertical line 52 is unified into black pixels. Conversely, if the number of white pixels is larger, all black pixels included in the vertical line are converted into white pixels, and the entire vertical line is unified into white pixels. By these processes, the unevenness of the vertical line is eliminated, and the vertical outline of the character (linear in the sub-scanning direction) is corrected.

  Returning to FIG. 1, the main scanning correction unit 22 targets the converted image 49 (see FIG. 5B) generated by the sub-scanning conversion unit 12 of the pixel number conversion processing unit 4. An uneven horizontal line detection unit 22a that detects a horizontal line in which white pixels are mixed), a counter 22b that counts the number of white and black pixels included in the uneven horizontal line detected by the uneven horizontal line detection unit 22a, and an output of the counter 22b Is provided with a linear correction unit 22c for correcting the uneven horizontal line detected by the uneven horizontal line detection unit 22a.

  As shown in FIG. 9, the pattern for collation used in the uneven horizontal line detection unit 22a has four patterns of “end pattern C”, “end pattern D”, “continuous pattern C”, and “continuous pattern D”. Consists of patterns.

  The end pattern C in FIG. 9A is used when the pixel of interest is a white pixel in detecting the left end (starting point) of the horizontal line, and the end pattern in FIG. 9B. D is used when the pixel of interest is a black pixel in detecting the left end of the horizontal line. 9C is used when a pixel of interest is a white pixel in detecting a portion other than the left end of the horizontal line, and the continuous portion pattern C of FIG. The pattern D is used when the pixel of interest is a black pixel in detecting a portion other than the left end portion of the horizontal line.

  As in the case of the correction process in the sub-scanning direction, as shown in FIGS. 10A and 10B, in the main scanning direction character shape correction process, “the unevenness in which black pixels and white pixels in the converted image 49 are mixed. "Detection processing of horizontal line 49a", "Counting processing of the number of pixels of white and black pixels in uneven horizontal line 49a", "Process to unify the pixel values in uneven horizontal line 49a to the value of the pixel type on the larger side" I do.

  In FIG. 1, for the sake of convenience of explanation, the correction processing unit for the sub-scanning direction (sub-scanning correction unit 21) and the correction processing unit for the main-scanning direction (main scanning correction unit 22) are completely separated. However, since the functions of both are common except for the difference in the scanning direction, for example, the counter and the linear correction unit can be unified and shared by the sub-scanning correction unit 21 and the main scanning correction unit 22.

  Next, operation | movement of the converter 1 which has the said structure is demonstrated, referring FIGS.

  In the conversion of the number of pixels, as shown in FIG. 11, the line width conversion table creation unit 3 performs the number of pixels in the main scanning and sub-scanning directions of the input original image and the main image in the main scanning and sub-scanning directions of the converted image to be output. Based on the number of pixels, a line width conversion table 3c (see FIG. 3) in the main scanning direction and the sub-scanning direction is created (step S1). Next, the conversion mode is set, and it is selected whether to execute the conversion process in mode 1 (thickening process), mode 2 (thinning process), or mode 3 (optimum destination width process) (step S2).

  By the way, when printing a binary image, the binary image (digital image) is transferred to the printing plate through the plate making process, and then the ink is printed on the paper surface. In the printing process, dots due to ink bleeding or the like are printed. A gain phenomenon (a phenomenon in which the halftone dot on the printing paper becomes larger than the halftone dot of the printing plate) always occurs.

  The amount of dot gain varies not only with the amount of ink, etc., but also with the presence or absence of pixel number conversion, so in the case of pixel number conversion on the premise of printing, it is necessary to consider the effect on dot gain given by the conversion process . When other conditions (such as ink characteristics and paper quality) are ignored and only the characteristics of binary image data (images to be printed) are considered, the amount of dot gain generated during printing depends on the pixel size of the binary image. fluctuate.

  For example, in the case of resolution conversion of a text image (when only the number of pixels is changed while maintaining the size of the entire image), the size of one pixel changes before and after the conversion, which affects the dot gain amount in the printing process. Will give. For example, when the resolution is increased from 909 dpi to 1200 dpi, the size of one pixel is reduced from 28 μm × 28 μm to 21 μm × 21 μm, and the area of one pixel in the image after resolution conversion is the area of one pixel of the original image To 56%.

  Since the dot gain amount tends to increase as the size of one pixel of the binary image increases, printing is performed without converting the resolution when the size of one pixel is reduced by resolution conversion as in the above example. Compared with the case, the dot gain amount becomes small. For this reason, when the resolution of the binary image is increased and printed, the appearance of the printed image becomes light, and there is a problem that the impression of the image is changed.

  On the other hand, the pixel number conversion includes not only the above-described resolution conversion but also enlargement / reduction of the image (a process of increasing / decreasing the number of pixels while maintaining the size of one pixel). Therefore, there is no dot gain variation as in the case of resolution conversion. In other words, depending on the type of pixel number conversion processing, it may or may not affect the dot gain at the time of printing, so it is important to be able to acquire an appropriate print image in either case. Become.

  Therefore, in consideration of the influence of the conversion process on the dot gain amount at the time of printing, the mode selection in step S2 is preferably performed as follows.

  For example, when converting from a low-resolution image to a high-resolution image (when increasing the resolution without changing the overall image size), the size of one pixel of the converted image is larger than the size of one pixel of the original image. The dot gain amount that occurs when the converted image is printed becomes smaller than the dot gain amount when the original image is printed.

  For this reason, when performing conversion to increase the resolution, the conversion in mode 1 (thickening process) is selected, and the number of pixels to increase the line width is converted, and the density of the entire image is higher than that of the original image. It is preferable to generate a converted image. In this case, the density of the image to be printed on the paper surface can be increased by the amount of increase in the density of the converted image (the image to be printed), and the influence of the decrease in the dot gain amount resulting from the pixel number conversion. Can be reduced. That is, the decrease in the density of the print image due to the decrease in the dot gain amount can be offset by the increase in the density due to the thickening process at the time of pixel number conversion, and the print image can be prevented from appearing thin.

  Also, when converting from a high-resolution image to a low-resolution image (when reducing the resolution without changing the size of the entire image), the size of one pixel of the converted image becomes large, and dot gain generated when the converted image is printed The amount is larger than that of the original image. For this reason, it is preferable to select the conversion in mode 2 (thinning process) and implement the pixel number conversion to reduce the line width. In this case, the increase in the density of the print image due to the increase in the dot gain amount can be offset by the decrease in the density due to the thinning process when converting the number of pixels, and the print image can be prevented from appearing dark. .

  Furthermore, when the size of the entire image is enlarged or reduced without changing the resolution, the number of images of the entire image is only increased or decreased before and after conversion, and the size of one pixel itself does not change. There is no. For this reason, it is preferable to select the conversion in mode 3 (optimal line width processing), and to perform the pixel number conversion while suppressing the variation in line width while faithfully reflecting the density of the original image.

  It is convenient to be able to manually change the threshold value t when selecting the conversion by mode 3 (conversion in which rounding up / down of the decimal point is performed with the threshold value t as a boundary). That is, it is preferable to set the optimal value t = 0.5 in the apparatus as a default value so that the value of the threshold value t can be increased or decreased when the user desires. As a result, for example, it is possible to respond to a request to intentionally change the density at the time of pixel conversion, for example, to increase the density of the converted image.

  The mode selection may be performed manually by the user. However, as described above, since there is regularity between the conversion type and the appropriate mode, depending on the specified conversion type. The mode may be automatically selected on the device side.

  In this way, as the conversion line width Dx ′, the pure conversion line width Wx is rounded down by rounding down the rounded line width Wtx rounded down to the whole number by rounding down the decimal point, rounded down to the whole line by rounding up the decimal point to the whole number Wtx, and the threshold value t. Since the three line widths of the distributed intermediate line width Wmx are prepared and can be selected according to the conversion conditions, resolution conversion that causes fluctuations in the dot gain amount and enlargement of an image that does not cause fluctuations in the dot gain amount In any case where reduction is performed, a print image having an appropriate density can be obtained, and adverse effects on the print density due to pixel number conversion can be suppressed.

  When the mode selection is completed, the main scanning conversion unit 11 (see FIG. 1) of the pixel number conversion processing unit 4 performs a pixel number conversion process in the main scanning direction (step S3). In the process of converting the number of pixels in the main scanning direction, as shown in FIG. 12, first, the original image 40 (see FIG. 4A) is scanned (step S11), and the horizontal line 41 of the original image 40 (FIG. 4A). )) Is detected (step S12).

  Next, the coordinate Ax of the pixel 42 located at the left end of the horizontal line 41 is detected, and the length Dx of the horizontal line 41 is detected (steps S13 and S14). Next, the coordinate Ax ′ on the converted image 43 (see FIG. 4B) corresponding to the coordinate Ax on the original image 40 is calculated (step S15).

  Next, a conversion line width Dx ′ corresponding to the length Dx is obtained using the line width conversion table 3c (see FIG. 3) in the main scanning direction created in step S1 of FIG. 11 (step S16). At this time, which of “thick line width Whx”, “thin line width Wtx”, and “intermediate line width Wmx” in FIG. 3 is acquired as the conversion line width Dx ′ is the mode selected in step S2 in FIG. Follow.

  Next, the horizontal line 44 (see FIG. 4B) of the conversion line width Dx ′ acquired in step S16 is written in the converted image 43, starting from the coordinates Ax ′ calculated in step S15 of FIG. 12 (step S17). Thereafter, the processes in steps S12 to S17 are repeated until the remaining unscanned horizontal line does not exist in the original image 40 (step S18), and the entire image is processed.

  By the above processing, as shown in FIGS. 13A and 13B, a converted image 63 is obtained in a state in which the unevenness of the horizontal outline (linear in the main scanning direction) of the character is eliminated. However, in the portion where the length of the horizontal line is different on the original image 60 (for example, the horizontal line 61 of length Dx1 and the horizontal line 62 of length Dx2), the pure conversion line width and the conversion line width (Whx of FIG. 3) , Wtx, or Wmx) (rounding error) is different. This difference in rounding error appears as a difference in the number of pixels when writing a horizontal line as shown in FIG. 13B. Therefore, in the converted image 63, the outline of the character in the vertical direction (linear in the sub-scanning direction). May occur (mainly in the region 64 at the end (right end) of the written horizontal line).

  Therefore, as shown in FIG. 11, following the pixel number conversion processing in the main scanning direction, the sub-scanning correction unit 21 (see FIG. 1) of the character-shaped correction processing unit 5 performs character correction processing in the sub-scanning direction (step). S4). In the character correction processing in the sub-scanning direction, as shown in FIG. 14, first, the converted image 63 (see FIG. 13B) in the main scanning direction is scanned in units of one pixel in the sub-scanning direction (step S21). It is determined whether or not the scanned pixel (target pixel) is a white pixel (step S22).

  For example, when the target pixel is a white pixel as in the pixels 71a, 71b, 71c, 71d in FIG. 15A (step S22: Y), the target pixel, the upper adjacent pixel, and the left adjacent pixel are used. These three pixel values are compared with the end pattern A shown in FIG. 6A to determine whether or not they match (step S23).

  In the case of the white pixel 71a in FIG. 15A, as shown in FIG. 16A, the pixel 72a adjacent to the upper side of the target pixel 71a is a white pixel, and the pixel 72b adjacent to the left side of the target pixel 71a is a white pixel. Therefore, the pixel values of the pixels 71a to 72b do not match the end pattern A (step S23: N). In this case, no process is performed, and the next pixel is scanned in the order of scanning in the sub-scanning direction (step S38). Also, when scanning the white pixels 71b and 71c in FIG. 15A, the three pixels including the target pixel and the peripheral pixels do not match the edge pattern A, so that no processing is performed as in the above case, The next pixel is scanned in the order of scanning in the sub-scanning direction.

  On the other hand, in the case of the white pixel 71d in FIG. 15A, as shown in FIG. 16B, the pixel 73a adjacent to the upper side of the target pixel 71d is a white pixel, and the pixel 73b adjacent to the left side of the target pixel 71d is Since it is a black pixel, the pixel values of these pixels 71d to 73b match the end pattern A (step S23: Y). In this case, the counter 21b (see FIG. 1) of the sub-scanning correction unit 21 is reset and the number of white pixels is counted up (steps S24 and S25).

  Next, a pixel located next to the scanned pixel (next pixel in the sub-scanning direction) is scanned (step S26), and it is determined whether or not the pixel is a white pixel (step S27). As a result, when the pixel is a black pixel as in the pixel 71e in FIG. 15A (step S27: N), the values of two pixels including the target pixel and the pixel on the right side thereof are set as shown in FIG. (Step S28).

  In the case of the pixel 71e, as shown in FIG. 17A, since the pixel 74 located on the right side of the target pixel 71e is a white pixel, it matches the continuous pattern B (step S28: Y). Then, the number of black pixels is counted up (step S29). Thereafter, the next pixel (the pixel 71f in FIG. 15A is scanned (step S26), and the processing from step S27 onward is repeated. In the pixel 71f, the arrangement of the pixel of interest and the pixel on the right is the pixel. Since it is the same as 71e, the number of black pixels is counted up as in the case of the pixel 71e.

  When the processing further proceeds and the pixel 71g in FIG. 15A is scanned, a white pixel is detected as before (step S27: Y). In this case, the values of the two pixels including the target pixel and the pixel adjacent to the left are collated with the continuous pattern A shown in FIG. 7A (step S30). Then, when the value of the target pixel and the adjacent pixel on the left side match the continuous part pattern A as in the pixel 71g (see FIG. 17B), the number of white pixels is counted up (step S31).

  Thereafter, the processing proceeds in the same manner as described above. As a result, the pixels 71h and 71i in FIG. 15A are detected as target pixels for counting up the number of black pixels, and the pixel 71j is counted up for counting the number of white pixels. It is detected as a target pixel (steps S26 to S31).

  Then, when the scan advances to the pixel 71k next to the pixel 71j in FIG. 15A, in the case of the pixel 71k, as shown in FIG. 17C, the values of the target pixel and the pixel on the left are the continuous portion pattern A. (Step S26, step S27: Y, step S30: N). In this case, the pixel 71k is not targeted for counting up the number of white pixels, and the process proceeds to step S35.

  On the other hand, in the pixel type determination in step S22, for example, when the target pixel is a black pixel, such as the pixels 81a and 81b in FIG. 15B (step S22: N), it is shown in FIG. As described above, the values of the three pixels including the target pixel, the upper adjacent pixel, and the right adjacent pixel are collated with the end pattern B shown in FIG. 6B to determine whether or not the two match. Step S32).

  For example, in the case of the pixel 81a in FIG. 15B, as shown in FIG. 18A, the value of three pixels including the target pixel, the upper adjacent pixel, and the right adjacent pixel is the edge pattern B. Does not match (step S32: N). In this case, no process is performed, and the next pixel is scanned in the order of scanning in the sub-scanning direction (step S38).

  On the other hand, in the case of the pixel 81b in FIG. 15B, as shown in FIG. 18B, the value of three pixels including the target pixel, the upper adjacent pixel, and the right adjacent pixel is the edge pattern B. (Step S32: Y). In this case, the counter 21b (see FIG. 1) of the sub-scanning correction unit 21 is reset and the number of black pixels is counted up (steps S33 and S34).

  Next, the pixel located next to the scanned pixel (the next pixel in the sub-scanning direction) is scanned (step S26), and thereafter, in the same manner as when the target pixel is a white pixel in the discrimination process in step S22. Then, the number of black pixels and the number of white pixels are sequentially detected, and the number of black pixels and white pixels is counted (steps S26 to S31).

  As a result of the processing in steps S21 to S34 described above, as shown in FIG. 19A, it is included in the region 64 (see FIG. 13B) which is the end of the horizontal line written in the conversion processing in the main scanning direction. Two vertical lines 83 and 84 are detected as vertical lines for counting the number of black pixels and the number of white pixels, and the number of black pixels and white pixels is counted in each vertical line.

  Then, following the above “detection of vertical lines to be counted” and “counting processing of black pixels and white pixels”, as shown in FIG. 14, the numbers of black pixels and white pixels are determined based on the count value of the counter 21b. It is determined whether or not both are “1” or more (step S35).

  For example, as shown by the vertical line 83 in FIG. 19A, when only black pixels are included and the number of white pixels is “0” (less than “1”) (step S35: N), no processing is performed. The next pixel is scanned along the order of scanning in the sub-scanning direction (step S38).

  On the other hand, when black pixels and white pixels are mixed and the number of black pixels and the number of white pixels are both “1” or more as indicated by the vertical line 84 in FIG. 19A (step S35: Y). The number of black pixels and the number of white pixels are compared (step S36). Then, when there are more black pixels like the vertical line 84 (step S36: Y), the white pixels 71d, 71g, 71j in the vertical line 84 are converted into black pixels, and FIG. As shown, the entire vertical line 84 is unified with black pixels (step S37).

  After the above processing, it is determined whether or not there are any remaining unscanned pixels in the input converted image 63 in the main scanning direction (step S38), and if there are remaining pixels, step until no unscanned pixels exist. The processes of S21 to S38 are repeated.

  In the above description, as shown in FIGS. 15 and 19, an image in which the number of black pixels is larger with respect to the vertical line in which black pixels and white pixels are mixed is illustrated, but the vertical line 85 in FIG. As shown in FIG. 14, when the number of white pixels increases, the black pixels 86a to 86c in the vertical line 85 are converted into white pixels, and the entire vertical line 85 is unified into white pixels (step S39 in FIG. 14). .

  Theoretically, the number of black pixels and the number of white pixels may be the same, but in this case, the pixel values may be converted so as to be unified to either black pixels or white pixels.

  As described above, after converting the number of pixels in the main scanning direction, a vertical line in which black pixels and white pixels are mixed is detected, and the number of pixels of the black pixels and white pixels is compared and the pixel type on the larger side is detected. In order to unify, the unevenness on the sub-scanning direction side caused by the conversion of the number of pixels in the main scanning direction can be eliminated, and the disturbance of the contour on the sub-scanning direction side can be corrected.

  Returning to FIG. 11, when the character correction processing in the sub-scanning direction is completed, the sub-scanning conversion unit 12 (see FIG. 1) of the pixel number conversion processing unit 4 performs the pixel number conversion processing in the sub-scanning direction (step S5). In the pixel number conversion process in the sub-scanning direction, as shown in FIG. 21, first, the correction image 55 subjected to the character correction process in the sub-scanning direction is scanned (step S51), and the vertical line of the correction image 55 is detected. (Step S52).

  Next, the coordinate Ay of the pixel located at the upper end of the vertical line is detected, and the length Dy of the vertical line is detected (steps S53 and S54). Next, the coordinate Ay ′ on the converted image 49 (see FIG. 5B) corresponding to the coordinate Ay on the corrected image 55 is calculated (step S55), and the sub-scanning direction created in step S1 of FIG. Using the line width conversion table 3c (see FIG. 3), a conversion line width Dy ′ corresponding to the length Dy is acquired (step S56).

  Next, using the coordinate Ay 'calculated in step S39 as a starting point, the vertical line 50 (see FIG. 5B) of the conversion line width Dy' acquired in step S46 is written in the converted image 49 (step S57). Thereafter, the processes in steps S52 to S57 are repeated until the remaining unscanned horizontal line does not exist in the corrected image 55 (step S58), and the entire image is processed.

  By the above processing, a converted image in a state in which the unevenness of the vertical outline of the character (linear in the sub-scanning direction) is eliminated is acquired. However, this time, due to the difference in rounding error that occurs at the time of converting the number of pixels in the sub-scanning direction, there is a possibility that irregularities occur in the horizontal outline of the character (linear in the main scanning direction) in the converted image (main scanning direction). To the end (lower end) of the written vertical line. For this reason, as shown in FIG. 11, a character shape correction process in the main scanning direction is performed on an image that has been subjected to pixel number conversion in the sub-scanning direction (step S6).

  The main scanning direction character shape correction process is performed by the main scanning correction unit 22 of the character shape correction processing unit 5, and in this process, as shown in FIG. In step S61, it is determined whether the scanned pixel (target pixel) is a white pixel (step S62).

  As a result of the determination, if the target pixel is a white pixel (step S62: Y), the values of three pixels including the target pixel, the left adjacent pixel, and the upper adjacent pixel are shown in FIG. It is checked against the end pattern C to determine whether or not both match (step S63). If they match, the counter 22b (see FIG. 1) of the main scanning correction unit 22 is reset and the number of white pixels is counted up (steps S64 and S65).

  Next, a pixel (next pixel in the main scanning direction) located to the right of the scanned pixel is scanned (step S66), and it is determined whether or not the pixel is a white pixel (step S67). If the pixel is a white pixel (step S67: Y), the two pixels including the pixel (target pixel) and the upper adjacent pixel are collated with the continuous pattern C shown in FIG. 9C (step S68). If they match, the number of white pixels is counted up (step S69).

  On the other hand, when the right adjacent pixel is a black pixel in step 67 (step S67: N), the continuous pixel pattern D shown in FIG. (Step S70), and if both match, the number of black pixels is counted up (step S71).

  Thereafter, the processes in steps S66 to S71 are repeated until a pixel that does not match any of the continuous portion patterns C and D appears.

  On the other hand, if the pixel of interest is a black pixel in the pixel type determination in step S62 (step S62: N), the value of three pixels consisting of the pixel of interest, the pixel adjacent below and the pixel adjacent to the left is determined. 9 is collated with the end pattern D shown in FIG. 9B (step S72). If they match, the counter 22b is reset and the number of black pixels is counted (steps S73 and S74).

  Next, the pixel located immediately to the right of the scanned pixel is scanned (step S66), and thereafter the pixel for counting up the number of black pixels and the number of white pixels are counted in the same manner as when the target pixel is a white pixel. While sequentially detecting the pixels to be counted up, the number of black pixels and white pixels is counted (steps S66 to S71).

  After performing “detection of horizontal lines to be counted” and “counting process of black pixels and white pixels” as described above, it is determined whether or not the number of black pixels and the number of white pixels are both “1” or more. Then, a horizontal line in which black pixels and white pixels are mixed is detected (step S75). Then, the number of black pixels and the number of white pixels are compared for the detected horizontal line (step S76), and the value of the pixel belonging to the smaller pixel type is converted to the value of the larger pixel (steps S77 and S78). .

  After the above processing, it is determined whether or not there are remaining unscanned pixels in the input converted image in the sub-scanning direction (step S79). If there are remaining pixels, step S61 is performed until no unscanned pixels exist. The process of ~ S79 is repeated.

  In this way, after the number of pixels in the sub-scanning direction is converted, a horizontal line in which black pixels and white pixels are mixed is detected, and the number of black and white pixels is compared and unified to a larger pixel type. Therefore, the unevenness on the main scanning direction side caused by the conversion of the number of pixels in the sub-scanning direction can be eliminated, and the disturbance of the contour on the main scanning direction side can be corrected.

  Further, in the present embodiment, conversion processing in the main scanning direction is performed, and character shape correction processing in the sub scanning direction is performed.Subsequently, conversion processing in the sub scanning direction is performed on an image that has undergone these processes, Finally, since the shape correction processing in the main scanning direction is performed, the unevenness of the top, bottom, left, and right contours can be eliminated while correlating the horizontal and vertical lines. For this reason, as shown in FIG. 23, it is possible to prevent unevenness at the intersection of the horizontal line and the vertical line, and it is possible to avoid the disorder of the character shape.

  Furthermore, in the present embodiment, the conversion processing in the main scanning direction and the sub scanning direction shown in FIGS. 4 and 5 is performed on a line-by-line basis using the line width conversion table. It is possible to avoid the variation in the line width that occurs (see FIG. 26) and the contour noise that occurs in the second prior art.

  That is, in the first and second prior arts, each pixel is sequentially converted while interpolating, and in the process, it is affected by the values of surrounding pixels, resulting in line width variations and contour noise. In this embodiment, in the present embodiment, the conversion line width is acquired using the line width conversion table, and the conversion process is performed in units of lines using the conversion line width and the conversion coordinates of the tip pixel. The conversion line width corresponding to the line width on the original image is uniquely determined regardless of the position on the image without being affected by the values of surrounding pixels.

  For this reason, as shown in FIGS. 24A and 24B, regarding the two horizontal lines 101 and 102 (or vertical lines 103 and 104) located at different locations, the same line width (thickness) on the original image. If so, the line widths of both can always be output (see horizontal lines 101 ′ and 102 ′ (or vertical lines 103 ′ and 104 ′)). In this case, if the line image is linear on the original image, the converted line image is always linear, so the outline of the line image (however, if it is a horizontal line, it is a horizontal outline, and if it is a vertical line, it is vertical. There is no extra unevenness in the direction profile. Accordingly, it is possible to avoid variations in line width due to pixel number conversion and noise in the outline portion, and conversion processing can be performed while maintaining the character shape.

  The embodiment of the present invention has been described above, but the present invention is not limited to the above-described configuration, and various modifications can be made within the scope of the invention described in the claims.

  For example, in the above-described embodiment, the image in which characters are formed by the black pixel group on the background configured by the white pixel group is illustrated as the target image for pixel number conversion. The present invention can also be applied to an image in which characters are formed by a white pixel group.

  In the above embodiment, when generating the line width conversion table 3c, three types of line widths, a thin line width, a thick line width, and an intermediate line width, are provided as conversion line widths. This is not essential because it takes into account fluctuations in the amount of gain and does not directly control the distortion of the character shape. That is, any line width conversion table may be used as long as it specifies the black line width (converted line width) on the converted image corresponding to the black image line width of the original image. It may have a kind of line width of only the line width, or it may have two kinds of line widths, a thin line width and a thick line width, as in the pixel number conversion method described in Patent Document 1. There may be.

  Further, in the above embodiment, the processing proceeds in the order of conversion processing in the main scanning direction → character shape correction processing in the sub scanning direction → conversion processing in the sub scanning direction → character shape correction processing in the main scanning direction. The processing may proceed in the order of: conversion processing in the main scanning direction → character shape correction processing in the main scanning direction → conversion processing in the main scanning direction → character shape correction processing in the sub scanning direction. In this case, the conversion process in the main scanning direction is performed on an image subjected to the conversion process in the sub-scanning direction and the character shape correction process in the main scanning direction.

DESCRIPTION OF SYMBOLS 1 Pixel number conversion apparatus 2 Memory 3 Line width conversion table preparation part 3a Conversion magnification calculation part 3b Table preparation part 3c Line width conversion table 4 Pixel number conversion process part 5 Character form correction process part 11 Main scanning conversion part 11a Horizontal line detection part 11b Tip Coordinate detection unit 11c Black line length detection unit 11d Conversion tip coordinate calculation unit 11e Line width acquisition unit 11f Line unit conversion unit 12 Sub-scanning conversion unit 12a Vertical line detection unit 12b Tip coordinate detection unit 12c Black line length detection unit 12d Conversion tip coordinate Calculation unit 12e Line width acquisition unit 12f Line unit conversion unit 21 Sub-scanning correction unit 21a Uneven vertical line detection unit 21b Counter 21c Linear correction unit 22 Main scanning correction unit 22a Uneven horizontal line detection unit 22b Counter 22c Linear correction unit 30 Original image 31 Conversion Image 40 Original image 41 Horizontal line 42 Tip pixel 43 Converted image 44 in the main scanning direction Horizontal line 45, 46 Horizontal line 47 Vertical line 48 Tip Pixel 49 Converted image 49a in main scanning and sub-scanning direction Uneven horizontal line 50 Vertical line 51 (51a to 51j) Pixel 52 Vertical line 55 Character correction processing image 56 in sub-scanning direction Vertical line 57 Horizontal line 60 Original image 61, 62 Horizontal line 63 Conversion Image 64 Area 71 (71a-71k) Pixel 72 (72a, 72b) Pixel 73 (73a, 73b) Pixel 74 Pixel 81 (81a, 81b) Pixel 83, 84 Vertical line 85 Vertical line 86 (86a-86c) Pixel 101, 102 Horizontal lines 101 ′ and 102 ′ of the original image Horizontal lines 103 and 104 after the conversion Vertical lines 103 ′ and 104 ′ of the original image Vertical lines after the conversion

Claims (5)

A pixel number conversion method for generating a binary converted image by converting the number of pixels of a binary original image,
A first step of creating a line width conversion table for specifying a conversion line width on the binary conversion image corresponding to a line width of the line image of the binary original image based on a conversion magnification;
A second step of detecting the coordinates of the leading end pixel of the line image extending in one of the main scanning direction or the sub-scanning direction with respect to the binary original image, and detecting the length of the line image;
A third step of calculating conversion coordinates on the conversion image corresponding to the detected coordinates, and acquiring a conversion line width corresponding to the length of the detected line image using the line width conversion table;
A fourth step of creating, on the converted image, a line image having the calculated conversion line width and extending in one of the main scanning direction or the sub-scanning direction with the calculated coordinates as a starting point;
The converted image created in the fourth step is scanned in the other of the main scanning direction or the sub-scanning direction, and a line image extending in the other of the main scanning direction or the sub-scanning direction and mixed with black and white pixels is detected A fifth step to:
For the line image in which black pixels and white pixels detected in the fifth step are mixed, the number of black pixels is compared with the number of white pixels, and the value of the pixel belonging to the smaller side is converted to the value of the larger side. and a sixth step of,
In the fifth step, a pixel pattern formed by the scanned target pixel and peripheral pixels adjacent to the target pixel, and the target pixel is a starting point of a line image in which the black pixel and white pixel are mixed. The pixel pattern and the end pattern or the continuous part pattern are compared with a predetermined end pattern indicating the above and / or a predetermined continuous part pattern indicating a part other than the starting point of the line image. converting the number of pixels and wherein that you detect a line image in which the black pixels and white pixels are mixed depending on whether matching. Targeting the converted image created in the sixth step, the coordinates of the leading end pixel of the second line image extending in the other of the main scanning direction or the sub-scanning direction are detected, and the length of the second line image is detected. A seventh step of detecting the thickness;
The conversion coordinates on the conversion image corresponding to the coordinates of the leading end pixel of the detected second line image are calculated, and the length of the detected second line image is corresponded using the line width conversion table. An eighth step of obtaining a conversion line width;
A line image having the length of the converted line width obtained in the eighth step is created on the converted image, starting from the coordinates calculated in the eighth step and extending in the other of the main scanning direction or the sub-scanning direction. A ninth step;
The converted image created in the ninth step is scanned in either the main scanning direction or the sub-scanning direction, and a line image that extends in one of the main scanning direction or the sub-scanning direction and contains black pixels and white pixels is detected. A tenth step,
For the line image in which black pixels and white pixels detected in the tenth step are mixed, the number of black pixels is compared with the number of white pixels, and the value of the pixel belonging to the smaller side is converted to the value of the larger side. and a eleventh step of,
The tenth step compares a pixel pattern formed by the scanned target pixel and peripheral pixels adjacent to the target pixel with a predetermined end pattern and / or continuous pattern, and the pixel pattern and the pixel conversion method according to claim 1, wherein the black pixels and white pixels, characterized that you detect line images mixed depending on whether or not the end pattern or the continuous part pattern matching. In the fifth step or the tenth step,
When the scanning direction of the scan is a sub-scanning direction and the target pixel is a white pixel, the target pixel, a white pixel located above the target pixel, and a left side of the target pixel Any one of the first end pattern composed of black pixels and / or the pixel of interest and the first continuous portion pattern composed of the pixel of interest and the black pixel located on the left of the pixel of interest And compare
When the scanning direction of the scan is the sub-scanning direction and the target pixel is a black pixel, the target pixel, a white pixel positioned above the target pixel, and a right side of the target pixel Any one of the second end pattern composed of white pixels and / or the second pixel of the second continuous pattern composed of the pixel of interest and a white pixel located right next to the pixel of interest and the pixel pattern And compare
When the scanning direction of the scan is the main scanning direction and the target pixel is a white pixel, the target pixel, a white pixel located to the left of the target pixel, and an upper side of the target pixel Any one of a third end pattern composed of black pixels and / or a third continuous portion pattern composed of the pixel of interest and a black pixel located adjacent to the pixel of interest and the pixel pattern And compare
When the scanning direction of the scan is the main scanning direction and the target pixel is a black pixel, the target pixel, a white pixel located on the left side of the target pixel, and a lower side next to the target pixel Any one of a fourth end pattern composed of white pixels and / or a fourth continuous portion pattern composed of the target pixel and a white pixel located immediately below the target pixel and the pixel pattern The pixel number conversion method according to claim 1 or 2, characterized in that 2 NeHara pixel number conversion method for generating a binary converted image by converting the number of pixels of the image to be because of a program is executed by a computer device,
A program for causing a computer device to execute the pixel number conversion method according to claim 1, 2 or 3. A pixel number conversion device that converts the number of pixels of a binary original image to generate a binary converted image,
A line width conversion table creating unit for creating a line width conversion table for designating a conversion line width on the binary conversion image corresponding to the line width of the line image of the binary original image based on the conversion magnification;
A conversion processing unit that changes a line width of a line image existing in the binary original image using the line width conversion table, and generates the binary converted image;
A character shape correction unit that performs character shape correction processing on the image generated by the conversion processing unit,
The conversion processing unit
A detection unit for detecting the coordinates of the leading end pixel of the line image extending in one of the main scanning direction or the sub-scanning direction for the binary original image, and detecting the length of the line image;
A conversion tip coordinate calculation unit that calculates conversion coordinates on a conversion image corresponding to the coordinates detected by the detection unit;
A line width acquisition unit that acquires a conversion line width corresponding to the length of the line image detected by the detection unit using the line width conversion table;
A line image having the length of the conversion line width acquired by the line width acquisition unit and extending in one of the main scanning direction and the sub-scanning direction is created on the conversion image, starting from the coordinates calculated by the tip coordinate calculation unit. A line unit conversion unit,
The character shape correction unit is
The converted image created by the conversion processing unit is scanned in the other of the main scanning direction or the sub-scanning direction, and a line image in which black pixels and white pixels are mixed is detected while extending in the other of the main scanning direction or the sub-scanning direction. An uneven line image detection unit;
The line image detected by the concavo-convex line drawing detection unit, and a correction unit that compares the number of black pixels and the number of white pixels and converts the value of the pixels belonging to the smaller side to the value of the larger side ,
The concavo-convex line drawing detection unit is a pixel pattern formed by the scanned target pixel and peripheral pixels adjacent to the target pixel, and the target pixel is a starting point of a line image in which the black pixel and white pixel are mixed. The pixel pattern and the end pattern or the continuous part pattern are compared with a predetermined end pattern indicating the above and / or a predetermined continuous part pattern indicating a part other than the starting point of the line image. matching whether the pixel number conversion apparatus characterized that you detect a line image in which the black pixels and white pixels are mixed in accordance with the.

Images