JPH0244432B2 - - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION (Field of Industrial Application) The present invention relates to an image reduction conversion method with little deterioration in image quality in pixels expressed in black and white binary.

(Conventional technology) For conventional reduction conversion of binary images, OPC
(Ordinal Processing Conversion) method, SPC
(Selective Processing Conversion) method, (both published in Proceedings of the 7th National Conference of the Institute of Image Electronics Engineers, Proceedings No.
10, May 19, 1975) and area discrimination method (IEICE Technical Report
PRL81-92, February 19, 1982), etc. have been announced.The OPC method is a method that sequentially extracts pixels of the original image (original pixels) and replaces them with pixels of the converted image (converted pixels), and the SPC method extracts the pixels of the original image (original pixels) and replaces them with the pixels of the converted image (converted pixels). This is a method of replacing the pixel with the closest original pixel on the original image.

In addition, in the area discrimination method, the nearest 4
In this method, the average density projected by each original pixel onto the converted pixel position is calculated, and threshold processing is performed to obtain the value of the converted pixel.

(Problems to be Solved by the Invention) In these conversion methods, a maximum of four original pixels are selected in order from the one closest to the converted pixel position, and the value of the converted pixel is calculated by a logical operation that calculates their value or density. It has been decided.

Therefore, if there are black or white lines on the original image and those lines are not the closest pixels to the converted pixel position, or the result of threshold processing of the average density projected by those lines to the converted pixel position is If the value of the line did not match, the corresponding line would sometimes disappear in the converted image.

For this reason, these conversion methods have the disadvantage that the black lines disappear, causing cuts and blurring, and the white lines disappear, causing blurring, resulting in deterioration of image quality.

In addition, in order to prevent image quality deterioration such as cuts and blurring caused by the disappearance of the black line, it is possible to consider a method in which the value of the converted pixel is set to black even when the black line is not at the closest position to the converted image. .

However, with this method, for example, when the line width is 1 pixel, two converted pixels located on both sides of the line may both become black, and as a result, the line width on the converted image may change. This has the disadvantage that the line width becomes thicker than the line width on the original image, and the image quality deteriorates due to collapse.

Structure of the Invention (Means for Solving the Problems) In order to solve these drawbacks, the present invention identifies a thin line that is likely to disappear among the thin lines on the original image, and causes the thin line to disappear. This method prevents deterioration of image quality by saving the converted image on the converted image. Conventional methods have proposed methods that give priority to black pixels and preserve them, but there is no method that determines black and white thin lines and preserves both white and black thin lines. By preserving both the white thin line and the black thin line, cutting, blurring, and blurring can be prevented, and a converted image of good quality can be obtained.

(Function) In this way, it is possible to identify lines that are likely to disappear during reduction conversion processing among the lines that exist on the original image, and to save the lines on the converted image without disappearing. Therefore, a converted image with less image quality deterioration can be obtained.

(Example) FIG. 1 is a diagram for explaining the principle of the present invention and reference pixels, and shows the arrangement of original pixels and reduced conversion pixels, and also shows the horizontal conversion ratio B/A= An example of 3/5 is shown. Here, P(m, n), P(m+1, n), P(m+2,
n), P (m, n+1), P (m+1, n+1), P
(m+2, n+1) ○, ◎ represent the original pixel Q(k,
l) X in Q(k+1,l) represents a converted pixel. Further, the original pixels having the shortest distance are shown connected by a solid line, and the center line between the original pixels is shown by a dashed line. In FIG. 1, the distance between the original pixels is normalized to 1, and the distance between the original pixel and the dashed dotted line is displayed as 1/2 as shown in FIG.
At this time, the distance between the converted pixels is expressed as A/B (5/3 in this case). Moreover, ~ is a number representing an area. The regions of ~ are delimited by dashed lines. In addition, in FIG. 1, the diagonal line indicates P(m+
1, n) and P(m+1, n+1) are black, the area in which a black value is assigned to the converted pixel is shown.

In order to prevent thin lines from disappearing, it is necessary to identify thin lines that are at risk of disappearing.

As shown in Figure 1, original pixel P(m+1,n)
and P(m+1,n+1) are black and have the same color, and P
(m, n), P (m+2, n), P (m, n+1), P
Assuming that (m+2, n+1) is white, which is the opposite color of P(m+1, n), in the conventional conversion method, the original pixel closest to the converted pixel position (in the case of Figure 1, Q(k,
Since the value of P(m,n) for l) and P(m+2,n) for Q(k+1,l) is the most weighted to determine the value of the converted pixel, Q
(k,l), Q(k+1,l) and P(m+1,n)
is substituted with the opposite color, white, and the black lines of P(m+1, n) and P(m+1, n+1) disappear.

In the present invention, the original pixel closest to the converted pixel is the starting point, and the horizontal direction approaches the converted pixel position.
The individual original pixels are added to the reference image to determine thin lines that may disappear. In the case of Figure 1, Q(k,l)
On the other hand, the above reference pixel is P(m+2,n),
For Q(k+1,l), it is P(m,n).

A specific example of the thin line discrimination method in which the above reference pixels are added is shown below. Q(k,
l), use the above P as a reference pixel.
In addition to (m+2, n), P(m, n), P(m+1,
Using P(m+1,n+1), P(m,n)=P(m+2,2) ≠P(m+1,n)=P(m+2,n+1)
...When (1) is satisfied, it is determined that there is a thin line that may disappear, and the value of P(m+1,n) is substituted for Q(k,l). If equation (1) is not satisfied, a value determined by a conventional conversion method (for example, the value of the shortest distance original pixel in the SPC method) is substituted for Q(k,l). In addition, in equation (1), P(m, n+1) and P
The value of (m+2, n+1) is not referenced. This is the vertical thin line P(m+1,n)-P(m+1,n
+1) and the horizontal thin line P(m, n+1)-P(m+1,
n+1)-P(m+2,n+1) also exists, the vertical thin line P(m+1,n)-P(m+1,
As in the case where only n+1) exists, Q(k,l)
This is to assign the value of P(m+1,n) to .
In other words, even if the horizontal and vertical thin lines are perpendicular, P(m, n+1) and P(m+2, n+1) are used to prevent the thin lines from being cut.
Determine whether a thin line exists without referring to the value of . Through the above processing, it is possible to prevent thin lines from disappearing. In the example of FIG. 1, the horizontal and vertical position (x,
y) is in region 1, that is, m≦x<m+
The case where 1/2 and n≦y<n+1/2 is satisfied is explained.

Similarly, if the position (x, y) is a region, m+1/2≦x<m+1 and n≦y<n
+1/2, P(m+1,n)=P(m
-1, n) ≠ P (m, n) = P (m, n + 1), and if this conditional expression is satisfied, a thin line exists, and the converted pixel of interest (Q) = P (m, n). , and when the position (x, y) is a region, that is, m≦x<m+1/2 and n+1/2≦y<
When n+1, P(m, n+1)=P(m
+2, n+1)≠P(m+1,n+1)=P(m+
1, n), and if this condition is satisfied, a thin line exists, and the converted pixel (Q)P(m
+1, n+1), and if the position (x, y) is a region, that is, m+1/2≦x<m+
1 and n+1/2≦y<n+1,
P(m+1,n+1)=P(m-1,n+1)≠P
Using (m, n+1) = P (m, n), if this conditional expression is satisfied, a thin line exists, and by setting the converted pixel of interest (Q) = P (m, n + 1), each disappears. Identify possible thin lines,
It is possible to prevent thin lines from disappearing.

FIG. 3 shows a flowchart of the above-described thin line detection and thin line storage operations.

Above, we have established a conversion method that preserves vertical lines during horizontal reduction conversion, but by using the same method for vertical reduction conversion,
It is possible to preserve horizontal lines.

FIG. 2 is a block diagram of an embodiment showing the configuration of the present invention, in which 1 is a sub-scanning counter of a converted pixel, 2 is a sub-scanning counter of a converted pixel;
is a main scanning counter of the converted pixel; 3 and 4 are memories, which give position information of the converted pixel in the sub-scanning direction and the main scanning direction, respectively; 5 is an area determining unit, which has sub- and main-scanning counters 1 and 4 indicating the converted pixel position; 2 outputs the type of area where the converted pixel is located, and 6 and 7 are an original pixel sub-scanning counter and an original pixel main scanning counter, respectively, based on the outputs of memories 3 and 4 indicating the original pixel position information. count the original pixel positions in the sub-scanning direction and the main scanning direction, and
is the line memory of original pixels to refer to, #1 to #k
10 constitutes a line memory group, and 10 is a conversion pixel determination logic unit that executes a logical operation to determine a converted pixel from an area and a plurality of original pixels, or stores the result of a logical operation. The value of the converted pixel is output by reading the storage device.

In the reduction conversion process, values of converted pixels are sequentially determined according to a fixed rule such as raster scanning, and converted pixels are created by repeating this process.

First, the positions of the converted pixel of interest in the sub-scanning direction and the main scanning direction are counted by the sub-scanning counter 1 and the main-scanning counter 2 of the converted pixel, respectively.

Next, the position of the original pixel corresponding to the converted pixel, for example, the original pixel closest to the converted pixel position, is determined.

The memories 3 and 4 store positional information of the original pixel corresponding to the converted pixel of interest, for example, the difference value between the position of the original pixel corresponding to the previous converted pixel and the converted pixel of interest, and store the original pixel sub-scanning counter 6 and Together with the original pixel main scanning counter 7, the original pixel positions corresponding to the converted pixels are counted.

The line memories 8 and 9 output a plurality of original pixels including the specified original pixel as reference pixels. On the other hand, the area determination unit 5 determines an area indicating a line to be saved and its storage condition from the converted pixel position.

Finally, the conversion pixel determination logic unit 10 checks whether the conditional expression for determining a line is satisfied based on the area and the reference pixel, and if it is satisfied, it is assumed that a line exists and the value of the line is converted to Assign to .

If not satisfied, the value of the converted pixel is determined based on a predefined logical operation method. An example of a simple logical operation method is the above-mentioned SPC method.

Next, a specific processing example will be described below.

(a) Relationship between sub-scanning counter 1 and main-scanning counter 2 of converted pixels, memories 3 and 4, and sub-scanning counter 6 and main-scanning counter 7 of original pixels Original image is divided into 2000 pixels in the main scanning direction x 2000 pixels in the sub-scanning direction
As an example, assume that the original image is converted with a conversion ratio α = 1/4 in both the main scanning and sub-scanning directions, and a converted image with a size of 500 pixels in the main scanning direction x 500 pixels in the sub-scanning direction is created. Let me explain.

A converted pixel sub-scanning counter 1 and a converted pixel main-scanning counter 2 count the positions of converted pixels in the sub- and main-scanning directions. If the positions of the converted pixel in the sub- and main-scanning directions are expressed as j' and i', respectively, the outputs of the converted pixel sub-scanning counter 1 and the converted pixel main-scanning counter 2 are j' and i'. That is, (i′, j′)
=(1,1),(2,1),(3,1),...,(500,
500) is output.

Memories 3 and 4 provide corresponding original pixels when the converted pixels are mapped onto the original image. The information stored in the memories 3 and 4 is the difference value between the positions of the original pixels corresponding to the previous conversion pixel and the conversion pixel of interest. This difference value is 4 when the conversion ratio α=1/4. A difference value 4 is output from the memory 3 each time j' increases, and a difference value 4 is output from the memory 4 each time i' increases.

The position of the original pixel in the sub- and main-scanning directions is counted by the original pixel sub-scanning counter 6 and the original pixel main-scanning counter 7. If the positions of the original pixel in the sub- and main-scanning directions are expressed as j and i, respectively, the outputs of the original pixel sub-scanning counter 6 and the original pixel main-scanning counter 7 are j.
and i. The position of the original pixel in the sub- and main-scanning directions when converting at the conversion ratio α=1/4 is (i,
j)=(4,1),(8,1),(12,1),...,
(2000, 2000). Based on the difference value 4 output from the memories 3 and 4, the original pixel sub-scanning counter 6 and the original pixel main-scanning counter 7 calculate (i,
j)=(4,1),(8,1),(12,1),...,
It can be counted as (2000, 2000). Based on the position information of this original pixel, the original pixel required for thin line discrimination is designated and the corresponding original pixel is read out.

(b) Operation of area determination unit 5 The area determination unit 5 outputs the type of area in which the converted pixel is located. This area indicates the type of thin line to be saved and the conditions for whether or not to save it. In order to output the area type, the type of the area where the converted pixel is located is calculated in advance and stored in correspondence with the converted pixel i', j'. As an input to the fixed section 5, an area indicating the thin line to be saved and the preservation conditions can be output.

Effects of the Invention As explained above, according to the present invention, lines existing on an original image that are likely to disappear during reduction conversion processing are identified, and the lines are transferred onto the converted image without disappearing. Because it is possible to save
A converted image with less deterioration in image quality can be obtained.

Further, it is easy to implement in hardware, and image reduction conversion can be performed at high speed.

Furthermore, since the conditional expression for line identification and the logical operation method used when the conditional expression is not satisfied are independent, by adopting an excellent logical operation method, converted images with even less image quality deterioration can be obtained. There are advantages that can be obtained.

According to the present invention, reduction conversion can be performed not only at the same magnification in the main scanning direction and in the sub-scanning direction, but also at different magnifications.

The present invention can be applied to various image processing systems, including document editing systems that reduce the size of images input by facsimile, and communication between different types of image terminals having different resolutions.

[Brief explanation of the drawing]

FIG. 1 is a layout diagram of original pixels and converted pixels for explaining the principle of the present invention and reference pixels, FIG. 2 is a block diagram of an embodiment showing the configuration of the present invention, and FIG. 3 is a diagram showing the detection of thin lines and This is a flowchart for preserving thin lines. 1... Converted pixel sub-scanning counter, 2... Converted pixel main-scanning counter, 3, 4... Memory, 5... Area determination section, 6... Original pixel sub-scanning counter, 7... Original pixel main-scanning counter, 8, 9... Line Memory, 10...Conversion pixel determination logic unit.

Claims (1)

[Claims] 1. An original image expressed by a pixel P (m, n) having two values of black and white is reduced and converted in the horizontal direction at an arbitrary magnification smaller than 1, and Q (k, l) is obtained. In the reduction transformation method that creates a transformed image expressed as , when Q(k,l) is mapped onto P(m,n),
Letting the distance between pixels of the original image be 1, Q(k,l)
The first step is to find the horizontal and vertical position (x, y) of
and the position (x,
y) is m≦x<m+1/2 and n≦y<n+1/2, m+1/2≦x<m+1 and n
a second region where ≦y<n+1/2, m≦x<m
+1/2 and n+1/2≦y<n+1, the third
area, m+1/2≦x<m+1 and n+1/2
a second means for determining which region of the fourth region where y<n+1 is located, and when the region is located in the first region, P from the original image
(m, n), P (m+2, n), P (m+1, n), P
Select (m+1, n+1) and determine whether the conditional expression P(m, n)=P(m+2,n) ≠P(m+1,n)=P(m+1,n+1) is satisfied. Then, assuming that a thin line exists, Q(k,l)=P(m+1,
a third means for outputting P from the original image when located in the second area;
(m+1,n),P(m-1,n),P(m,n),P
(m, n+1) and determine whether the conditional expression P(m+1, n)=P(m-1, n) ≠P(m, n)=P(m, n+1) is satisfied, If satisfied, assuming that a thin line exists, Q(k,l)=P(m,n)
a fourth means for outputting P from the original image when located in the third area;
(m, n+1), P(m+2, n+1), P(m+1,
n+1), P(m+1, n) and determine whether the conditional expression P(m, n+1) = P(m+2, n+1) ≠ P(m+1, n+1) = P(m+1, n) is satisfied. If satisfied, assuming that a thin line exists, Q(k,l)=P(m+1,
a fifth means for outputting P from the original image when located in the fourth area;
(m+1, n+1), P(m-1, n+1), P(m,
n+1), P(m, n) and whether the conditional expression P(m+1, n+1)=P(m-1, n+1) ≠P(m, n+1)=P(m, n) is satisfied. If it is satisfied, a thin line exists, and Q(k,l)=P(m,n+
1), and if all the above conditional expressions are not satisfied, P(m,n), P(m+1,n), P(m,n+
1), seventh means for determining the value of Q(k,l) from P(m+1,n+1), and the third, fourth, fifth, sixth and seventh means determine the value of Q(k,l). , l) to create a converted image by repeating the process of determining . 2. The line determination and storage method is also applied on the vertical axis to prevent thin lines from disappearing in the horizontal and vertical directions that occur when scaling down in the horizontal and vertical directions. image reduction method.