JPH01238275A - Device and method for contracting image - Google Patents

Abstract

(57) [Summary] This bulletin contains application data before electronic filing, so abstract data is not recorded.

Description

[Detailed description of the invention] [Table of Contents] Overview Background of the industrial application field technology (Figure 8.9) Prior art (Figure 20) Problem to be solved by the invention (Figure 10) Solving the problem Means for (Figs. 1 and 2) Effects (Figs. 1 and 2) First embodiment (Figs. 3 to 10) Second embodiment (Figs. 11 to 19) Effect [Summary] Set a divided area determined by a projection method according to the desired reduction ratio on a plane where pixels to be converted are arranged, and apply logic corresponding to the divided area where the converted pixels projected on the plane are located. Regarding an image reduction device and method for calculating the density of a converted pixel from the density of neighboring pixels to be converted according to the formula, the purpose of this article is to provide an image reduction device and method that performs conversion pixels that do not eliminate thin lines. a converted pixel input means for sequentially inputting density data of a predetermined group of located pixels to be converted; and a means for detecting thinning target pixels on a thin line to be thinned out by reduction from among the group of converted pixels based on the reduction rate. a thinned-out data generating means for generating data necessary for the conversion, a divided area determining means for determining in which divided area the converted pixel is located based on a desired reduction ratio, and a divided area determining means for determining in which divided area the converted pixel is located based on the thinned-out data and the converted pixel. The conversion pixel density calculation means calculates the density of the converted pixel based on the density data of the area and the group of pixels to be converted.

[Industrial application field]

Is this invention G4 to G? It relates to an image conversion device and method for pixel density conversion from high density to low density such as facsimile or image editing, and in particular, division determined by a projection method according to a desired reduction ratio on a plane on which pixels to be converted are arranged. The present invention relates to an image reduction apparatus and method for setting a region and calculating the density of a converted pixel from the density of neighboring pixels to be converted according to a logical formula corresponding to a divided area in which a converted pixel projected onto the plane is located.

[Technology background]

Conventionally, when reducing an image, SPC (Select
veProcessing Conversion) method, logical sum method, nine-part method, high-speed projection method (Jikai 358-97
958 Image Electronics Academic Journal Vol. 11, No. 2, 72-83 (
1982)) and the like have been proposed, and both methods determine the density of a converted pixel from four nearby pixels to be converted.

In the SPC method, the image tends to become thinner when reduced, and missing white and black pixels are noticeable.

In addition, the logical sum method and the nine-part method have the disadvantage that the image becomes thicker and the black image becomes more blurred.

On the other hand, the high-speed projection method is known as a method that reduces omissions and collapse of black and white pixels.

The high-speed projection method realizes high-speed processing by replacing arithmetic operations with logical operations when determining the density of converted pixels. In the high-speed projection method, as shown in Fig. 8, divided regions G to G8 are set in the plane of the converted image according to the conversion magnification, and the density of the converted pixel is determined from the converted pixel within the divided region. A logical formula is prepared in advance, and the density of the converted pixel is determined based on the logical formula assigned to the divided area where the converted pixel is located.

Here, Table 1 shows the logical formula.

Table 1. High-speed projection method (conventional method) Logical formula FIG. 8 shows the principle of the projection method. As shown in the figure, first, four pixels A, B, and C near a predetermined converted pixel.

Divide the converted image plane including D into four equal parts. The pixel densities of the four neighboring pixels are set as IA, IB, IC, 1, and when the converted pixel R is projected onto the converted pixel surface, the area ratios projected to each of the four equally divided areas are WA and WB.

Let W,,WD be.

In the projection method, on the conversion side, the density of the element R is IR=Σ1. *W.

(i = A, B, C, D), so ■
8≧0.5→IR=1 (black), IR<0.5→■□
= 0 (white). Note that p, , and q in FIG. 8 are conversion/reduction magnifications in the X-axis and y-axis directions.

Furthermore, since this projection method requires arithmetic operations to obtain the density of the converted pixel R, high-speed processing is not possible. Therefore,
The high-speed projection method replaces arithmetic operations with logical operations to increase speed. The curve in Figure 9 is one of four pixels.
It shows the boundary where the density of the converted pixel R can be determined only by pixels. For example, the curve separating G3 and 07 is given by the following equation.

(px+0.5)(qy+0.5)=0.5 Therefore, as shown in Table 1, only one pixel exists in areas 01 to G4;
5 to G8 are logical expressions that take into consideration the combination with the other three pixels.

[Conventional technology]

Conventionally, when reducing an image using the above projection method, as shown in FIG. 20, converted pixel input is performed in which each density data of a group of converted pixels located in the vicinity of a converted pixel is sequentially input by projection. means 201, a divided area determining means 203 that determines which divided area the converted pixel belongs to based on a desired reduction rate, and density data of the converted pixel group inputted by the converted pixel manual single stage 201, and each The image is reduced using an image reduction device having a converted pixel density calculation means 205 that calculates the density of the converted pixel based on the divided area to which the converted pixel belongs.

[Problem to be solved by the invention]

By the way, in conventional image reduction devices, during reduction, the line width 1/I)(+) perpendicular to the X-axis direction is
This has the problem that vertical lines whose width is less than 1/q (where q is the reduction ratio in the y-axis direction) and horizontal lines whose width is less than 1/q (where q is the reduction ratio in the y-axis direction) are eliminated.

That is, FIG. 10 shows the positional relationship between the pixel to be converted and the converted pixel in the case of reduction to 273 on both the X-axis and the y-axis. For the pixel to be converted, 5oot~S2□, the converted pixel ROOe
Flos, R□. , R, are located in each region G
5. G a , G s , G 7. Therefore, the value of the converted pixel R□1 is determined by the logical formula in Table 1, that is, AI:(B+
C+D)+B*C*D becomes "0". The following converted pixels R1□, R2□.

Similarly, R2□ becomes “0'', so the converted pixel S1□→S22→S3□ or S21→S2□→S23
There is a problem in that when a black and white pattern with a line width of 1 appears in the converted pixel, the above-mentioned thin line pattern is not expressed in the converted pixel and becomes a thinned-out pattern. Therefore, the technical object of the present invention is to solve the above-mentioned problems, and detects converted pixels that are to be thinned out (thin lines may disappear) according to the conversion reduction rate. The object of the present invention is to provide an image reduction device and method that performs conversion pixel processing that prevents thin lines from disappearing.

[Means to solve the problem]

In order to solve the above technical problems, the first invention sets divided regions determined by a projection method according to a desired reduction ratio on a plane in which pixels to be converted are arranged as shown in FIG.
In an image reduction device that calculates the density of each converted pixel from the density of neighboring converted pixels according to a logical formula corresponding to a divided area in which a converted pixel projected onto the plane is located, a converted pixel located in the vicinity of the converted pixel conversion pixel input means 1 for sequentially inputting density data of each group; and thinning necessary for detecting thinning target pixels on a thin line to be thinned out by reduction from among the conversion pixel group based on the reduction rate. a thinned-out data generating means 2 that generates data; a divided area determining means 3 that determines in which divided area the converted pixel is located based on the reduction rate; and the thinned-out data;
Image reduction characterized by comprising a divided area in which a converted pixel is located and a converted pixel density calculation means 4 that calculates the density of the converted pixel by detecting whether or not it is a thinning target based on the density data of the converted pixel. Device.

On the other hand, the second invention sets divided regions determined by a projection method according to a desired reduction ratio on a plane in which pixels to be converted are arranged as shown in FIG. In an image reduction method that calculates the density of a converted pixel from the density of neighboring pixels to be converted according to a logical formula corresponding to a divided area in which the pixel is located, density data of a predetermined group of converted pixels located in the vicinity of the converted pixel is sequentially input. The process of doing (S1)
, a process (S2) of generating thinning data necessary for detecting thinning target pixels to be thinned out by reduction from among the converted pixel group based on a desired reduction rate; Process of determining area (S3)
and a step of calculating the density of the converted pixel by detecting whether or not it is a thinning target based on the thinning data, the divided area in which the converted pixel is located, and the density data of the converted pixel group (S4
).

[Effect]

Detailed explanation of the present invention (first and second inventions).

When reducing an image according to the present invention, in step S1, the density of a predetermined group of converted pixels located near the converted pixel is determined by projection for the converted pixel whose density is currently being determined. Data is sequentially inputted by the converted pixel input means 1. Here, "predetermined" refers to a group of converted pixels that affects the density of the converted pixel, and for example, when projected onto the converted pixel plane, the group within a certain distance from the converted pixel. This is a pixel to be converted.

In step S2, the thinning data generating means 2 generates thinning data necessary for detecting whether the pixel to be converted is a pixel to be thinned out based on the desired reduction rate.

Here, the pixel to be thinned out means that when the pixel to be converted forms a thin line (black pixels or white pixels arranged continuously in a line) with a line width less than one of the reduction weight, A pixel to be converted is a pixel whose thin line of reduction is erased by the projection method of It is determined based on the reduction rate. This is because whether or not a certain pixel to be converted is thinned out is determined by the position of the adjacent converted pixel, and the position of the converted pixel is uniquely determined by the desired reduction rate.

In step S3, the divided area determining means 3 determines in which divided area the current converted pixel is located based on the reduction ratio.

Further, in step S4, the converted pixel density calculation means 4 calculates the density of the converted pixel based on the thinned-out data, the divided area where the determined converted pixel is located, and the density data of the converted pixel group, The above procedure is repeated for each converted pixel.

That is, in step s4, even if it is determined based on the thinning data that there is a pixel to be thinned out among the pixels to be converted, if the pixel to be thinned out does not form a thin line based on the density data of the group of pixels to be converted. The density of the converted pixel is determined using the same method as before, but in order to prevent the thin line that would be thinned out due to reduction from the density data of the converted pixel group, the thin line pattern is converted. so that it is reproduced to the density of the pixel,
If the pixel to be converted is not to be thinned out, or if it is to be thinned out but does not form a thin line, the density of the converted pixel is output according to a normal projection method.

〔Example〕

Next, a first and second embodiment of the present invention will be described.

First Embodiment> An image reduction device according to a first embodiment is shown in FIG. 3.

This device is a device that reduces the image while preventing thin lines from disappearing in the X-axis direction, and the reduction rate p is 1>p≧173.
This corresponds to

This device includes a shift register 11a as a converted pixel input means 1 that sequentially inputs density data of a group of converted pixels located in the vicinity of the converted pixel when the current converted pixel is projected onto the converted pixel plane; llb, and the reduction ratios p and q in the horizontal direction (X-axis direction) and vertical direction (y-axis direction) as the thinned-out data generation means 2 and the divided area determination means 3. In addition to determining the divided area to which it belongs, the converted pixel input means 1
1, and a logic operation circuit 13 as the converted pixel density calculation means 4.

The shift registers lla and llb sequentially input and move and hold the upper and lower converted pixels of the converted pixel group located in the vicinity of the current converted pixel.

Each stage A-L of each shift register lla, b is shown in FIG.
), and can correspond to the case where the reduction ratio p is 1>p≧173. Incidentally, in the case of reduction exceeding the reduction rate, the number of stages of each shift register may be increased.

The control unit 12 generates thinning data and determines division areas based on the reduction ratios p and q in the X and y axis directions, and in order for the converted pixel density calculation means 4 to detect pixels to be thinned out, Regardless of the reduction rate, the following 5
It is sufficient to generate thinned data of

Here, the thinned-out data is as shown in FIG. 4. For example, considering the X-axis direction, the coordinate position of the immediately previous converted pixel R80 on the converted pixel plane is Xo, and the converted pixel of the current converted pixel R81 is The coordinate position on the plane is xl, and the coordinate position of the next converted pixel R82 on the converted pixel plane is x2.
Regarding the five parameters, xshiftO, the number of converted pixels (shift number) that pass when moving from converted pixel R80 to R81, and xshiftl, the number of converted pixels that pass when moving from converted pixel R8 to RO2, Xo >
0.0. x, >0.0°x2>0.0. xshift
o=1 (or 2), xshiftl=1 (or 2
), and if the condition is satisfied, it is 1'', and if it is not, it is 1-bit data of 110+1 or the xshiftO.

This is 2-bit data necessary to distinguish xshiftl (note that the condition of X□>0.0 is used to determine the divided area, so it can be omitted).

The contents of the calculation performed by the logic operation circuit 13 are performed based on the contents shown in Tables 1.2 and 3.4, as will be described later.

The operation of this embodiment will be explained below.

When reducing the original image by 273 using the device,
For a predetermined converted pixel, a group of converted pixels that belong to the vicinity of the converted pixel, that is, a group of converted pixels surrounding the converted pixel, are sequentially input to the converted pixel input means 11 by projection.

At this time, the pixels to be converted in the upper row of the converted pixels shown in FIG. 4 are sequentially moved and held in the shift register lla,
The shift register llb sequentially moves and holds the converted pixels located below the converted pixel. The timing of the movement is performed for each converted pixel based on the signal from the control section 12.

On the other hand, the control unit 12 controls the previous converted pixel R8°, the current converted pixel R8! , the fourth for the next converted pixel R82
It is judged whether or not the above conditions are satisfied for the three parameters of the coordinate values x0, xl, x2 shown in the figure.
Generate 1-bit data of lj, xshiftO,
Regarding the two-line property of xshiftl, thinned data expressed in 2 bits is generated.

Further, the control unit 12 as the divided area determination means 3 determines the converted pixels Roo, Ro□ based on the reduction ratios p and q.
, Ro1, the division area to which the converted pixel belongs is determined sequentially using the above-mentioned formula (px+0.5) (qy+o, 5)=o, from each coordinate value of the converted pixel.
This is performed based on s, and the divided area from G□ to G8 is expressed as data of 3 and . and is output.

Based on the thinning data thus obtained, the density data of the converted pixel group input to the shift 1 registers lla and b, and the divided area to which the converted pixel belongs, the thinning target is detected, and the density of each converted pixel is detected. will be calculated.

Here, the logical operation circuit 13 calculates the density of the converted pixel as follows.

First, based on the thinning conditions input to the logic operation circuit 13, it is determined according to Table 2 or Table 3 whether the pixel to be converted located near the current converted pixel is to be thinned out.

Note 11 Xo >0.0. X+ >0.0. X Cao>o,
11" of o indicates that the inequality is true, and "o" indicates that the inequality is false.

Note that Table 2 corresponds to the reduction ratio of 1>p≧172, and Table 3 corresponds to 1/2>p≧173.

Now, if the desired reduction rate is 273, this corresponds to the case in Table 2.

Further, it is assumed that an array of pixels to be converted as shown in FIG. 10 is formed.

That is, in FIG. 10, the pixel S to be converted. →S2
Consider a case where the density of □→S3□ is "1" (black) and all other pixels to be converted are 11011 (white).

Assume that the current converted pixel is R1□ in FIG. 10 (corresponding to upper R81 in FIG. 4).

The converted pixel R80 is connected to four nearby converted pixels A (S), B (S21), C (S2□), and D (S).
t
2, the X coordinate value of the converted coordinate has a reduction ratio of 273, so as shown in FIG. 10 or FIG. A signal representing thinned data xshiftl=1 is input.

Then, the logic operation circuit 13 performs a logic operation as shown in Table 2 from the thinned-out data, so that the converted pixel R corresponds to condition 10 in Table 2 and is one of the nearby converted pixels on the diagram. It is detected that the CD is to be thinned out and that the pixels to be converted form a thin line pattern No. 2 as shown in FIG.

Based on the information obtained in this way, that is, the information that the converted pixel R11 is located in the divided area G5, it is determined that there is a converted pixel to be thinned out in the vicinity of the converted pixel, based on the logical formula shown in Table 4. By substituting the information and the density data of the converted pixel into the logical formula, the density of the converted pixel R is calculated.

PI-Xi to book eight B $ 5 $ 5 E car F book H 5 Corresponds to thin line pattern 4 in Figure 5 P2 x 2 zero personnel n book C o D zero E 粁 machine zero H, l ] 5 thin line pattern Corresponds to 3 Q1・×1 Principal n Zero 〇 Owner Rei ti palm mouth; Corresponds to thin line pattern 2 in Figure 5 02~×2 Kyo At8 rich 5i car Pt5*R: To thin line pattern 1 in Figure 5 Compatible Xl: IF (xt>0.0)
OR(xshiftl-2) TIIEN I ELS
E O; From condition 10.13 in Table 2, X2:IF (xs
hifto-2) TIIEN I ELSE Q
According to condition 8 in Table 2, if the converted pixel R is not subject to normal thinning, the density will be 0゛° according to the logical formula in Table 4 (same as Table 1), but under thinning conditions When forming a thin line pattern by substituting , the density of "1" will be calculated from the logical formula in Table 3.

Similarly, the above-described procedure is repeated for the next converted pixel R1□.

In the case of the conversion pixel R 2, the control section 12 sends
. <0.0 and Xi >o, o and shi
Thinned data representing the condition 'ft0=1 is generated,
Further, the control unit 12 determines that the converted pixel is located in the divided region G8.

These information and the density of the converted pixel input to the converted pixel input means 11 are input to the logic operation circuit 13.

The logic operation circuit 13 detects that the thinned-out data corresponds to condition 2 of Table 2, and converts the pixel to be converted S□2. $2□
are the converted pixels A and B to be thinned out. However, since the '1°' information of S1□ and S2□ has already been reproduced in R1□, although it corresponds to thin line pattern No. 011 in FIG. 5, it is not regarded as a thin line pattern, and the corresponding converted pixels in Table 3 are 0'° from the corresponding logical formula of the divided area to which it belongs
becomes.

Conversely, the pixel to be converted S1□→S2□→S3□ is “0″′
(white) and all other converted pixels are "1'' (black), the condition io in Table 2 applies to the converted pixel R11, and the thin line pattern No. 4 in FIG. Therefore, since thin line pattern No. 4 has appeared, the density of the converted pixel R□ is set to 0''' from the logical formula in Table 4. In addition, the density of the converted pixel R□2 is not regarded as a thin line pattern as in the previous example, but is
”.

Note that when the reduction ratio is 1/2>p≧173, the logical operation circuit 13 uses Tables 3 and 5 described above to calculate the density of the converted pixel.

PI-A Sui C attack fnE grass FtG request 1 rule attack and heat L
1 Corresponds to thin line pattern 7 of 61 section 7 P2-XltA
Lazy B zero CtO want F book ridge to nationals L; Figure 61 Corresponds to thin line pattern 8 in Figure 7 p3-XZtA old E tl
lJtK bill! (Go, 5.a, R), Figure 6. Thin line of lZ7 corresponds to 9-:/30-36 d P4-λ rippling ridge 0 F wealth G junior H special I hot J customer ridge 8 Figure 61 Corresponds to thin line pattern 5 in Figure 7 ps- x 3 house At attackctoxEtP*a
*o*+t, +txtL; I2e, corresponds to Fig. 7 tv thin line pattern 9-8 P6-X4tljDtGtkawatJtK
Ridge 1 (person, fi, eJ); [26, Corresponding to thin line patterns 23 to 29 in Fig. 7 Ql-λ zeron heat o*1tPxa
trItThj house RYE; Fig. 60 Corresponds to thin line pattern 3 in Fig. 7 Q2-X1 line λ quasi i hard *ill curtain io'
*c request 11 zero■寞] Summer R book [: Figure 61 Corresponds to thin line pattern 4 in Figure 7 q3-x2 zero XonutF*it kotRtE
I (C, D, G, +1); Fig. 6, 1lZ7 thin line pattern 16-22 corresponds to Q4-A, 5 o 1tptj
*j (zero Y book book R thing [6th ward 61 thin line pattern l
Compatible with QS-X3 Wealth λ2 Silence C Less Heat Etnati+Zero■
This code corresponds to the thin line pattern 2 in Figures 6 and 7.
-x+ti:z5ta*fit? xjtRtE family (^,
n, F, r), Figure 63 Thin line patterns 9 to 15 in Figure 7
Compatible with Xl:IF (xshikawa-3) TIIEN
I ELSE O, from condition 15.16 in Table 3, X2:I
F((xshiftl-12) AND (X,)0.
[1)) 0R((xshiftl-3) AND (
X, (0,0)) TIIEN I ELSE0; Condition 10.13 of Table 3 X3:IF (xshift
-3) Eight no (xo(Q,Q) THEN I E
LSE O; From condition 5 in Table 3, X4+lF (xshi
ftO-3) TIIEN I ELSE Q
; From Condition 8 in Table 3 Note) The logical formula for the pixels in parentheses is as follows.

(A, B, E, F)・A book F order BtE◆A year [lKyoE◆
A2BtF◆A Kyo E Chapter F+〇ZeroE Book F◆^Book B Customer E House F
Although only the X-axis direction has been described above, the same applies to the y-axis direction.

Further, in this embodiment, the reduction ratio has been explained using a table only when 1>p≧173, but it is not limited to this range, and the reduction ratio is not limited to this range. As the number of pixels involved in conversion increases, it is necessary to create a table in a similar manner to Tables 2.3 and 4. It is sufficient to generate thinned data that satisfies the five conditions described above as the thinned data used for the determination.

In addition, FIG. 6 shows a thin line pattern to be thinned out with a line width of 1 when the reduction rate p is 1/2>p≧173,
FIG. 7 shows a thin line pattern to be thinned out with a line width of 2.

Next, unlike the first embodiment, this embodiment differs from the first embodiment in that when detecting a thinning target pixel for the current conversion pixel from among the conversion pixel group, the conversion pixel is The area formed by the four converted pixels that are located near the converted pixel when projected onto the converted pixel plane is divided into four divided areas, and the converted pixel at the current moment is divided into which divided area. This is intended to reduce the number of pixels to be converted that are required when determining which pixels are to be thinned out by detecting whether the pixels belong to the .

FIG. 11 shows an image reduction device according to this embodiment.

This apparatus includes a converted pixel input means 111 which sequentially inputs density data of the converted pixel group located in the vicinity of the current converted pixel by projection, and a converted pixel input means 111 in which density data of the converted pixels of the original image is stored. a two-dimensional image memory 110, and access control to the two-dimensional image memory 110;
A control unit 112 that controls the converted pixel input means 111, determines the divided area to which the current converted pixel belongs as the divided area determination means 112 and the thinned-out data generation means 2, and generates the thinned-out data, and the converted pixel group. and a logic operation circuit 113 as a converted pixel density calculating means 4 for calculating the density of the converted pixel based on the density of the converted pixel, the divided area, and the thinned-out data.

As shown in FIG. 11, the converted pixel input means 111 has eight shift registers 111a to 111h.

The shift registers 1lla and 1lla and b hold pixels to be converted on the first image line.

Similarly, shift register 1llc,d is on the second line,
The shift registers 1lle and 111f sequentially move and hold the pixels to be converted on the third line, and the shift registers 111g and h on the fourth line.

The control unit 112 determines the divided area to which the current converted pixel belongs based on the desired reduction ratios p and q in the X-axis direction and the y-axis direction, and also determines the divided area to which the currently converted pixel belongs It generates thinning data necessary for detecting whether or not a pixel is to be thinned out, and determines which of the above-mentioned four-section regions the current converted pixel belongs to.

Here, as shown in FIG. 12, the thinned-out data is the coordinate position X of the immediately preceding converted pixel R1o within the converted image plane in the X-axis direction. >0.0, the coordinate position of the current converted pixel R, x1>0.0, the coordinate position of the next converted pixel R, x2>0.0, and whether or not the following conditions are satisfied, and the converted pixel R □. Number of pixels to be converted that pass when transitioning from to R1 □
This is 1-bit or 2-bit data that displays the value of iftl, and similarly in the y-axis direction, it is the value of the immediately preceding converted pixel R.
Engineering. The coordinate position y in the transformed image plane. , the coordinate position y of the current converted pixel R□1, the coordinate position y2 of the next converted pixel R21 on the converted image plane, the converted pixel R8□ to Ro
The number of converted pixels passed when moving to l yshift
This is data that displays the value of yshiftl, the number of converted pixels that pass when moving from the converted pixel R0, to R2□.

Furthermore, in this embodiment, 4 pixels surrounding the current converted pixel are used.
The data as to which area of a four-part area in which the square area formed by the pixels to be converted is divided into four areas is used as the thinning data for detecting the thinning target.

This is to enable the thinning target to be detected without significantly increasing the number of reference pixels to be converted.

That is, as shown in FIG. 14, the converted pixel is g□+g
If it is located at 4, then b, n+1. n+2. The n+3rd line data is sufficient, and the converted pixel is g2. When located in the g:+ segmented area, it is sufficient to read the data on the n-2, n-1, n, and n+1 lines.

In other words, when the converted pixel is located in the segmented area of gx and g4, the pixels are A, B, C, D, G, and H.

Referring to 12 pixels of I, J, O, P, Q, and R, the converted pixel is g2. g Pixel A if located in the 3-section area
, B, C, E, F, 12 pixels L, M, N, S, and T are referenced.

Further, the logic operation circuit 113 calculates the current density of the converted pixel based on the thinned-out data, the density of the pixel to be converted, and the divided area, and for this purpose, for example, Tables 6 to 9, which will be described later, are used. (Reduction rate 1>p, Q≧17
Calculations will be made based on the details listed in 3).

Next, the operation of the apparatus according to this embodiment will be explained.

A case will be described in which the converted pixels stored in the two-dimensional image memory 110 are reduced at a reduction rate of p=2/3.

The control unit 112 reads out the image memory 110 in which the original image to be subjected to reduction conversion is stored in units of a certain word length (4 bits in the case shown in FIG. 11), A shift register 111 serving as the converted pixel input means 111 holds each density data of a group of converted pixels located in the vicinity of the converted pixel for a maximum of four lines sequentially.
Input a, c, e, and g.

That is, as shown in FIG. 14, in the case of 1>p and q≧172, and when the current converted pixel is projected into a grid formed by the converted pixels ABCD in the converted image plane, , each density data of each converted pixel group A-L for the conversion pixel is stored in each shift register as C' for each line (n-1st line, nth line, n+1st line, n+2nd line). +e* Input to g.

The control unit 112 moves the density data of each pixel to be converted, which is input and held in the shift registers 1lla, c, e, and g, at a predetermined timing, and moves and holds them in the shift registers b, d, f, and h. The logic operation circuit 113
It will be sent to

At this time, as shown in FIG. 13, the control unit 112 divides the lattice formed by the four pixel groups located near the current pixel to be converted into four sections. Segmented area gxtg2+g3*g4 That is, as shown in FIG. 13, the minimum lattice unit formed by four pixels to be converted in the plane where the pixels to be converted are arranged is four rectangles (including squares) of equal area. 3-bit data for distinguishing the area is sent to the logic operation circuit 113.

At the same time, the control unit 112 controls the thinning data necessary for determining thinning target pixels from the group of converted pixels read out from the two-dimensional image memory 110 and held in the converted pixel input means 111. is generated and sent to the logic operation circuit 113, and it is determined to which region of the divided regions G□ to G8 the current converted pixel belongs, and a signal indicating the region is sent to the logic operation circuit 113.

Then, the hanging chain arithmetic circuit 113 receives the information about the four divided areas g-g4 outputted by the control unit 112,
The density of the pixel to be converted is calculated by detecting whether or not the pixel is to be thinned out based on the thinning data, information on the divided area, and density data of the pixel group to be converted.

Below, the case where the reduction ratio p (reduction ratio in the x-axis direction) and q (reduction ratio in the y-axis direction) are both 273 will be explained by dividing it into ■ Thin line pattern perpendicular to the X axis and ■ Thin line pattern perpendicular to the X axis. do.

■For a thin line pattern perpendicular to the X-axis According to Figure 19, consider the case where the converted pixel S□2→S2□→S3□ is “1pa (black)” and all other converted pixels are (white). In this case, the conditions in Table 6 determined by the thinning data and the four-part area (reduction rate is 273, so Table 6-1
or using Table 6-2), based on the thin line pattern of the converted pixel group to be erased as pixels to be thinned out as shown in FIG. 15 and the logical formula corresponding to each divided area shown in Table 7. The logical operation circuit 113 calculates the density of the converted pixel at the current time.

Table 6-1.1 Detection conditions for converted pixels to be thinned out when p, Q≧D! Table 6-2.1 Detection conditions for converted pixels to be thinned out when p, Q≧D 1 piece Pi-Y (pattern so, 1); Figure 1
Corresponding to thin line pattern 1 of 5, P2 tone Y 1 line (pattern No.
, 2) H Corresponding to thin line pattern 2 in Figure 15 P: 1-Xl!
(Pattern No. 3); Corresponds to thin line pattern 3 in Fig. 15 P4-XI car (Pattern No. 4); f315 (
Q thin line) < Corresponds to turn 4 711 According to the conditions in Table 2, if pattern No. 1.2 is the target -1, etc.! 0x16
According to the conditions in Table 2, pattern No. 3.4 is the target proverb 1. 1110 Now, for the current converted pixel R□1, the surrounding four pixels ABCD located near the pixel R□1 on the pixel plane to be converted are S-process □ and S-process, respectively, as shown in FIG. 2. S2. , S2□.

The position of the converted pixel R1□ belongs to the divided area G5 xshif
tl=1. yshiftl=1. x2>0.0. y2>
0.0 corresponds to condition g□-NO04 in Table 6-1.

Therefore, the target thin line pattern is g□-N in Figure 15.
o, 1 to No, 4.

The thin line pattern in this example corresponds to gl-No.3.

Therefore, the density of the converted pixel R11 is set to "0" from the logical formula 05 in Table 7 when the converted pixel group is not a normal thinning target and the converted pixel group does not form a thin line pattern. However, if there is a pixel to be thinned out in the pixel group to be converted, and the thin line pattern g knee No.
3 appears, so it is set as R11l from the logical formula 01 in Table 7.

Here, since the reduction ratios p and q are 2/3, Table 7 is used.

Similarly, for the converted pixel R1□, the condition g in Table 6-2 is satisfied 4-
This corresponds to No. 4, and g 4-No. 3 in FIG. 15 is a thin line pattern.

However, Sl. , S2□ has already been expressed by pixel R□1, so the current converted pixel R0
The density of □ is set to "0" by adopting the value of S13.

The above process is the same for pixels R2□ and R22.

Conversely, among the pixel group to be converted, S 2 → S 22 → S 3
□ is “0” (white) and all other pixels to be converted are 1 (II
In the case of I (black), conditions g, - for the converted pixel R1□
According to No. 4, the thin line pattern g, -No. 4 corresponds to the thin line pattern g.

Therefore, the density of the converted pixel R11 is set to 0' from the logical formula in Table 7.

As in the previous example, the density of the converted pixel R1□ is not regarded as a thin line pattern, but is 1gIII according to the logical formula 08 in Table 7 mentioned above.
The value of is adopted.

The above process is the same for the converted pixels R2□ and R2□.

■In the case of a thin line pattern perpendicular to the y-axis From FIG. 19, the converted pixel S21 among the converted pixel group
→S2□→Consider the case where S23 is “1” (black) and all other pixels to be converted are “0” (white).In this case, the conditions in Table 6, the thin line pattern in FIG. 15, and the logic in Table 7 are considered. The expressions correspond.

The four surrounding converted pixel groups A, B, C, and D located in the vicinity of the converted pixel on the converted image plane projected to the current converted pixel R11 are S, S1, and S2, respectively.
, S2□. Conversion pixel R engineering.

The position belongs to the divided area G, and xshifto=1
.

yshiftl=1, x2>0.0, y2>0
．． 0 corresponds to the conditions in Table 6-1.

Therefore, the target thin line pattern is g□-No, 1~
It is N004. The thin line pattern in this example is g□
- No, corresponds to 1. Therefore, the density of the converted pixel R0□ falls under Table 7 when there is no pixel to be thinned out in the converted pixel located near the converted pixel, or when a thin line pattern is not formed even if it is thinned out. The logical expression corresponding to the divided area G5 gives 0'', but in this example, since the thin line pattern g□-NO91 has appeared, the corresponding logical expression in Table 7 mentioned above gives 1.

Similarly, for the converted pixel R2□, the condition g 2 in Table 6-2 is satisfied.
-No. 4 corresponds to g2-No. 1 in FIG. 15.

However, since the thin line pattern existing in the converted pixel S2□+S2□ is already expressed by the pixel R□1, the
From the corresponding logical formula, the density of the pixel R□1 is set to 0% by adopting the value of the density data in S31.

The above process is for pixel R□2. The same applies to R2□.

Conversely, the converted pixel S2□→S2□→S23 is 0゛'
(white) and all other converted pixels are 11111 (black), for the converted pixel R1□, the condition g knee No.
4, the thin line pattern corresponds to g, -No., and 2.

Therefore, the density of converted pixel R1□ is 0'' from the logical formula in Table 7.
shall be.

As in the previous example, the density of the converted pixel R2□ is not regarded as a thin line pattern and is set to "1". The above process is for pixel R□2. R2□
The same applies to

As described above, according to this embodiment, thin line white or black patterns can be made equally, without increasing the line width, and with fewer converted pixels (
1) 8 pixels when p, q≧1/2, 1/2>p, (
If 1≧1/3, then 12 pixels), it is possible to reproduce a thin line in the converted image.

In addition, although the case where the reduction rate is 1>p, q≧172, and especially the case of 273 has been explained above, the case is not limited to the reduction rate, and the reduction rate is 1/2>p, q≧173.
In the case of 1>p, it is possible to calculate using the device according to Table 8-1, Table 8-2, and Table 9.
N+1 lines of image data and 4N for q≧1/H
Image data can be reduced by referring to the thinned-out data described above without erasing the thin line pattern just by referring to the pixel groups to be converted.

Table 8-1. ->p, pixel detection condition for conversion to be thinned out when q≧th PL-Yl Gin-shai, ■) 2 Figure 171 Corresponds to thin line pattern 1 in Fig. 18 P2-Yl Gin-shai, 2
): Figure 17. Corresponds to thin line pattern 2 in Figure 18 P3-Y2
t(A>-7No, )); Figure 170
FIG. 1v) Corresponding to thin line bar 9-3 P4-Y2t()sha-bansha, 4): FIG. P5 corresponding to thin line pattern 4 in Figure 18
-Y3o (the most embarrassing, odd number lton of 1-18); Figure 17. Corresponds to thin line patterns 1 to 18 (odd numbers) in Figure 18 P
6-Y3 groaning No. 1-1&7) even aA9-y)'
; Figure 176 Figure 18 Thin line butter:, t 1-18
(Even number) corresponds to P7 ~
;Figure 17. Figure 18171 Thin line pattern 19
Compatible with P8-X11 (+G-n1io, 20)
; I2117. Figure 18 Thin wire bar 9-20
Corresponds to P9-X2 Utsunosha, 21)
:Figure 17. Figure 1 1'IQ corresponding to small line pattern 21
・X9 (#-7+io, 22); Figure 1
7. [Corresponds to thin line pattern 22 of 2111 P11-X3
Book (Ruto:/No, +9-36 odd ashy) r Figure 17
Compatible with thin line patterns 19 to 36 (odd numbers) of 11 cases P12
・X31 (If) Ao, 19-36 even number 1 ton):
Figure 17. Corresponding to thin line patterns 19 to 36 (even numbers) in Figure 1 718 From the conditions in Table 3, when targeting thin line patterns of 1Ij-7No, 1, 2 -1, etc! OY2・Table 3
Based on the conditions, if the target is the thin line pattern of bar←:/No, j, 4, then 1. Others = OY31 From the conditions in Table 3, reward 7
No. 1-18 when targeting thin line patterns
．． Others -0x1=From the conditions in Table 3, B←;4io, +9.
When targeting 20 thin line patterns - 1, etc! Ox2
: According to the conditions in Table 3, if the thin line pattern of Bar needle near + 10, 21, 22 is targeted -1, other = Ox33 From the conditions of Table 3, the target is the thin line pattern of No. 3, 19-11i Case=1. Other=O Incidentally, in the past, 1>p,
2 to detect a thin line pattern for q≧1/N
It was necessary to refer to N lines of image data and 8N+4 pixels to be converted.

As described above, according to this embodiment, the thin line can be reproduced in the converted pixel equally with the white or black pattern of the thin line, without increasing the line width.

(Effects of the Invention) As explained above, in the present invention, thinning data necessary for detecting pixels to be converted having a thin line pattern whose line width is less than one of the reduction weight is thinned out by reducing the original image. is generated based on the desired reduction ratio, and the thin line pattern is automatically reproduced in the converted pixels based on the data.

Therefore, according to the present invention, reliable reduced images can be obtained at high speed.

[Brief explanation of the drawing]

FIG. 1 is a block diagram of the principle of the first invention, FIG. 2 is a flowchart of the principle of the second invention, FIG. 3 is a diagram showing an image reduction device according to the first embodiment, and FIG. 4 is a diagram of the principle of the second invention. FIG. 5 is an explanatory diagram of the thinning data necessary for detecting the converted image to be thinned out according to the embodiment, and FIG. 5 shows that the reduction rate according to the first embodiment is 1>p≧
172, FIG. 6 is a diagram showing a thin line pattern when the reduction ratio is 1/2>p≧173 according to the first embodiment, and FIG. FIG. 8 is an explanatory diagram showing the principle of the projection method. FIG. 9 is a diagram showing the divided regions of the high-speed projection method. FIG. A diagram showing the positional relationship between the converted pixel and the converted pixel when the reduction rate is 2/3 according to the first embodiment, FIG. 11 is a diagram showing the image reduction device according to the second embodiment, and FIG. 12 is an explanatory diagram of thinning data necessary for detecting a converted image to be thinned out according to the second embodiment, FIG. 13 is a diagram showing four divided regions, and FIG. 14 is a diagram according to the second embodiment. A diagram showing reference pixels in the case of reduction ratio 1>p and q≧1/2, FIG. 15 is the reduction ratio i according to the second embodiment.
>p, q≧172, first diagram showing a thin line pattern
Figure 6 shows the reduction ratio 1/2>I) and q≧1 according to the second embodiment.
/3, FIG. 17 is a diagram showing a thin line pattern when reduction ratio 1/2>p and q≧1/3 according to the second embodiment, and FIG. 18 is a diagram showing the reference pixel in the case of the second embodiment. FIG. 19 shows the thin line pattern when the reduction rate is 1/2>p and q≧1/3 according to the second embodiment, and FIG. FIG. 20, a diagram showing the positional relationship with converted pixels, is a block diagram related to a conventional example.

Claims (3)

[Claims] (1) Set a divided area determined by a projection method according to the desired reduction ratio on the plane where the pixels to be converted are arranged, and use a logical formula corresponding to the divided area where the converted pixels projected onto the plane are located. In an image reduction device that determines the density of each converted pixel from the density of neighboring converted pixels according to the method, converted pixel input means (1) sequentially inputs density data of a predetermined group of converted pixels located in the vicinity of a converted pixel; , a thinning data generating means for generating thinning data necessary for detecting thinning target pixels on a thin line thinned out by reduction from among the group of pixels to be converted based on the reduction ratio (
2); divided area determining means (3) for determining in which divided area the converted pixel is located based on the reduction ratio; the thinned-out data, the divided area where the converted pixel is located, and the converted pixel group. An image reduction device comprising a converted pixel density calculation means (4) that calculates the density of the converted pixel by detecting whether or not the converted pixel is to be thinned out based on density data of the converted pixel. (2) Set a divided area determined by a projection method according to the desired reduction ratio on the plane where the pixels to be converted are arranged, and use a logical formula corresponding to the divided area where the converted pixels projected onto the plane are located. In an image reduction method for determining the density of a converted pixel from the density of neighboring converted pixels according to the method, the step (S1) of sequentially inputting density data of a predetermined group of converted pixels located in the vicinity of the converted pixel; a step (S2) of generating thinning data necessary for detecting thinning target pixels on a thin line thinned out by reduction from among the converted pixel group based on the converted pixel group; The process of making a judgment (S
3), and a step (S4) of detecting whether or not the pixel is to be thinned out based on the thinning data, the divided area in which the converted pixel is located, and the density data of the converted pixel and calculating the density of the converted pixel. An image reduction method characterized by: (3) As the thinning data generating means (2), when the horizontal axis or the vertical axis is set as the x-axis direction as a coordinate plane based on the four pixels of the converted pixels surrounding the converted pixel, the immediately preceding The position coordinate x_0 of the converted pixel on the converted image plane, the current position coordinate x_1, and the position coordinate x_2 of the next converted pixel on the converted image plane, so that x_0>coordinate reference value, x_1>coordinate reference value, x_2 >The truth or falsity of the inequality of the coordinate reference value and the number of converted pixels that exist between the immediately preceding and current converted pixels or the number of converted pixels that exist between the current converted pixel and the next converted pixel. 2. The image reduction apparatus according to claim 1, wherein the image reduction apparatus generates the data as thinned-out data.