JP3409713B2 - Image processing device - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image processing apparatus, and more specifically, to "color primary colors (R, G, B)" handled by a color printer, color monitor, color image data, etc. The three primary colors (C, M, Y
(, K)) ”for performing color conversion processing.

[0002]

2. Description of the Related Art In recent years, with a computer, anyone can easily handle a color image on the screen. The color image is easily input / output between various devices. By the way, when each signal of the three primary colors (R, G, B) of light read by the color scanner is recorded on the recording paper by the color printer, each signal of the three primary colors (C, M, Y (, K)) of the color is recorded. Must be converted to. In televisions and displays, various colors can be expressed by combining the three primary colors of light (R, G, B). This method starts from a state where no color is shining (all are 0: black), and a state where all colors are shining (all are 100: white), by adding three colors, various colors are added. In order to express, it is called additive color mixture (color additive method).

However, color can be expressed by additive color mixing when it can emit light by itself. For this reason, when light is not emitted by itself, for example, when recording is performed on recording paper, the reflected light expresses a color. To be precise, it expresses a color by absorbing (decreasing) a part of the color of the incident light and reflecting the remaining color. Such a method is called subtractive color mixing (subtraction method),
Ink and paint express various colors by this subtractive color mixture.

In this subtractive color mixture, the basic color is not the additive color mixture of red (R), green (G) and blue (B), but when mixed, it becomes black (a color that does not reflect light at all) cyan ( The three colors are C), magenta (M), and yellow (Y). Here, theoretically, if the three colors of C, M, and Y are mixed,
It will be "black", but in order to express "black" more clearly,
A method of recording with four colors of C, M, Y, and K using black (K) ink is generally performed. By the way, conventionally, in the color conversion processing from the RGB signal to the CMYK signal,
An interpolation calculation method is used. Since the interpolation calculation requires high-speed processing, it has been executed by hardware.

[0005]

However, if the interpolation calculation processing is configured by hardware, a dedicated LSI is required, the circuit becomes complicated, and it is not possible to reduce the size and cost of the LSI. Further, since it is configured by hardware, there is a problem that a dedicated circuit is required and it is not possible to support various interpolation calculation methods. On the other hand, in recent years, since the processing speed of the MPU has improved, it is becoming possible to perform high-speed interpolation calculation using software. However, since the required processing speed of the color printer is also increasing, it is necessary to perform the color conversion processing at a higher speed than the processing speed of the color printer. The present invention has been made to meet such market demands, and an object of the present invention is to provide an image processing apparatus capable of high-speed color conversion processing even using software.

[0006]

In order to achieve the above object, according to the invention of claim 1, in an image processing apparatus for converting a color space, at least grid point data of grid points surrounding a conversion point are calculated. In addition to calculating the reference address of the grid point data at the reference grid point and sequentially incrementing from the reference address, the grid point necessary for the interpolation calculation of the conversion point And a control means for reading out the data and performing the interpolation calculation of the conversion points.

According to a second aspect of the present invention, in the image processing apparatus according to the first aspect, the storage means stores whether or not the exception processing is performed, and the control means is a case where the exception processing is performed. Interpolation calculation of the conversion point is performed depending on the case where it does not apply.

According to a third aspect of the present invention, in the image processing apparatus according to the first aspect, the control means performs the interpolation calculation of the conversion points without determining whether or not the exceptional processing is applied.

In the embodiments of the invention described below, the "storage means" described in the claims or means for solving the problems corresponds to the ROM 12, and the "control means" also includes the MPU 11, the ROM 12, and the ROM 12. RAM13
Equivalent to.

[0010]

BEST MODE FOR CARRYING OUT THE INVENTION [First Embodiment] Case of Cube Interpolation (8-Point Interpolation) A first embodiment in which the present invention is embodied in a facsimile machine will be described below with reference to the drawings.

As shown in FIG. 1, a facsimile apparatus 1 as an image processing apparatus includes an MPU 11, a ROM 12 and an RA.
The M11, the reading unit 14, the recording unit 15, the operation unit 16, the display unit 17, the image memory 18, the codec 19, the modem 20, and the NCU 21, and each unit 11 to 21.
Are respectively connected via the bus 22.

The MPU 11 controls each part of the facsimile machine 1. The ROM 12 stores a program for controlling the facsimile device 1. Also, RO
M12 stores various data necessary for interpolation calculation. Specifically, a look-up table (LUT for short) of grid point data shown in FIG. 5 described later and FIG.
The LUT of the weighting coefficient W (w1 to w8) shown in FIG.
The RAM 13 temporarily stores various data related to the facsimile device 1. The reading unit 14 includes a color scanner, reads image data on a document, and outputs RGB signals.

The recording unit 15 is composed of an electrophotographic color printer, a thermal printer or an ink jet color printer, and receives image data of a document read by the reading unit 14 on a recording paper in a received image data or a copy operation. Record. The operation unit 16 includes a ten-key pad (including * and # keys) 16a for inputting a FAX number, abbreviated number registration, and an abbreviated key 16 for transmitting from an abbreviated number.
b, various operation keys such as a start key 16c for starting a document reading operation and a communication / copy key 16d for setting a "communication (FAX)" operation or a "copy" operation. The display unit 17 including an LCD or the like displays various information such as the operating state of the facsimile device 1.

The image memory 18 temporarily stores the received image data and the image data read by the reading unit 14 and compression-coded by the codec 19 by the MH, MR or MMR system. The codec 19 encodes the image data read by the reading unit 14 by the MH, MR, MMR method, or the like for transmission. The codec 19 also decodes the received image data. The modem 20 uses the ITU-T recommendation T.264. 30 based on the facsimile transmission control procedure according to V.30. 17, V.I. 27te
r, V. The transmission / reception data is modulated and demodulated according to any one of 29 or the like. The NCU 21 has a function of controlling connection with the telephone line L and detecting sending and receiving of a dial signal corresponding to the FAX number of the other party.

Next, a color conversion process by an interpolation calculation method using a look-up table (LUT) will be described.
First, cubic interpolation (8-point interpolation) will be described with reference to the drawings. The RGB signals output from the reading unit 14 are each composed of 8 [bit] data.
Therefore, each RGB signal has data of 256 gradations.

As shown in FIG. 2, in RGB space,
For each lower 4 [bit] of the RGB signal, in other words, when each axis is divided into 16, the number of intersections between each axis and each grid point is "1.
7 ”. That is, the intersection of each axis and each grid point is
"0, 16, 32, ..., 240, 255".
Here, when the intersection points of each axis and each grid point are expressed by hexadecimal numbers, “00h, 10h, 20h, ..., F0h, FF
h ". In addition, "h" shows a hexadecimal number. same as below.

That is, "0-240 (00h-F0
Up to “h)”, there is an intersection every time “16” increases,
"15" up to "240-255 (F0h-FFh)"
When it increases, it becomes the intersection. Therefore, if each grid point data corresponding to the RGB signal is stored in advance as a “three-dimensional LUT”, in the case of the RGB signal corresponding to the grid point,
The color conversion process can be performed quickly from the grid point data without any calculation.

However, in the actual color conversion processing, it is extremely rare that the conversion points correspond to the grid points. Therefore, when the conversion point does not correspond to the grid point, the interpolation calculation is performed based on the grid point data of the grid point in the cube surrounding the conversion point. Therefore, each upper 4 [bit] of the RGB signal at the conversion point is (RH, GH, B
H), it is possible to determine in which cube the conversion point is located based on the respective values of (RH, GH, BH).

Therefore, as shown in FIG. 3, "C, M,
L in which each grid point data corresponding to the “Y, K” component is stored
Assume a UT. That is, "C" component, "M" component,
The upper 4 [bit] of each RGB signal at the conversion point P corresponding to the “Y” component and the “K” component in this order are represented by (R
H, GH, BH), (0,0,0), (1,
0,0), (2,0,0), ..., (16,16,1)
5), the grid point data is stored in the order of the grid points of (16, 16, 16).

Here, as shown in FIG. 4, the color image data output from the reading unit 14, that is, the conversion point represented by the RGB signal is defined as "P", and each lattice point of the cube surrounding the conversion point P. Is “A to H”, the reference address HADR (N) of the grid point data of the “C” component in the grid point data of the “C, M, Y, K” components shown in FIG. = A to H) can be shown as follows.

[0021]   HADR (A) = (RH) + (GH) * 17 + (BH) * 17 ∧2 ……… (Equation 1-1)   HADR (B) = (RH) + (GH) * 17 + (BH + 1) * 17∧2          = (RH) + (GH) * 17 + (BH) * 17∧2 + 17∧2          = HADR (A) +289 ……… (Equation 1-2)   HADR (C) = (RH) + (GH + 1) * 17 + (BH) * 17∧2          = (RH) + (GH) * 17 + (BH) * 17 ∧2 + 17          = HADR (A) +17 ……… (Equation 1-3)   HADR (D) = (RH) + (GH + 1) * 17 + (BH + 1) * 17∧2          = (RH) + (GH) * 17 + (BH) * 17∧2 + 17 + 17∧2          = HADR (A) +306 ……… (Equation 1-4)   HADR (E) = (RH + 1) + (GH) * 17 + (BH) * 17∧2          = (RH) + (GH) * 17 + (BH) * 17∧2 + 1          = HADR (A) +1 ……… (Equation 1-5)

HADR (F) = (RH + 1) + (GH) * 17 + (BH + 1) * 17 ∧ 2 = (RH) + (GH) * 17 + (BH) * 17 ∧2 + 1 + 17∧2 = HADR (A) +290 ……… (Equation 1-6) HADR (G) = (RH + 1) + (GH + 1) * 17 + (BH) * 17∧2 = (RH) + (GH) * 17 + (BH) * 17 ∧2 + 1 + 17 = HADR (A) +18 ……… (Equation 1-7) HADR (H) = (RH + 1) + (GH + 1) * 17+ (BH + 1) * 17∧2 = (RH) + (GH) * 17 + (BH) * 17∧2 + 1 + 17 + 17∧2 = HADR (A) +307 ……… (Equation 1 -8) Note that "*" indicates multiplication and "^" indicates exponentiation. Less than,
the same.

Here, the grid point "A" in the "C" component
When the offset to the next component “M, Y, K” is examined with reference to, the offset M (os) to the next component “M”
Can be shown as: M (os) = 17∧3 = 4913 (Equation 1-9) Further, the offset Y (os) to the next component “Y” can be expressed as follows. Y (os) = 17∧3 * 2 = 4913 * 2 = 9826 (Equation 1-10) Further, the offset K (os) to the next component “K” can be expressed as follows. K (os) = 17∧3 * 3 = 4913 * 3 = 14739 ……… (Equation 1-11)

Therefore, when the grid point data is stored as shown in FIG. 3, the above (formula 1-2) to (formula 1-8) are stored.
As is clear from the above, when the reference address of the grid point "A" in the "C" component is used as a reference, the grid point "B ~
The offset of the reference address in “H” is the grid point “B
~ H ". Therefore, in order to obtain the grid point data of the grid points “B to H”, the address must be calculated for each grid point. In addition, the above (Formula 1-9) to
As is clear from (Equation 1-11), the next component “M,
The reference address of the grid point “A” in “Y, K” also differs from the reference address of the grid point “A” in the reference “C” component by the next component “M, Y, K”. Therefore, in order to obtain the grid point data of the grid points "A to H" in the next component "M, Y, K", the address is calculated every time the next component "M, Y, K" is changed. There must be.

Therefore, as shown in FIG. 5, each upper 4 [bit] of the RGB signal at the conversion point P is (RH, GH,
BH), (0,0,0), (1,0,0),
(2,0,0), ..., (15,15,14), (1
5, 15 and 15), the grid point data of eight grid points “A to H” are stored for each CMYK. Specifically, for (0,0,0), the grid point data of grid points “A to H” in the “C” component, the grid point data of grid points “A to H” in “M”, and “ The grid point data of the grid points “A to H” in the “Y” component and the grid point data of the grid points “A to H” in the “K” component are stored.

Then, for (1,0,0), "C"
Grid point data of grid points “A to H” in the component, “M”
Grid point data of grid points “A to H” in the component, “Y”
Grid point data of grid points “A to H” in the component, “K”
The grid point data of grid points “A to H” in the component is stored. And finally, for (15,15,15),
Grid point data of grid points “A to H” in the “C” component,
Grid point data of grid points “A to H” in the “M” component,
Grid point data of grid points “A to H” in the “Y” component,
The grid point data of grid points “A to H” in the “K” component is stored.

If the data is stored in this way, the grid point data to be stored is increased, but the grid point "A" in the "C" component is increased.
If the reference address HADR (A) of is determined, the reference addresses of the lattice points “A to H” can be obtained only by sequentially incrementing. In other words, by sequentially incrementing from the reference address HADR (A) of the grid point “A” in the “C” component, grid point data of grid points “A to H” can be obtained. Specifically, when the reference address HADR (A) of the grid point “A” in the “C” component is incremented seven times, all grid point data in the “C” component can be obtained. Then, by incrementing eight times, all grid point data in the “M” component can be obtained. In addition, if it is incremented eight times, all grid point data in "Y" can be obtained. Further, if it is incremented eight times, all grid point data in "K" can be obtained.

Therefore, the grid point "A" in the "C" component
If the reference address HADR (A) is known, all grid point data in the “C, M, Y, K” components can be sequentially obtained by incrementing 31 times from the reference address HADR (A). Therefore, (RH, GH,
BH) grid point data includes (0,0,0) to (1
6, 16, 16), not (0,
It is only necessary to store the grid point data from 0,0) to (15,15,15).

Therefore, by storing the grid point data as shown in FIG. 5, the reference address HADR (A) of the grid point "A" in the "C" component can be shown as follows. HADR (A) = (RH + GH * 16 + BH * 16∧2) * 4 * 8 = (RH + GH * 16 + BH * 256) * 32 ……… (Equation 2-1) Also, “C” The reference address HADR (N) (where N = B to H) of the lattice points “B to H” in the component can be shown as follows. HADR (B) = (RH + GH * 16 + BH * 16∧2) * 4 * 8 + 1 = HADR (A) +1 ……… (Equation 2-2) HADR (C) = (RH + GH * 16 + BH * 16∧2) * 4 * 8 + 2 = HADR (A) +2 ……… (Equation 2-3) HADR (D) = (RH + GH * 16 + BH * 16∧2) * 4 * 8 + 3 = HADR (A) +3 ……… (Equation 2-4)

HADR (E) = (RH + GH * 16 + BH * 16∧2) * 4 * 8 + 4 = HADR (A) +4 ……… (Equation 2-5) HADR (F) = (RH + GH * 16 + BH * 16∧2) * 4 * 8 + 5 = HADR (A) +5 ……… (Equation 2-6) HADR (G) = (RH + GH * 16 + BH * 16∧2 ) * 4 * 8 + 6 = HADR (A) +6 ……… (Equation 2-7) HADR (H) = (RH + GH * 16 + BH * 16∧2) * 4 * 8 + 7 = HADR ( A) +7 ... (Equation 2-8) Similarly, the grid points “A to
Reference address HADR (N) of “H” (where N = A to
H) can also be shown using an expression that represents an increment.

Next, the principle of interpolation in cubic interpolation will be described with reference to the drawings. As shown in FIG. 6A, in the cube, when the conversion point indicated by the color image data output from the reading unit 14, that is, the RGB signal is “P”, the cube is converted with the conversion point P as a reference. It can be divided into eight rectangular parallelepipeds (eight cubes when the conversion point P is located at the center of the cube). Therefore, FIG.
As shown in (b), the volume of the obtained eight rectangular parallelepipeds is w
1 to w8, the interpolated value IPL (C) of the “C” component at the conversion point P can be expressed as follows. IPL (C) = (w1 * c8 + w2 * c7 + w3 * c6 + w4 * c5 + w5 * c4 + w6 * c3 + w7 * c2 + w8 * c1) / (w1 + w2 + w3 + w4 + w5 + w6 + w7 + w8) (Equation 3-1) In addition, “c1 to c8” are in the “C” component obtained by increment based on the reference address HADR (A) of the above (Equation 2-1). It is the grid point data of the grid points “A to H”.

Also, the "M" component at the conversion point P,
Interpolated values IPL (M), IPL of the “Y” component and the “K” component
Similarly, (Y) and IPL (K) can be shown as follows. IPL (M) = (w1 * m8 + w2 * m7 + w3 * m6 + w4 * m5 + w5 * m4 + w6 * m3 + w7 * m2 + w8 * m1) / (w1 + w2 + w3 + w4 + w5 + w6 + w7 + w8) ……… (Equation 3-2) IPL (Y) = (w1 * y8 + w2 * y7 + w3 * y6 + w4 * y5 + w5 * y4 + w6 * y3 + w7 * y2 + w8 * y1) / (w1 + w2 + w3 + w4 + w5 + w6 + w7 + w8) ……… (Equation 3-3) IPL (K) = (w1 * k8 + w2 * k7 + w3 * k6 + w4 * k5 + w5 * k4 + w6 * k3 + w7 * k2 + w8 * k1) / (w1 + w2 + w3 + w4 + w5 + w6 + w7 + w8) ……… (Equation 3-4)

Incidentally, "m1 to m8", "y1 to y8" and "k1 to k8" are also obtained by incrementing "M, Y," based on the reference address HADR (A) of (Formula 2-1). It is grid point data of grid points “A to H” in “K”. As described above, it is also clear from (Equation 3-1) to (Equation 3-4) showing the interpolated value IPL (N) of the “C, M, Y, K” component (where N = C, M, Y, K). As described above, the volumes w1 to w8 of the rectangular parallelepiped are the “weighting coefficient W” in the interpolation value IPL (N). Moreover, "C, M, Y, K"
The weighting factors W (w1 to w8) used when calculating the interpolated value IPL (N) of the component are all the same. In other words, the weighting factor W is independent of the "C, M, Y, K" components.

Therefore, the weighting factors W (w1 to w8) irrelevant to the "C, M, Y, K" components are continuously stored. That is, as shown in FIG. 7, color image data output from the reading unit 14, that is, RGB at the conversion point P.
If the lower 4 bits of the signal are (RL, GL, BL), (0, 0, 0), (1, 0, 0), (2, 0,
0, ..., (15,15,14), (15,15,
15) In the order of each grid point, the corresponding weighting factor W
(W1 to w8) are stored continuously.

Specifically, the weighting coefficient W (w1 to w8) is stored for (0, 0, 0). Then, (1,
The weighting coefficient W (w1 to w8) is stored for 0, 0). And finally, for (15,15,15),
The weighting factor W (w1 to w8) is stored. Therefore, the reference address LADR (w1) of the weighting factor w1 can be shown as follows. LADR (w1) = (RL + GL * 16 + BL * 16∧2) * 8 = (RL + GL * 16 + BL * 256) * 8 ……… (Equation 4-1) Weighting factors w2 to w8 Reference address LADR
(N) (however, N = w2-w8) can be shown as follows. LADR (w2) = (RL + GL * 16 + BL * 16∧2) * 8 + 1 = LADR (w1) +1 ……… (Equation 4-2) LADR (w3) = (RL + GL * 16 + BL * 16∧2) * 8 + 2 = LADR (w1) +2 ……… (Equation 4-3) LADR (w4) = (RL + GL * 16 + BL * 16∧2) * 8 + 3 = LADR (w1) +3 ……… (Equation 4-4)

LADR (w5) = (RL + GL * 16 + BL * 16∧2) * 8 + 4 = LADR (w1) +4 ... (Equation 4-5) LADR (w6) = (RL + GL * 16 + BL * 16∧2) * 8 + 5 = LADR (w1) +5 ……… (Equation 4-6) LADR (w7) = (RL + GL * 16 + BL * 16∧2) * 8 + 6 = LADR (w1) +6 ……… (Equation 4-7) LADR (w8) = (RL + GL * 16 + BL * 16∧2) * 8 + 7 = LADR (w1) +7 ……… ( Therefore, if stored in this manner, if the reference address LADR (w1) of the weighting factor w1 is known, all the weighting factors W (w1 to w8) can be sequentially obtained by incrementing by 7 times. it can. Similarly, the reference address LAD of the weighting factor W (w1 to w8) in “M, Y, K”.
R (N) (however, N = w1 to w8) can also be shown by the formula showing increment.

Next, the interpolation value I of the "C, M, Y, K" components
A method of calculating PL (N) (however, N = C, M, Y, K) will be described. When the color image data output from the reading unit 14, that is, the respective lower 4 [bits] of the RGB signal at the conversion point P is (RL, GL, BL),
Interpolated value IPL (N) of "C, M, Y, K" components (however,
N = C, M, Y, K) is expressed by (Formula 3-1) to (Formula 3-4).
Therefore, it can be shown as follows. IPL (N) = (w1 * f (HADR (A)) + w2 * f (HADR (B)) + w3 * f (HADR (C)) + w4 * f (HADR (D)) + w5 * f ( HADR (E)) + w6 * f (HADR (F)) + w7 * f (HADR (G)) + w8 * f (HADR (H))} / (16 * 16 * 16) ……… (Equation 5 )

Note that f (HADR shown in (Equation 5)
(N)) (where N = A to H) is obtained by incrementing “C, M,
It is the grid point data of the grid points "A to H" in the "Y, K" component.

Here, as shown in FIG. 6 (b),
The volumes of eight rectangular parallelepipeds based on the conversion point P, that is, the weighting factors W (w1 to w8) are set to the lower 4 [bi of the RGB signals.
t], that is, (RL, GL, BL), can be expressed as follows. w1 = (16-RL) * (16-GL) * (16-BL) ……… (Equation 6-1) w2 = (16-RL) * (16-GL) * BL ……… (Equation 6- 2) w3 = (16-RL) * GL * (16-BL) ……… (Equation 6-3) w4 = (16-RL) * GL * BL ……… (Equation 6-4) w5 = RL * (16-GL) * (16-BL) ……… (Equation 6-5) w6 = RL * (16-GL) * BL ……… (Equation 6-6) w7 = RL * GL * (16-BL ) ……… (Equation 6-7) w8 = RL * GL * BL ……… (Equation 6-8)

Therefore, in (Equation 5), (Equation 6-1)-
Substituting (Equation 6-8), the interpolated value IPL (N) of the “C, M, Y, K” component (where N = C, M, Y, K) is
It can be shown as follows. IPL (N) = ((16-RL) * (16-GL) * (16-BL) * f (HADR (A)) + (16-RL) * (16-GL) * BL * f (HADR ( B)) + (16-RL) * GL * (16-BL) * f (HADR (C)) + (16-RL) * GL * BL * f (HADR (D)) + RL * (16-GL ) * (16-BL) * f (HADR (E)) + RL * (16-GL) * BL * f (HADR (F)) + RL * GL * (16-BL) * f (HADR (G) ) + RL * GL * BL * f (HADR (H))} / (16 * 16 * 16) ……… (Equation 7) Note that f (HADR (N)) shown in (Equation 7) (however, N
= A to H) is the grid point data of the grid points "A to H" in the "C, M, Y, K" components obtained by the increment based on the (formula 2-1).

Next, the case where the conversion point P corresponds to exceptional processing in cubic interpolation will be described. In the RGB space shown in FIG. 2, each RGB signal is 8 [bi
t], each axis is "0 to 255 (00
h to FFh) ”. Therefore, each axis has 16
When divided, not all have the same spacing. That is, in each axis, the intersection between "0 to 240 (00h to F0h)" is increased by "16".
The portion "55 (F0h to FFh)" becomes an intersection when "15" is increased. Therefore, the size of the interpolation calculation is "2
40-255 (F0h-FFh) ".

Therefore, the conversion point P is "240-255 (F
0h to FFh) ”must be determined. However, when the interpolation calculation is configured by software, the interpolation calculation process is performed after determining whether the conversion point P corresponds to the exceptional process. Since the determination process of whether or not it corresponds to the exception process is performed for all pixels, even if the time required for the determination process for one pixel is a minute time, the determination process is performed for many pixels. If so, the determination process takes time. Therefore, as shown in FIG.
Consider the case of exception handling in a cube in GB space. In other words, the case where the exception processing is applied means that each upper 4 [bit] of the RGB signal at the conversion point P is (R
H, GH, BH), then (RH, GH, BH)
This is a case where any one of the values of is 15.

Therefore, when the conversion point P corresponds to exceptional processing, there are the following "7 ways". [1] RH = 15, GH <15, BH <15 (E1 shown in FIG. 8) [2] RH <15, GH = 15, BH <15 (E2 shown in FIG. 8) [3] RH <15, GH <15, BH = 15 (E3 shown in FIG. 8) [4] RH = 15, GH = 15, BH <15 (E4 shown in FIG. 8) [5] RH <15, GH = 15, BH = 15 (FIG. 8 E5) [6] RH = 15, GH <15, BH = 15 (E6 shown in FIG. 8) [7] RH = 15, GH = 15, BH = 15 (E7 shown in FIG. 8)

Therefore, the weighting factor W (w1 to w8) in the case of the exceptional processing of the above [1] to [7] is
From the above (formula 6-1) to (formula 6-8), it can be shown as follows. [1] In the case of RH = 15, GH <15, BH <15 (E1 shown in FIG. 8) w1 = (15-RL) * (16-GL) * (16-BL) ... (Equation 8-1 ) W2 = (15-RL) * (16-GL) * BL ……… (Equation 8-2) w3 = (15-RL) * GL * (16-BL) ……… (Equation 8-3) w4 = (15-RL) * GL * BL ……… (Equation 8-4) w5 = RL * (16-GL) * (16-BL) ……… (Equation 8-5) w6 = RL * (16- GL) * BL ……… (Equation 8-6) w7 = RL * GL * (16-BL) ……… (Equation 8-7) w8 = RL * GL * BL ……… (Equation 8-8)

[0045]   [2] In the case of RH <15, GH = 15, BH <15 (E2 shown in FIG. 8)   w1 = (16-RL) * (15-GL) * (16-BL) ……… (Equation 9-1)   w2 = (16-RL) * (15-GL) * BL ……… (Equation 9-2)   w3 = (16-RL) * GL * (16-BL) ……… (Equation 9-3)   w4 = (16-RL) * GL * BL ……… (Equation 9-4)   w5 = RL * (15-GL) * (16-BL) ……… (Equation 9-5)   w6 = RL * (15-GL) * BL ……… (Equation 9-6)   w7 = RL * GL * (16-BL) ……… (Equation 9-7)   w8 = RL * GL * BL ……… (Equation 9-8)

[0046]   [3] In case of RH <15, GH <15, BH = 15 (E3 shown in FIG. 8)   w1 = (16-RL) * (16-GL) * (15-BL) ……… (Equation 10-1)   w2 = (16-RL) * (16-GL) * BL ……… (Equation 10-2)   w3 = (16-RL) * GL * (15-BL) ……… (Equation 10-3)   w4 = (16-RL) * GL * BL ……… (Equation 10-4)   w5 = RL * (16-GL) * (15-BL) ……… (Equation 10-5)   w6 = RL * (16-GL) * BL ……… (Equation 10-6)   w7 = RL * GL * (15-BL) ……… (Equation 10-7)   w8 = RL * GL * BL ……… (Equation 10-8)

[0047]   [4] In the case of RH = 15, GH = 15, BH <15 (E4 shown in FIG. 8)   w1 = (15-RL) * (15-GL) * (16-BL) ……… (Equation 11-1)   w2 = (15-RL) * (15-GL) * BL ……… (Equation 11-2)   w3 = (15-RL) * GL * (16-BL) ……… (Equation 11-3)   w4 = (15-RL) * GL * BL ……… (Equation 11-4)   w5 = RL * (15-GL) * (16-BL) ……… (Equation 11-5)   w6 = RL * (15-GL) * BL ……… (Equation 11-6)   w7 = RL * GL * (16-BL) ……… (Equation 11-7)   w8 = RL * GL * BL ……… (Equation 11-8)

[0048]   [5] In the case of RH <15, GH = 15, BH = 15 (E5 shown in FIG. 8)   w1 = (16-RL) * (15-GL) * (15-BL) ……… (Equation 12-1)   w2 = (16-RL) * (15-GL) * BL ……… (Equation 12-2)   w3 = (16-RL) * GL * (15-BL) ……… (Equation 12-3)   w4 = (16-RL) * GL * BL ……… (Equation 12-4)   w5 = RL * (15-GL) * (15-BL) ……… (Equation 12-5)   w6 = RL * (15-GL) * BL ……… (Equation 12-6)   w7 = RL * GL * (15-BL) ……… (Equation 12-7)   w8 = RL * GL * BL ……… (Equation 12-8)

[0049]   [6] In the case of RH = 15, GH <15, BH = 15 (E6 shown in FIG. 8)   w1 = (15-RL) * (16-GL) * (15-BL) ……… (Equation 13-1)   w2 = (15-RL) * (16-GL) * BL ……… (Equation 13-2)   w3 = (15-RL) * GL * (15-BL) ……… (Equation 13-3)   w4 = (15-RL) * GL * BL ……… (Equation 13-4)   w5 = RL * (16-GL) * (15-BL) ……… (Equation 13-5)   w6 = RL * (16-GL) * BL ……… (Equation 13-6)   w7 = RL * GL * (15-BL) ……… (Equation 13-7)   w8 = RL * GL * BL ……… (Equation 13-8)

[0050]   [7] In the case of RH = 15, GH = 15, BH = 15 (E7 shown in FIG. 8)   w1 = (15-RL) * (15-GL) * (15-BL) ……… (Equation 14-1)   w2 = (15-RL) * (15-GL) * BL ……… (Equation 14-2)   w3 = (15-RL) * GL * (15-BL) ……… (Equation 14-3)   w4 = (15-RL) * GL * BL ……… (Equation 14-4)   w5 = RL * (15-GL) * (15-BL) ……… (Equation 14-5)   w6 = RL * (15-GL) * BL ……… (Equation 14-6)   w7 = RL * GL * (15-BL) ……… (Equation 14-7)   w8 = RL * GL * BL ……… (Equation 14-8)

Next, a method of determining whether or not the conversion point P corresponds to exceptional processing will be described. Each upper 4 [bit] of the RGB signal at the conversion point P is (RH, GH, B
H), the interpolation calculation process may be performed by determining whether any of the values (RH, GH, BH) is “15”. However, in order to perform the interpolation calculation processing at higher speed, and thus to improve the processing speed of the entire apparatus, the cases where the exception processing is applied to (RH, GH, BH) and the cases where the exception processing is not applied In order to determine that, a value of "0 to 7" is stored for (RH, GH, BH).

That is, as shown in FIG. 9, assuming that each of the upper 4 bits of the RGB signal is (RH, GH, BH) and the value indicating whether or not the exceptional processing is applied is "Z", In the case of [1] to [7] corresponding to the exception processing of
When the value of “Z” is set to “1 to 7,” it does not correspond to exception processing, that is, each value of (RH, GH, BH) is “1”.
If it is not “5”, in other words, if all the values of (RH, GH, BH) are “0 to 14”, the value of “Z” is stored as “0”. Therefore, whether or not the exception processing is applicable can be determined based on (RH, GH, BH) by the value “Z” indicating whether or not the exception processing is applied. In addition, it is possible to determine which of the "7 types" of exception processing shown in [1] to [7] corresponds to.

Therefore, the reference address ZHADR (Z) of "Z" indicating whether or not it corresponds to the exception processing is (RH, G
H, BH) can be used to show as follows. ZHADR (Z) = RH + GH * 16 + BH * 16∧2 = RH + GH * 16 + BH * 256 ……… (Equation 15)

Next, another method for determining whether or not the conversion point P corresponds to exceptional processing will be described. That is,
In the “7 ways” exception handling shown in [1] to [7],
A value “Z” indicating whether or not it corresponds to exception processing is respectively set to [1] RH = 15, GH <15, BH <15 Z = 1 [2] RH <15, GH = 15, BH <15 Z = 2 [3] RH <15, GH <15, BH = 15 Z = 4 [4] RH = 15, GH = 15, BH <15 Z = 3 [5] RH <15, GH = 15, BH = 15 Z = 6 [6] RH = 15, GH <15, BH = 15 Z = 5 [7] RH = 15, GH = 15, BH = 15 Z = 7 If defined in advance, the value “Z” indicating whether or not it corresponds to the exception processing can be indicated as follows. Z = (RH == 15) + (GH == 15) * 2 + (BH == 15) * 4 ……… (Equation 15-1) Here, “RH == 15” means RH = 15 In the case of, the rule is “1”, and in the case of RH ≠ 15, it is “0”. In addition,
“GH == 15” and “BH == 15” also have “RH ==
15 ”.

Next, the order of storing the weighting factors W (w1 to w8) in the case of exceptional processing will be described with reference to the drawings. As shown in FIG. 10, “Z” is a value indicating whether or not exception processing is performed, and color image data output from the reading unit 14, that is, each lower 4 [bit] of the RGB signal is (RL, GL). , BL) and the weighting factor is “W”, and corresponds to each Z (= 0 to 7) (RL, G
L, BL) are stored. Also, (RL, GL, BL)
The weighting factor W corresponding to is sequentially stored in the order of w1 to w8.

Specifically, the weighting coefficient W (w1 to w8) is stored for (0, 0, 0) when Z = “0”. Then, the weighting factor W (w1 to w8) is stored for (1, 0, 0). And finally, Z
The weighting factor W (w1 to w8) is stored for (15, 15, 15) in the case of “7”.

Therefore, the reference address ELADR (w1) of the weighting factor w1 including the case corresponding to the exceptional processing can be shown as follows. ELADR (w1) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 = (RL + GL * 16 + BL * 256) * 8 + Z * 4096 * 8 ...... (Equation 16-1) In addition, the weighting factors w2 to w including the case corresponding to exception processing
8 reference address ELADR (N) (where N = w2−
w8) can be shown as follows. ELADR (w2) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 + 1 = ELADR (w1) +1 ……… (Equation 16-2) ELADR (w3) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 + 2 = ELADR (w1) +2 ……… (Equation 16-3) ELADR (w4) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 + 3 = ELADR (w1) +3 ……… (Equation 16-4)

ELADR (w5) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 + 4 = ELADR (w1) +4 ……… (Equation 16-5) ELADR (w6) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 + 5 = ELADR (w1) +5 ……… (Equation 16-6) ELADR (w7) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 + 6 = ELADR (w1) +6 ……… (Equation 16-7) ELADR (w8) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 8 + 7 = ELADR (w1) +7 (Equation 16-8) Therefore, if stored in this way, the weighting factor w1 Based on the reference address ELADR (w1), all the weighting factors W (w1 to w8) can be sequentially obtained only by incrementing seven times.

Next, in the exception processing, "C, M, Y,
Interpolated value IPL (N) of the “K” component (where N = C, M,
A method of calculating Y, K) will be described. Reading unit 14
If the lower 4 [bits] of the RGB image data, that is, the lower 4 bits of the RGB signal are (RL, GL, BL), the interpolated value IPL (N) of the “C, M, Y, K” components (however, N = C, M, Y, K) is similar to the above (Equation 5),
It can be shown as follows. IPL (N) = (w1 * f (HADR (A)) + w2 * f (HADR (B)) + w3 * f (HADR (C)) + w4 * f (HADR (D)) + w5 * f ( HADR (E)) + w6 * f (HADR (F)) + w7 * f (HADR (G)) + w8 * f (HADR (H))} / (16 * 16 * 16) ……… (Equation 17 )

Note that f (HADR shown in (Expression 17) is
(N)) (where N = A to H) is obtained by incrementing “C, M,
It is the grid point data of the grid points "A to H" in the "Y, K" component.

Here, the weighting factor W (w1 to w8) is RG
Each lower 4 [bit] of the B signal, that is, (RL, GL,
When expressed by BL), (Equation 8-1) to (Equation 14-)
From 8), interpolation value IPL (N) (where N = C, M,
The above (formula 17) indicating Y, K) can be expressed as follows. [1] When RH = 15, GH <15, BH <15 IPL (N) = {(15-RL) * (16-GL) * (16-BL) * f (HADR (A)) + (15 -RL) * (16-GL) * BL * f (HADR (B)) + (15-RL) * GL * (16-BL) * f (HADR (C)) + (15-RL) * GL * BL * f (HADR (D)) + RL * (16-GL) * (16-BL) * f (HADR (E)) + RL * (16-GL) * BL * f (HADR (F)) + RL * GL * (16-BL) * f (HADR (G)) + RL * GL * BL * f (HADR (H))} / (15 * 16 * 16) ……… (Equation 18) Note that ( Formula 18) can also be obtained by substituting “16 * RL / 15” for “RL” shown in (Formula 7).

[2] When RH <15, GH = 15, BH <15: IPL (N) = {(16-RL) * (15-GL) * (16-BL) * f (HADR (A)) + (16-RL) * (15-GL) * BL * f (HADR (B)) + (16-RL) * GL * (16-BL) * f (HADR (C)) + (16-RL) * GL * BL * f (HADR (D)) + RL * (15-GL) * (16-BL) * f (HADR (E)) + RL * (15-GL) * BL * f (HADR (F )) + RL * GL * (16-BL) * f (HADR (G)) + RL * GL * BL * f (HADR (H))} / (16 * 15 * 16) ……… (Equation 19) Note that (Equation 19) can also be obtained by substituting “16 * GL / 15” into “GL” shown in (Equation 7).

[3] When RH <15, GH <15, BH = 15: IPL (N) = {(16-RL) * (16-GL) * (15-BL) * f (HADR (A)) + (16-RL) * (16-GL) * BL * f (HADR (B)) + (16-RL) * GL * (15-BL) * f (HADR (C)) + (16-RL) * GL * BL * f (HADR (D)) + RL * (16-GL) * (15-BL) * f (HADR (E)) + RL * (16-GL) * BL * f (HADR (F )) + RL * GL * (15-BL) * f (HADR (G)) + RL * GL * BL * f (HADR (H))} / (16 * 16 * 15) ……… (Equation 20) Note that (Equation 20) can also be obtained by substituting “16 * BL / 15” into “BL” shown in (Equation 7).

[4] When RH = 15, GH = 15, BH <15 IPL (N) = {(15-RL) * (15-GL) * (16-BL) * f (HADR (A)) + (15-RL) * (15-GL) * BL * f (HADR (B)) + (15-RL) * GL * (16-BL) * f (HADR (C)) + (15-RL) * GL * BL * f (HADR (D)) + RL * (15-GL) * (16-BL) * f (HADR (E)) + RL * (15-GL) * BL * f (HADR (F )) + RL * GL * (16-BL) * f (HADR (G)) + RL * GL * BL * f (HADR (H))} / (15 * 15 * 16) ……… (Equation 21) In addition, in (Expression 21), "RL" shown in (Expression 7) is "16 * RL / 15", and "GL" is "16 * GL / 1".
5 ”can also be obtained by substituting each of them.

[5] When RH <15, GH = 15, BH = 15 IPL (N) = {(16-RL) * (15-GL) * (15-BL) * f (HADR (A)) + (16-RL) * (15-GL) * BL * f (HADR (B)) + (16-RL) * GL * (15-BL) * f (HADR (C)) + (16-RL) * GL * BL * f (HADR (D)) + RL * (15-GL) * (15-BL) * f (HADR (E)) + RL * (15-GL) * BL * f (HADR (F )) + RL * GL * (15-BL) * f (HADR (G)) + RL * GL * BL * f (HADR (H))} / (16 * 15 * 15) ……… (Equation 22) In addition, in (Expression 22), "16 * GL / 15" is added to "GL" and "16 * BL / 1" is added to "BL" shown in (Expression 7).
5 ”can also be obtained by substituting each of them.

[6] When RH = 15, GH <15, BH = 15 IPL (N) = {(15-RL) * (16-GL) * (15-BL) * f (HADR (A)) + (15-RL) * (16-GL) * BL * f (HADR (B)) + (15-RL) * GL * (15-BL) * f (HADR (C)) + (15-RL) * GL * BL * f (HADR (D)) + RL * (16-GL) * (15-BL) * f (HADR (E)) + RL * (16-GL) * BL * f (HADR (F )) + RL * GL * (15-BL) * f (HADR (G)) + RL * GL * BL * f (HADR (H))} / (15 * 16 * 15) ……… (Equation 23) In addition, (Expression 23) is changed to "16" in "RL" shown in (Expression 7).
* RL / 15 "can also be obtained by substituting" 16 * BL / 15 "for" BL ".

[7] When RH = 15, GH = 15, BH = 15 IPL (N) = {(15-RL) * (15-GL) * (15-BL) * f (HADR (A)) + (15-RL) * (15-GL) * BL * f (HADR (B)) + (15-RL) * GL * (15-BL) * f (HADR (C)) + (15-RL) * GL * BL * f (HADR (D)) + RL * (15-GL) * (15-BL) * f (HADR (E)) + RL * (15-GL) * BL * f (HADR (F )) + RL * GL * (15-BL) * f (HADR (G)) + RL * GL * BL * f (HADR (H))} / (15 * 15 * 15) ……… (Equation 24) In addition, in (Formula 24), "16 * RL / 15" is added to "RL" and "16 * GL / 1" is added to "GL" shown in (Formula 7).
5 ”can also be obtained by substituting“ 16 * BL / 15 ”into“ BL ”.

In addition, f shown in (Equation 18) to (Equation 24)
(HADR (N)) (where N = A to H) is the value of the grid point "A to H" in the "C, M, Y, K" components obtained by increment based on the above (Equation 2-1). This is grid point data. Therefore, as shown in the LUT of the weighting coefficient including the case corresponding to the exceptional processing of FIG. 10, [1] to
According to the “7 ways” exception handling shown in [7], “RL,
"GL, BL" to "16 * RL / 15, 16 * G" respectively
L / 15, 16 * BL / 15 ”substituting weighting factor W
When (w1 to w8) is stored, (Equation 18) to (Equation 2)
Without performing the calculation of 4), from the above (Formula 7), “C,
Interpolated value IPL (N) of the “M, Y, K” component (where N =
C, M, Y, K) can be calculated.

Next, in the facsimile apparatus 1, 1 when the RGB signal of the conversion point P is output from the reading unit 14.
The operation of color conversion processing for pixels will be described with reference to the flowchart shown in FIG. In addition, this operation is RO
It is executed under the control of the MPU 11 based on the program stored in M12.

In step S1 shown in FIG. 11, each lower 4 bits, that is, (RL, GL, BL) is extracted from the RGB signal at the conversion point P. In step S2, (RL,
GL, BL), the reference address LADR (w1) of the weighting factor w1 is calculated. Specifically, the reference address LA of the weighting factor w1 is calculated based on the above (formula 4-1).
DR (w1) is calculated.

In step S3, the weighting factors W (w1 to w8) in the cube surrounding the conversion point P are stored in the ROM.
It is sequentially read from 12. Specifically, the reference address L of the weighting factor w1 calculated in step S2
The weighting factors W (w1 to w8) are sequentially incremented from ADR (w1), and are sequentially read from the LUT of the weighting factors W (w1 to w8) shown in FIG. Note that the weighting factors W (w1 to w8) may be calculated based on (RL, GL, BL) extracted in step S1 instead of the processing in steps S2 and S3. Specifically, the weighting factor W (w1 to w8) may be calculated based on the above (Equation 6-1) to (Equation 6-8).

In step S4, the RG of the conversion point P
Each upper 4 bits from the B signal, that is, (RH, G
H, BH) is extracted. In step S5, (RH, GH, B extracted in step S4 is extracted.
Based on H), the reference address HADR (A) of the grid point “A” in the “C” component is calculated. In particular,
The reference address HADR (A) of the grid point “A” in the “C” component is calculated based on the above (formula 2-1).

In step S6, the grid point data of grid points "A to H" in the cube surrounding the conversion point P are sequentially read from the ROM 12. Specifically, the grid points "A to H" in the "C" component are sequentially incremented from the reference address HADR (A) of the grid point "A" in the "C" component calculated in step S5.
Is the LU of the grid point data shown in FIG.
The data is sequentially read from T.

In step S7, interpolation calculation at the conversion point P is performed. Specifically, the step S3
Based on the weighting factor W (w1 to w8) read in step S6 and the grid point data of grid points “A to H” in the “C” component read in step S6, The interpolation calculation at the conversion point P is performed.

In step S8, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component. If there is a next component, the process proceeds to step S9. On the other hand, when there is no next component, that is, for all "C, M,
When the interpolation calculation is completed for the “Y, K” components,
This process ends.

In step S9, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 5, since the grid point data of the grid points “A to H” for the “C, M, Y, K” components are stored, the calculation is performed in step S5. "C"
Reference address HADR of grid point "A" in the component
It only increments (A) sequentially. And
Steps S6 to S9 are repeatedly executed to perform interpolation calculation of "M, Y, K" at the conversion point P.
That is, steps S6 to S9 are repeatedly executed to perform interpolation calculation of the "C, M, Y, K" components at the conversion point P, and the RGB signal at the conversion point P is converted into a CMYK signal. Note that generally, since there are many pixels to be converted, this processing is repeatedly executed for all the read pixels.

As described above, according to the first embodiment, the following actions and effects can be obtained. (1) As shown in FIG. 5, in order of (RH, GH, BH), "C, M, Y, K" components are arranged in the order of "C, M, Y,
The grid point data is stored in the ROM 12 in the order of grid points “A to H” for the “K” component. In other words, the grid point data of the grid points “A to H” in the cube are continuously stored. Therefore, if the reference address HADR (A) of the lattice point “A” in the cube surrounding the conversion point P is known based on the above (Equation 2-1), it is simply incremented sequentially and the other lattice points “B to H” are obtained. It is possible to obtain grid point data of ". As a result, the lattice point “A
It is possible to quickly obtain the grid point data of "~ H".

Further, as shown in FIG. 7, (RL, GL,
BL) in the order of the weighting factor W from w1 to w8, ROM
It is stored in 12. Therefore, the reference address LADR (w1) of the weighting factor w1 is calculated based on the above (formula 4-1).
, The other weighting factors w2 to w8 can be obtained only by sequentially incrementing. As a result, the weighting factor W (w1 to w8) can also be obtained quickly. Therefore, in the case of performing color conversion processing from RGB signals to CMYK signals, it is possible to perform color conversion processing at high speed even with software.

(2) Since the color conversion processing is performed using software, a dedicated LSI is unnecessary. Moreover, because it is not configured using hardware, the LSI
It is possible to easily reduce the size and reduce the cost. In addition, since a dedicated circuit is unnecessary, various interpolation calculation methods can be supported.

(3) In the flowchart shown in FIG. 11, the weighting factor W (w1 to w8) is first read.
This is because the weighting factor W (w1 to w8) is "C, M, Y,
This is because it is the same regardless of the “K” component. Therefore, the interpolation calculation process can be performed quickly. Therefore, the color conversion process can be performed at high speed.

(4) As is clear from (Equation 5), the product-sum operation is performed 8 times in the interpolation operation process. However, depending on the MPU, there is one that processes an instruction for repeatedly performing the product-sum operation at a higher speed.
Therefore, color conversion processing can be performed at high speed by using an MPU having an instruction for repeatedly performing the sum-of-products calculation.

(5) In the first embodiment, even when the conversion point P corresponds to an exceptional process, the color conversion process is performed using the normal weighting coefficient W. Therefore, when the conversion point P corresponds to the exceptional processing, the value after the color conversion processing includes some error. However, it is extremely rare that the conversion point P corresponds to exceptional processing. Therefore, the color conversion processing can be made even faster. Moreover, since the conversion point P rarely corresponds to exceptional processing, the error does not significantly affect the entire image on which the color conversion processing is performed.

[Second Embodiment] ... In the case of cubic interpolation (8-point interpolation) In the second embodiment, the same configuration as that of the first embodiment is the same as that of the first embodiment. The reference numerals are given and the description is omitted. The second embodiment takes into consideration the case corresponding to the exception processing in the first embodiment. Therefore, in the second embodiment, the ROM 12 stores the LUT of grid point data shown in FIG.
The LUT for determining whether the conversion point P shown in FIG. 9 corresponds to an exceptional process and the LUT of the weighting coefficient W (w1 to w8) including the case where it corresponds to the exceptional process shown in FIG. 10 are stored.

Next, in the facsimile apparatus 1, 1 when the RGB signal of the conversion point P is output from the reading unit 14.
Regarding the operation of color conversion processing for pixels, FIG. 12 and FIG.
This will be described with reference to the flowchart shown in. Note that this operation is based on a program stored in the ROM 12,
It is executed under the control of the PU 11.

In step S11 shown in FIG. 12,
From the RGB signal at the conversion point P, each upper 4 [bit], that is, (RH, GH, BH) is extracted. Step S12
In step S11, it is determined whether the conversion point P corresponds to an exceptional process, based on (RH, GH, BH) extracted in step S11. Specifically, the reference address ZHADR (Z) of “Z” indicating whether or not the exception processing is performed is calculated based on the above (Equation 15).
Then, the value of "Z" is extracted based on the reference address ZHADR (Z). Specifically, it is determined whether or not the conversion point P corresponds to the exceptional process based on the LUT for determining whether or not the conversion point P shown in FIG. 9 corresponds to the exceptional process. That is, when the value of “Z” indicating whether or not the exception processing is applied is “1 to 7”, it is determined that the exception processing is applied, and the process proceeds to step S31 shown in FIG. On the other hand, when the value of “Z” indicating whether or not the exception processing is applied is “0”, it is determined that the exception processing is not applied,
Control goes to step S13. In addition, the above (formula 15-1)
It may be possible to determine whether or not it corresponds to the exception processing based on.

In step S13, R of the conversion point P
Each lower 4 bits from the GB signal, that is, (RL, G
L, BL) are extracted. In step S14,
(RL, GL, extracted in step S13)
Based on BL), the reference address ELADR (w1) of the weighting factor w1 including the case corresponding to the exceptional process is calculated. Specifically, the reference address ELADR (w1) of the weighting factor w1 is calculated based on the above (Equation 16-1).

In step S15, the weighting coefficient W (w1 to w8) in the cube surrounding the conversion point P is RO.
The data is sequentially read from M12. Specifically, the reference address ELADR (w1) of the weighting factor w1 calculated in step S14 is sequentially incremented,
Weighting factors W (w1 to w8) including the case where all the weighting factors W (w1 to w8) correspond to the exceptional processing shown in FIG.
It is sequentially read from the LUT of 8). More specifically,
In the LUT of the weighting coefficient W (w1 to w8) including the case corresponding to the exceptional process shown in FIG. 10, “Z (= (Eq.15)” calculated based on the (Formula 15) in the process of step S12 shown in FIG. 0) ”weighting coefficient W according to the value
(W1 to w8) are sequentially read.

Incidentally, instead of the processing of steps S14 and S15, (R
Based on L, GL, BL), the weighting factor W (w1 to w
8) may be calculated. Specifically, the above (formula 6-1)
~ Based on (Equation 6-8), the weighting factor W (w1 to w8)
May be calculated.

In step S16, R of the conversion point P
Each upper 4 bits from the GB signal, that is, (RH, G
H, BH) is extracted. In addition, this upper 4 [bit]
Since the extraction processing of step S11 is the same as that of step S11, the top 4 [b
It] may be used as it is.

In step S17, the reference address HADR (A) of the lattice point "A" in the "C" component is calculated based on (RH, GH, BH) extracted in step S16. Specifically, the above (formula 2
The reference address HADR (A) of the grid point “A” in the “C” component is calculated based on −1).

The reference address HADR (A) of the lattice point "A" in the "C" component is the reference address ZHADR of "Z" calculated in step S12.
(Z), that is, the reference address ZHADR of "Z" indicating whether or not the exception processing shown in (Equation 15) is satisfied.
The reference address HADR (A) of the lattice point “A” in the “C” component may be calculated by multiplying (Z) by “32”.

In step S18, the grid point data of grid points "A to H" in the cube surrounding the conversion point P are sequentially read from the ROM 12. Specifically, it is sequentially incremented from the reference address HADR (A) of the grid point "A" in the "C" component calculated in step S17, and the grid point "A ~" in the "C" component is sequentially incremented.
The grid point data of "H" is sequentially read from the LUT of the grid point data shown in FIG.

In step S19, interpolation calculation at the conversion point P is performed. Specifically, the step S
Weighting coefficient W (w1 to w8) read in 15
And the grid point data of the grid points “A to H” in the “C” component read in step S18, the interpolation calculation at the conversion point P is performed by the above (formula 5).

In step S20, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component.
If there is a next component, the process proceeds to step S21. On the other hand, when there is no next component, that is, for all "C,
When the interpolation calculation is completed for the “M, Y, K” components, this process is completed.

In step S21, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 5, since the grid point data of the grid points “A to H” for the “C, M, Y, K” components are stored, the calculation is performed in step S7. "C"
Reference address HADR of grid point "A" in the component
It only increments (A) sequentially. And
Steps S18 to S21 are repeatedly executed, and the interpolation calculation of "M, Y, K" at the conversion point P is performed. That is, steps S18 to S21 are repeatedly executed, and “C, M, Y,
The K ”component is interpolated, and the RGB signal at the conversion point P is converted into a CMYK signal.

In step S31 shown in FIG. 13,
Each lower 4 bits, that is, (RL, GL, BL) is extracted from the RGB signal at the conversion point P. Step S32
, The reference address ELADR (w of the weighting factor w1 including the case corresponding to the exception processing is based on (RL, GL, BL) extracted in step S31.
1) is calculated. Specifically, based on (Equation 16-1), the reference address ELADR (w of the weighting factor w1
1) is calculated.

In step S33, the weighting factor W (w1 to w8) in the cube surrounding the conversion point P is RO.
The data is sequentially read from M12. Specifically, the weighting coefficient w1 calculated in step S32 is sequentially incremented from the reference address ELADR (w1),
Weighting factors W (w1 to w8) including the case where all the weighting factors W (w1 to w8) correspond to the exceptional processing shown in FIG.
It is sequentially read from the LUT of 8). More specifically,
In the LUT of the weighting coefficient W (w1 to w8) including the case corresponding to the exceptional process shown in FIG. 10, “Z (= (Eq.15)” calculated based on the (Formula 15) in the process of step S12 shown in FIG. 1 to 7) ”, the weighting factors W (w1 to w8) are sequentially read.

Note that, instead of the processing of steps S32 and S33, a value "Z" indicating whether or not the exceptional processing calculated in the processing of step S12 shown in FIG.
The weighting factor W (w1 to w8) may be calculated according to Specifically, depending on the case where the value “Z” indicating whether or not the exception processing is applicable is “1 to 7,” in other words, (R
H, GH, BH) depending on the above (formula 8-1 to 8) ~
The weighting coefficient W (w1 to w8) in the case of exception processing may be calculated by (Equations 14-1 to 8).

In step S34, R of the conversion point P
Each upper 4 bits from the GB signal, that is, (RH, G
H, BH) is extracted. In addition, this upper 4 [bit]
Since the extraction processing of step S11 is the same as the step S11 shown in FIG. 12, the upper 4 bits extracted in step S11 shown in FIG. 12 may be used as they are.

In step S35, the reference address HADR (A) of the lattice point "A" in the "C" component is calculated based on (RH, GH, BH) extracted in step S34. Specifically, the above (formula 2
The reference address HADR (A) of the grid point “A” in the “C” component is calculated based on −1).

The reference address HADR (A) of the lattice point "A" in the "C" component is the reference address ZHADR of "Z" calculated in step S12.
(Z), that is, the reference address ZHADR of "Z" indicating whether or not the exception processing shown in (Equation 15) is satisfied.
The reference address HADR (A) of the lattice point “A” in the “C” component may be calculated by multiplying (Z) by “32”.

In step S36, the grid point data of grid points "A to H" in the cube surrounding the conversion point P are sequentially read from the ROM 12. Specifically, the reference address HADR (A) of the grid point "A" in the "C" component calculated in step S35 is sequentially incremented to obtain the grid points "A ~" in the "C" component.
The grid point data of "H" is sequentially read from the LUT of the grid point data shown in FIG.

In step S37, interpolation calculation at the conversion point P is performed. Specifically, the step S
Weighting coefficient W (w1 to w8) read in 33
And the grid point data of the grid points “A to H” in the “C” component read in step S36, the interpolation calculation at the conversion point P is performed by the above (formula 5).

In step S38, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component.
If there is a next component, the process proceeds to step S39. On the other hand, when there is no next component, that is, for all "C,
When the interpolation calculation is completed for the “M, Y, K” components, this process is completed.

In step S39, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 5, since the grid point data of the grid points “A to H” are stored for the “C, M, Y, K” components, the calculation is performed in step S35. Address HAD of grid point "A" in the extracted "C" component
It only increments R (A) sequentially. Then, steps S36 to S39 are repeatedly executed, and the interpolation calculation of "M, Y, K" at the conversion point P is performed. That is, steps S36 to S39 are repeatedly executed, and "C, M, Y,
The K ”component is interpolated, and the RGB signal at the conversion point P is converted into a CMYK signal. Note that generally, since there are many pixels to be converted, this processing is repeatedly executed for all the read pixels.

As described above in detail, according to the second embodiment, the following operation and effect can be obtained in addition to the operation and effect of the first embodiment. (1) In the second embodiment, as shown in FIG. 10, the weighting coefficient W including the case corresponding to the exceptional process is w1.
The data are stored in the ROM 12 in the order of to w8. Therefore, the reference address ELADR of the weighting factor w1 including the case corresponding to the exceptional process is calculated based on the above (formula 16-1).
If you know (w1), you can simply increment it,
The weighting factors w2 to w8 can also be obtained. Therefore, even when the conversion point P corresponds to the exceptional process, the color conversion process can be performed at high speed even by using software. Moreover, even if the conversion point P corresponds to the exception processing,
Color conversion processing can be performed without including an error.

(2) When determining whether or not the exception processing is applicable, the exception processing is not performed based on whether or not each value of (RH, GH, BH) is "15". The value of “Z” indicating whether or not
The decision is made based on 5). Therefore, (RH,
It is possible to increase the processing speed of the entire apparatus as compared with the processing of determining whether or not any of the values of (GH, BH) is “15”. Moreover, depending on the value of "Z",
It is also possible to determine which exception processing is applicable. Therefore, even if the conversion point P corresponds to the exceptional process, the color conversion process can be performed at a higher speed.

[Third Embodiment] ... In the case of cubic interpolation (8-point interpolation) In this third embodiment, the same configuration as that of the first embodiment is the same as that of the first embodiment. The reference numerals are given and the description is omitted. The third embodiment is similar to the second embodiment in consideration of the exception processing in the first embodiment. In addition, the third embodiment is different from the second embodiment in that the color conversion processing is executed without making a determination as to whether or not the exception processing is applicable.

In the third embodiment, the ROM 12
Is the LUT of the grid point data shown in FIG. 5, and the LU for determining whether the conversion point P shown in FIG. 9 corresponds to the exceptional processing.
T and the LUT of the weighting factors W (w1 to w8) including the case corresponding to the exceptional processing shown in FIG. 10 are stored.

Next, in the facsimile apparatus 1, 1 when the RGB signal of the conversion point P is output from the reading unit 14
The operation of color conversion processing for pixels will be described with reference to the flowchart shown in FIG. In addition, this operation is RO
It is executed under the control of the MPU 11 based on the program stored in M12.

In step S41 shown in FIG. 14,
From the RGB signal at the conversion point P, each upper 4 [bit], that is, (RH, GH, BH) is extracted. Step S42
In, the reference address ZHADR (Z) of “Z” indicating whether or not the exception processing is performed is calculated based on (RH, GH, BH) extracted in step S41. Specifically, based on the above (Equation 15), the reference address Z of "Z" indicating whether or not it corresponds to exception processing.
HADR (Z) is calculated.

At step S43, the reference address ZHA of "Z" calculated at step S42.
The value of “Z” is extracted based on DR (Z). Specifically, the value of “Z” is extracted based on the LUT for determining whether or not the conversion point P shown in FIG. 9 corresponds to exceptional processing. Therefore, when the value of "Z" is "0",
It can be determined that the conversion point P does not correspond to exception processing. When the value of "Z" is "1 to 7," it can be determined that the conversion point P corresponds to exceptional processing, and
It is also possible to determine which of the seven exceptions [1] to [7] corresponds to the exception handling.

In step S44, R of the conversion point P
Each lower 4 bits from the GB signal, that is, (RL, G
L, BL) are extracted. In step S45,
Extracted in step S44 (RL, GL,
Based on BL), the reference address ELADR (w1) of the weighting factor w1 including the case corresponding to the exceptional process is calculated. Specifically, the reference address ELADR (w1) of the weighting factor w1 is calculated based on the above (Equation 16-1).

In step S46, the weighting factor W (w1 to w8) in the cube surrounding the conversion point P is RO.
The data is sequentially read from M12. Specifically, the weighting coefficient w1 calculated in step S45 is sequentially incremented from the reference address ELADR (w1),
Weighting factors W (w1 to w8) including the case where all the weighting factors W (w1 to w8) correspond to the exceptional processing shown in FIG.
It is sequentially read from the LUT of 8). More specifically,
In the weighting factor W (w1 to w8) including the case corresponding to the exceptional process shown in FIG. 10, the weighting factor W corresponding to “Z (= 0 to 7)” extracted in step S43.
(W1 to w8) is read.

In step S47, R of the conversion point P
Each upper 4 bits from the GB signal, that is, (RH, G
H, BH) is extracted. In addition, this upper 4 [bit]
Since the extraction processing of step S41 is similar to that of step S41, the top 4 [b
It] may be used as it is.

In step S48, the reference address HADR (A) of the grid point "A" in the "C" component is calculated based on (RH, GH, BH) extracted in step S47. Specifically, the above (formula 2
The reference address HADR (A) of the grid point “A” in the “C” component is calculated based on −1).

The reference address HADR (A) of the lattice point "A" in the "C" component is the reference address ZHADR of "Z" calculated in step S42.
(Z), that is, the reference address ZHADR of "Z" indicating whether or not the exception processing shown in (Equation 15) is satisfied.
The reference address HADR (A) of the lattice point “A” in the “C” component may be calculated by multiplying (Z) by “32”.

In step S49, the grid point data of grid points "A to H" in the cube surrounding the conversion point P are sequentially read from the ROM 12. Specifically, the reference address HADR (A) of the grid point "A" in the "C" component calculated in step S48 is sequentially incremented to obtain the grid point "A ~" in the "C" component.
The grid point data of "H" is the grid points "A to H" shown in FIG.
Are sequentially read from the LUT of the grid point data.

In step S50, the interpolation calculation at the conversion point P is performed. Specifically, the step S
Weighting coefficient W (w1 to w8) read at 46
And the grid point data of the grid points “A to H” in the “C” component read in step S49, the interpolation calculation at the conversion point P is performed by the above (formula 5).

In step S51, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component.
If there is the next component, the process proceeds to step S52. On the other hand, when there is no next component, that is, for all "C,
When the interpolation calculation is completed for the “M, Y, K” components, this process is completed.

In step S52, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 5, since the grid point data of the grid points “A to H” for the “C, M, Y, K” components are stored, the calculation is performed in step S48. Address HAD of grid point "A" in the extracted "C" component
It only increments R (A) sequentially. Then, steps S49 to S52 are repeatedly executed, and the interpolation calculation of "M, Y, K" at the conversion point P is performed. That is, steps S49 to S52 are repeatedly executed, and “C, M, Y,
The K ”component is interpolated, and the RGB signal at the conversion point P is converted into a CMYK signal. Note that generally, since there are many pixels to be converted, this processing is repeatedly executed for all the read pixels.

As described in detail above, according to the third embodiment, the following operation and effect can be obtained in addition to the operation and effect of the second embodiment. (1) The interpolation calculation is performed without determining whether the conversion point P corresponds to exceptional processing. Therefore, unlike the second embodiment (flowcharts shown in FIGS. 12 and 13), the same processing (flowchart shown in FIG. 14) is performed regardless of whether the conversion point P corresponds to exception processing. Color conversion processing can be performed. Therefore, even when the color conversion processing is performed by software, the processing can be performed at a higher speed. In addition, even if the color conversion process is configured by hardware, the color conversion process can be performed by the same process, so that the configuration can be simplified.

[Fourth Embodiment] ... Triangular Pyramid Interpolation (Four-Point Interpolation) Next, a fourth embodiment in which the present invention is embodied in a facsimile machine will be described with reference to the drawings. In the fourth embodiment, the same components as those in the first embodiment are designated by the same reference numerals as those in the first embodiment, and the description thereof will be omitted.

In the fourth embodiment, the ROM1
The LUT data stored in 2 is different. That is, data required for triangular pyramid interpolation (four-point interpolation) in which a cube is divided into six is stored. Specifically, the LUT of weighting factors W (w1 to w4) shown in FIG. 35 described later, and (R
L, GL, BL) LUT indicating the offset determined by the magnitude relationship of each value and LU of the grid point data shown in FIG.
Remember T.

Next, the principle of triangular pyramid interpolation (four-point interpolation) will be described with reference to the drawings. As shown in FIG. 15A, in the RGB space, each lower 4 [bi of RGB signals is
t] is (RL, GL, BL), and the cube is “RL = G
"L plane", "GL = BL plane" and "BL = RL plane"
When divided by, the cube is 6 as shown in FIG.
It is divided into three triangular pyramids.

As shown in FIG. 16, when a cube is divided into six triangular pyramids, one of the six triangular pyramids with the conversion point P among the six triangular pyramids is based on the magnitude relation of each value of (RL, GL, BL). It is possible to determine whether or not it exists in the triangular pyramid.

That is, in FIG. 16A, BL>GL> R
In the case of L, FIG. 16B shows the case of GL>RL> BL, and FIG. 16C shows the case of RL>BL> GL.
16D shows a case of GL ≧ BL ≧ RL (where RL ≠ GL), and FIG. 16E shows a case of RL ≧ GL ≧ BL.
(F) shows the case of BL ≧ RL ≧ GL (where GL ≠ BL). Note that GL = in FIG.
Including the planes of BL and BL = RL, RL = in FIG.
It includes a plane of GL and GL = BL and a line of RL = GL = BL, and FIG. 16 (f) includes a plane of BL = RL and RL = GL.

Here, FIGS. 16A to 16F are in contact with two faces of another triangular pyramid. Therefore, the calculation result is the same regardless of which triangular pyramid the boundary surface is included in. Therefore,
In the present embodiment, these boundary surfaces and boundaries are shown in FIG.
It is described as being included in (d) to (f). It can be considered as included in FIGS. 16 (a) to 16 (c).

In addition, the weighting coefficient W (w1 to w4) is calculated based on the magnitude relation of the respective values of (RL, GL, BL). In other words, assuming that the values of (RL, GL, BL) are large in the order of "LL, LM, LS", the weighting factor W (w1 ~ w1
w4) can be shown as follows. w1 = 16-LL ……… (Equation 30-1) w2 = LL-LM ……… (Equation 30-2) w3 = LM-LS ……… (Equation 30-3) w4 = LS ……… (Equation 30-4)

Therefore, the weighting factors W (w1 to w4) in the cases of FIGS. 16 (a) to 16 (f) are calculated. [A] Weighting coefficient W in the case of BL>GL> RL shown in FIG. 16 (a) FIG. 17 shows a triangle in the case of FIG. 16 (a), that is, when the conversion point P is BL>GL> RL. Shows a cone.

As shown in FIGS. 18A to 18D, the triangular pyramid shown in FIG. 17 is further divided into four triangular pyramids with the conversion point P as a reference, and the volumes Va to Vd
To calculate. First, the volume V of the triangular pyramid shown in FIG.
a can be shown as follows. Va = 1/3 * (16 * 16) / 2 * (16-BL) = 1/6 * 16 * 16 * (16-BL) ……… (Equation 35-1)

Next, the triangular pyramid shown in FIG.
9 (a). When the triangular pyramid shown in FIG. 19 (a) is viewed from above, it can be shown as shown in FIG. 19 (b). Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 16 * SQRT (2)) / 2 * (GL-RL) / SQRT (2) = 1/6 * 16 * 16 * (GL-RL) ……… (Equation 35- 2) “SQRT” indicates a square root. same as below.

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 16) / 2 * RL = 1/6 * 16 * 16 * RL ……… (Equation 35-3)

From the equations (35-1) to (35-3), the volume Vb of the triangular pyramid shown in FIG. 18 (b) can be expressed as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 16 * 16 * {16- (16-BL)-(GL-RL) -RL} = 1/6 * 16 * 16 * (BL- GL) ……… (Equation 35-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 18A to 18D can be shown as follows. Va: Vb: Vc: Vd = 16-BL: BL-GL: GL-RL: RL (Equation 35-5) Therefore, the lower 4 bits of the RGB signal, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained from the relation of BL>GL> RL and (Equation 30-1) to (Equation 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-BL: BL-GL: GL-RL: RL ……… (Equation 35-6)

[B] GL> RL shown in FIG. 16 (b)
Weighting coefficient W in the case of> BL FIG. 20 shows a triangular pyramid in the case of FIG. 16B, that is, when the conversion point P is GL>RL> BL.

As shown in FIGS. 21A to 21D, the triangular pyramid shown in FIG. 20 is further divided into four triangular pyramids on the basis of the conversion point P, and the volume Va of each triangular pyramid. Vd
To calculate. First, the volume V of the triangular pyramid shown in FIG.
a can be shown as follows. Va = 1/3 * (16 * 16) / 2 * (16-GL) = 1/6 * 16 * 16 * (16-GL) ……… (Equation 36-1)

Next, the triangular pyramid shown in FIG.
2 (a). When the triangular pyramid shown in FIG. 22 (a) is viewed from the front side, it can be shown as shown in FIG. 22 (b). Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 16 * SQRT (2)) / 2 * (RL-BL) / SQRT (2) = 1/6 * 16 * 16 * (RL-BL) ……… (Equation 36- 2)

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 16) / 2 * BL = 1/6 * 16 * 16 * BL ……… (Equation 36-3)

From the expressions (36-1) to (36-3), the volume Vb of the triangular pyramid shown in FIG. 21 (b) can be expressed as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 16 * 16 * {16- (16-GL)-(RL-BL) -BL} = 1/6 * 16 * 16 * (GL- RL) ……… (Equation 36-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 21 (a) to 21 (d) can be shown as follows. Va: Vb: Vc: Vd = 16-GL: GL-RL: RL-BL: BL (Equation 36-5) Therefore, the lower 4 bits of the RGB signal, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained from the relationship of GL>RL> BL and the above (formula 30-1) to (formula 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-GL: GL-RL: RL-BL: BL ……… (Equation 36-6)

[C] RL> BL shown in FIG. 16 (c)
Weighting coefficient W in case of> GL FIG. 23 shows a triangular pyramid in the case of FIG. 16C, that is, when the conversion point P is RL>BL> GL.

As shown in FIGS. 24 (a) to 24 (d), the triangular pyramid shown in FIG. 23 is further divided into four triangular pyramids with the conversion point P as a reference, and each triangular pyramid has a volume Va ~. Vd
To calculate. First, the volume V of the triangular pyramid shown in FIG.
a can be shown as follows. Va = 1/3 * (16 * 16) / 2 * (16-RL) = 1/6 * 16 * 16 * (16-RL) ……… (Equation 37-1)

Next, the triangular pyramid shown in FIG.
5 (a). When the triangular pyramid shown in FIG. 25 (a) is viewed from the right, it can be shown as shown in FIG. 25 (b). Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 16 * SQRT (2)) / 2 * (BL-GL) / SQRT (2) = 1/6 * 16 * 16 * (BL-GL) ……… (Equation 37- 2)

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 16) / 2 * GL = 1/6 * 16 * 16 * GL ……… (Equation 37-3)

From the above (Equation 37-1) to (Equation 37-3), the volume Vb of the triangular pyramid shown in FIG. 24 (b) can be expressed as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 16 * 16 * {16- (16-RL)-(BL-GL) -GL} = 1/6 * 16 * 16 * (RL- BL) ……… (Formula 37-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 24 (a) to 24 (d) can be shown as follows. Va: Vb: Vc: Vd = 16-RL: RL-BL: BL-GL: GL (Equation 37-5) Therefore, the lower 4 bits of the RGB signal, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained from the relationship of RL>BL> GL and the above (formula 30-1) to (formula 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-RL: RL-BL: BL-GL: GL ……… (Equation 37-6)

[D] GL ≧ BL shown in FIG. 16 (d)
Weighting coefficient W in the case of ≧ RL (provided that RL ≠ GL) FIG. 26 shows a triangular pyramid in the case of FIG. 16D, that is, when the conversion point P is GL ≧ BL ≧ RL.

As shown in FIGS. 27A to 27D, the triangular pyramid shown in FIG. 26 is further divided into four triangular pyramids on the basis of the conversion point P, and the volumes Va to Vd
To calculate. First, the volume V of the triangular pyramid shown in FIG.
a can be shown as follows. Va = 1/3 * (16 * 16) / 2 * (16-GL) = 1/6 * 16 * 16 * (16-GL) ……… (Equation 38-1)

Next, the triangular pyramid shown in FIG.
8 (a). When the triangular pyramid shown in FIG. 28 (a) is viewed from the front side, it can be shown as shown in FIG. 28 (b). Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 16 * SQRT (2)) / 2 * (BL-RL) / SQRT (2) = 1/6 * 16 * 16 * (BL-RL) ……… (Equation 38- 2)

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 16) / 2 * RL = 1/6 * 16 * 16 * RL ……… (Equation 38-3)

As described above, from (Equation 38-1) to (Equation 38-3), the volume Vb of the triangular pyramid shown in FIG. 27 (b) can be expressed as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 16 * 16 * {16- (16-GL)-(BL-RL) -RL} = 1/6 * 16 * 16 * (GL- BL) ……… (Equation 38-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 27A to 27D can be shown as follows. Va: Vb: Vc: Vd = 16-GL: GL-BL: BL-RL: RL (Equation 38-5) Therefore, the lower 4 bits of the RGB signal, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained from the relationship of GL ≧ BL ≧ RL and the above (formula 30-1) to (formula 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-GL: GL-BL: BL-RL: RL ……… (Equation 38-6)

[E] RL ≧ GL shown in FIG. 16 (e)
Weighting coefficient W in the case of ≧ BL FIG. 29 shows a triangular pyramid in the case of FIG. 16E, that is, when the conversion point P is RL ≧ GL ≧ BL.

As shown in FIGS. 30 (a) to 30 (d), the triangular pyramid shown in FIG. 29 is further divided into four triangular pyramids with the conversion point P as a reference, and each triangular pyramid has a volume Va ~. Vd
To calculate. First, the volume V of the triangular pyramid shown in FIG.
a can be shown as follows. Va = 1/3 * (16 * 16) / 2 * (16-RL) = 1/6 * 16 * 16 * (16-RL) ……… (Equation 39-1)

Next, the triangular pyramid shown in FIG.
1 (a). When the triangular pyramid shown in FIG. 31A is viewed from the right, it can be shown as shown in FIG. Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 16 * SQRT (2)) / 2 * (GL-BL) / SQRT (2) = 1/6 * 16 * 16 * (GL-BL) ……… (Equation 39- 2)

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 16) / 2 * BL = 1/6 * 16 * 16 * BL ……… (Equation 39-3)

From the above (Equation 39-1) to (Equation 39-3), the volume Vb of the triangular pyramid shown in FIG. 30 (b) can be expressed as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 16 * 16 * {16- (16-RL)-(GL-BL) -BL} = 1/6 * 16 * 16 * (RL- GL) ……… (Equation 39-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 30 (a) to 30 (d) can be shown as follows. Va: Vb: Vc: Vd = 16-RL: RL-GL: GL-BL: BL (Equation 39-5) Therefore, the lower 4 bits of the RGB signal, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained from the relationship of RL ≧ GL ≧ BL and the above (formula 30-1) to (formula 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-RL: RL-GL: GL-BL: BL ... (Equation 39-6)

[F] BL ≧ RL shown in FIG. 16 (f)
Weighting coefficient W in the case of ≧ GL (where GL ≠ BL) FIG. 32 shows a triangular pyramid in the case of FIG. 16 (f), that is, when the conversion point P is BL ≧ RL ≧ GL.

As shown in FIGS. 33A to 33D, the triangular pyramid shown in FIG. 32 is further divided into four triangular pyramids with the conversion point P as a reference, and the volumes Va to Vd
To calculate. First, the volume V of the triangular pyramid shown in FIG.
a can be shown as follows. Va = 1/3 * (16 * 16) / 2 * (16-BL) = 1/6 * 16 * 16 * (16-BL) ……… (Equation 40-1)

Next, the triangular pyramid shown in FIG.
4 (a). When the triangular pyramid shown in FIG. 34 (a) is viewed from above, it can be shown as in FIG. 34 (b). Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 16 * SQRT (2)) / 2 * (RL-GL) / SQRT (2) = 1/6 * 16 * 16 * (RL-GL) ……… (Equation 40- 2)

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 16) / 2 * GL = 1/6 * 16 * 16 * GL ……… (Equation 40-3)

From the above (Equation 40-1) to (Equation 40-3), the volume Vb of the triangular pyramid shown in FIG. 33 (b) can be expressed as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 16 * 16 * {16- (16-BL)-(RL-GL) -GL} = 1/6 * 16 * 16 * (BL- RL) ……… (Equation 40-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 33 (a) to 33 (d) can be shown as follows. Va: Vb: Vc: Vd = 16-BL: BL-RL: RL-GL: GL (Equation 40-5) Therefore, the lower 4 bits of the RGB signal, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained from the relation of BL ≧ RL ≧ GL and the above (formula 30-1) to (formula 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-BL: BL-RL: RL-GL: GL ... (Equation 40-6)

Next, as in the case of the cubic interpolation in the first embodiment, the weighting factor W (w1 ...
Consider the order in which w4) is stored. Also in the triangular pyramid interpolation, the weighting factors W (w1 to w4) are continuously stored, similarly to the cubic interpolation in the first embodiment. That is, as shown in FIG. 35, when the data of the color image output from the reading unit 14, that is, each of the lower 4 [bits] of the RGB signals is (RL, GL, BL),
(0,0,0), (1,0,0), (2,0,0),
.., (15,15,14), and (15,15,15) in the order of the grid points, the corresponding weighting factors W (w1 to w1).
Store w4) continuously.

Specifically, the weighting coefficient W (w1 to w4) is stored for (0, 0, 0). Then, (1,
The weighting factor W (w1 to w4) is stored for 0, 0). And finally, for (15,15,15),
The weighting factor W (w1 to w4) is stored.

Therefore, the reference address L of the weighting factor w1
ADR (w1) can be shown as follows. LADR (w1) = RL + GL * 16 + BL * 16∧2 * 4 = RL + GL * 16 + BL * 256 * 4 (Equation 41-1) Further, the reference address LADR of the weighting factors w2 to w4.
(N) (however, N = w2-w4) can be shown as follows. LADR (w2) = RL + GL * 16 + BL * 16∧2 * 4 + 1 = LADR (A) +1 ……… (Equation 41-2)

LADR (w3) = RL + GL * 16 + BL * 16 ∧2 * 4 + 2 = LADR (A) +2 (Formula 41-3) LADR (w4) = RL + GL * 16 + BL * 16∧2 * 4 + 3 = LADR (A) +3 (Equation 41-4) Therefore, if stored in this way, three times if the reference address LADR (w1) of the weight coefficient w1 is known. All the weighting factors W (w1 to w4) can be sequentially obtained only by incrementing.

Next, in the same way as the cubic interpolation in the first embodiment, in the case of triangular pyramid interpolation, the grid point "A
Consider the order in which the grid point data of "~ H" are stored. As shown in FIG. 16, the trigonal pyramid divided into six has eight lattice points “A to H” that form a cube.
The grid point “A” and the grid point “H” are always included. That is, all the triangular pyramids divided into six are composed of the grid points “A, H” and two grid points of the other grid points “B to G”.

In the triangular pyramid interpolation, grid points "A, H"
And the grid point data composed of two grid points of the grid points “B to G” are subjected to interpolation calculation processing. Therefore,
In order to determine one triangular pyramid from among the six divided triangular pyramids, the magnitude relationship of each lower 4 [bit] of the RGB signal, that is, each value of (RL, GL, BL) is determined. There is a need.

Therefore, the offset of the reference address at the lattice point of the triangular pyramid determined by the magnitude relation of the values of (RL, GL, BL) will be examined. [A] BL>GL> RL triangular pyramid (FIG. 16A)
Reference) The lattice points forming this triangular pyramid are "A, B, D, H". Therefore, each grid point "A, B, D," from the reference address of the grid point "A" in the reference "C" component
The offset of “H” is (Formula 1-1) to (Formula 1-8) above.
Therefore, they are “0,289,306,307”, respectively.

[B] In the case of a triangular pyramid of GL>RL> BL (see FIG. 16B), the lattice points forming this triangular pyramid are "A, C, G, H". Therefore, each grid point "A, C, G," from the reference address of the grid point "A" in the reference "C" component
The offset of “H” is (Formula 1-1) to (Formula 1-8) above.
Therefore, they are “0, 17, 18, 307”, respectively.

[C] In the case of a triangular pyramid of RL>BL> GL (see FIG. 16 (c)) The lattice points forming this triangular pyramid are "A, E, F, H". Therefore, each grid point “A, E, F,” from the reference address of the grid point “A” in the reference “C” component
The offset of “H” is (Formula 1-1) to (Formula 1-8) above.
As a result, it becomes “0, 1, 290, 307”, respectively.

[D] GL ≧ BL ≧ RL (where RL ≠ G
In the case of the L) triangular pyramid (see FIG. 16D), the lattice points forming this triangular pyramid are “A, C, D, H”. Therefore, each grid point “A, C, D,” from the reference address of the grid point “A” in the reference “C” component
The offset of “H” is (Formula 1-1) to (Formula 1-8) above.
Therefore, they are “0, 17, 306, 307”, respectively.

[E] In the case of a triangular pyramid of RL ≧ GL ≧ BL (see FIG. 16 (e)) The lattice points forming this triangular pyramid are “A, E, G, H”. Therefore, each grid point "A, E, G," from the reference address of the grid point "A" in the reference "C" component
The offset of “H” is (Formula 1-1) to (Formula 1-8) above.
Therefore, they are “0, 1, 18, 307”, respectively.

[F] BL ≧ RL ≧ GL (where GL ≠ B
In the case of the L) triangular pyramid (see FIG. 16 (f)), the lattice points forming this triangular pyramid are “A, B, F, H”. Therefore, each grid point "A, B, F, from the reference address of the grid point" A "in the reference" C "component
The offset of “H” is (Formula 1-1) to (Formula 1-8) above.
Therefore, they are respectively “0, 29, 289, 307”.

Therefore, also in the case of the triangular pyramids shown in [a] to [f], as in the case of the cubic interpolation in the first embodiment, the reference address of the lattice point "A" in the "C" component is used as a reference. Then, the offset of the reference address at the grid point “BH” differs depending on the grid point “BH”.
In addition, a method of determining the magnitude relationship of each value of (RL, GL, BL) by using a comparison instruction is conceivable, but if the comparison instruction is used when determining the magnitude relationship by using software, the entire device becomes Processing speed becomes slow.

Therefore, the above [a] to [f] (see FIG. 16)
The cases of the triangular pyramids (see (a) to (f)) are defined as follows by using the offset “X”. [A] In the case of a BL>GL> RL triangular pyramid, in other words, in the case of a triangular pyramid composed of lattice points “A, B, D, H”: X = 0 ... (Equation 42-1) [ b] In the case of a triangular pyramid of GL>RL> BL, in other words, in the case of a triangular pyramid composed of lattice points “A, C, G, H”: X = 1 ... (Equation 42-2) [c ] In the case of a triangular pyramid of RL>BL> GL, in other words, in the case of a triangular pyramid composed of lattice points “A, E, F, H”: X = 2 ... (Equation 42-3)

[0186]   [D] In the case of a triangular pyramid of GL ≧ BL ≧ RL and RL ≠ GL, in other words, the grid point “A , C, D, H ”in the case of a triangular pyramid: X = 3 ... (Equation 42-4)   [E] In the case of a triangular pyramid with RL ≧ GL ≧ BL, in other words, the grid points “A, E, G, H In the case of a triangular pyramid consisting of: X = 4 ... (Equation 42-5)   [F] In the case of BL ≧ RL ≧ GL and GL ≠ BL triangular pyramids, in other words, the grid point “A , B, F, H ”in the case of a triangular pyramid: X = 5 ... (Equation 42-6)

Then, as shown in FIG. 36, [a]-
In the case of [f], that is, the offset “X” determined by the magnitude relationship of (RL, GL, BL) is stored as the LUT. If stored in this way, the magnitude relationship of the respective values of (RL, GL, BL) can be used without using a comparison instruction, and
It can also be determined which triangular pyramid the conversion point P corresponds to.

Next, the order of storing the grid point data at the four grid points forming the triangular pyramid will be examined.
As shown in FIG. 37, each upper 4 [bit] of the RGB signal
Let (RH, GH, BH) be (0,0,0),
(1,0,0), (2,0,0), ..., (15,1)
5, 14) and (15, 15, 15) in the order of four grid points “A, B, D, CMYK”.
"H", "A, C, G, H", "A, E, F, H",
"A, C, D, H", "A, E, G, H", "A, B,
F, H ”grid point data is stored.

Specifically, first, for (0,0,0), the lattice point data of the lattice points "A, B, D, H" corresponding to the "C" component, and the "M" component. Lattice points “A, B,
"D, H" grid point data, "Y" component grid point data "A, B, D, H" grid point data, "K" component grid point "A, B, D, H" The grid point data of is stored.

Then, for (0,0,0), "C"
It corresponds to the grid point data of the grid points “A, C, G, H” corresponding to the component, the grid point data of the grid point “A, C, G, H” corresponding to the “M” component, and the “Y” component. Lattice points “A, C,
The grid point data of "G, H" and the grid point data of grid points "A, C, G, H" corresponding to the "K" component are stored.

Next, for (0,0,0), grid point data of grid points "A, E, F, H" corresponding to the "C" component and grid point "M" of the "M" component. "A, E, F, H" grid point data, grid points "A, E, corresponding to the" Y "component
The grid point data of "F, H" and the grid point data of "A, E, F, H" corresponding to the "K" component are stored.

Finally, for (15,15,15), the grid point "A, B, F, H" corresponding to the "C" component
Grid point data of the grid point “A,
"B, F, H" grid point data, "Y" component grid point data "A, B, F, H" grid point data, "K" component grid point "A, B, F, The “H” grid point data is stored.

If stored in this way, the grid point data to be stored increases, but the grid point "A" in the "C" component is increased.
If the reference address HADR (A) of the above is determined, the other three lattice points “X1, X2, H” forming the triangular pyramid are simply incremented (where X1 and X2 are the lattice points “B to G”). Of the two) can be obtained. Therefore, the grid point data of (RH, GH, BH) is not the grid point data of (0,0,0) to (16,16,16) but (0,0,0) to (15 ，
It is only necessary to store the grid point data up to 15, 15).

Therefore, the grid point "A" in the "C" component
The reference address HADR (A) can be shown as follows. HADR (A) = (RH + GH * 16 + BH * 16∧2) * 4 * 4 * 6 + 4 * 4 * X = (RH + GH * 16 + BH * 256) * 96 + 16 * X ...... ... (Formula 43)

By the way, as shown in FIG. 37, two of the four grid points "A,
"H" is irrelevant to the magnitude relationship of each lower 4 bits of the RGB signal, that is, each value of (RL, GL, BL). In other words, the grid point data of grid point “A, H” is
In the case of any of the pyramids [a] to [f],
This is grid point data that is essential for the calculation of triangular pyramid interpolation. Therefore, the remaining two grid points used for the triangular pyramid interpolation are selected from the six grid points “B to G”. That is, the grid point data necessary for the triangular pyramid interpolation is the grid points "A, H" and the grid points "X1, X2" (where X1, X2 are the grid points "B ~
2) of “G”). Specifically, among the grid points “B to G”, the grid points combined with the grid point “A, H” are “B, D”, “C, G”, “E,”.
“F”, “C, D”, “E, G”, “B, F”.

Therefore, first, the grid point data of the grid points "A, H" are stored as one set. Subsequently, the grid point data of the grid points "B, D" are stored as one set. Similarly, the grid points “C, G” and the grid points “E,
The grid point data of “F”, grid point “C, D”, grid point “E, G”, and grid point “B, F” are stored as one set.

That is, as shown in FIG. 38, assuming that the upper 4 bits of the RGB signal are (RH, GH, BH), (0, 0, 0), (1, 0, 0), (2 , 0,
0, ..., (15,15,14), (15,15,
15) In the order of the grid points, grid points “A, H”, grid points “B, D”, grid points “C, G”, grid points “E, F”, grid points “C, for each CMYK. D ", grid point" E,
The grid point data of “G” and the grid point “B, F” is stored as one set.

Specifically, first, for (0,0,0), the grid point data of the grid point "A, H" corresponding to the "C" component and the grid point "A" corresponding to the "M" component. , H ”grid point data, grid point data“ A, H ”grid point data corresponding to the“ Y ”component, and grid point data“ A, H ”grid point data corresponding to the“ K ”component.

Then, for (1, 0, 0), "C"
Grid point data of grid points “B, D” corresponding to the components,
The grid point data of the grid point “B, D” corresponding to the “M” component, the grid point data of the grid point “B, D” corresponding to the “Y” component, the grid point “B, corresponding to the K component” The grid point data of "D" is stored.

Finally, for (15,15,15), the grid point data of the grid point "B, F" corresponding to the "C" component, and the grid point "B, F" corresponding to the "M" component. The grid point data of "," the grid point data of grid points "B, F" corresponding to the "Y" component, and the grid point data of grid points "B, F" corresponding to the "K" component are stored.

If stored in this way, the grid point data of the four grid points "A, X1, X2, H" forming the triangular pyramid cannot be stored continuously, but the storage order shown in FIG. Compared with (1), only one grid point data of the grid point “A, H” is stored for (RH, GH, BH). As a result, the number of grid point data to be stored can be reduced as compared with the storage order shown in FIG.

Therefore, the grid point "A" in the "C" component
The reference address HADR (A) can be shown as follows. HADR (A) = (RH + GH * 16 + BH * 16∧2) * 16 = (RH + GH * 16 + BH * 256) * 16 ……… (Equation 44-1) Therefore, in the “C” component The reference address HADR (H) of the grid point “H” can be shown as follows. HADR (H) = (RH + GH * 16 + BH * 16∧2) * 16 + 1 = HADR (A) +1 ……… (Equation 44-2)

Also, the grid point "B ~
The reference address HADR (BG) of “G” can be indicated as follows. HADR (BG) = (RH + GH * 16 + BH * 16∧2) * 14 * 4 + 8 * (X + 1) = (RH + GH * 16 + BH * 256) * 56 + 8 * (X + 1) ……… (Equation 45)

Further, among the grid points “B to G”, (R
If two grid points determined by the magnitude relationship of L, GL, and BL) are “X1, X2”, reference addresses HADR (X1), HA of grid points “X1, X2” in the “C” component
DR (X2) can be shown as follows. HADR (X1) = (RH + GH * 16 + BH * 16∧2) * 14 * 4 + 8 * (X + 1) = (RH + GH * 16 + BH * 256) * 56 + 8 * (X + 1) ……… (Equation 46-1) HADR (X2) = (RH + GH * 16 + BH * 16∧2) * 14 * 4 + 8 * (X + 1) +1 = (RH + GH * 16 + BH * 256) * 56 + 8 * (X + 1) +1 ……… (Equation 46-2)

"X" in (Equation 45), (Equation 46-1) and (Equation 46-2) is (RL, GL, BL).
Of the values of (RL, GL, BL). Specifically, according to the magnitude relationship of the values of (RL, GL, BL), the definitions of (Expression 42-1) to (Expression 42-6) above are defined. Determined by the formula. The offset "X" is also determined according to each value of (RL, GL, BL) shown in FIG.

Therefore, the grid point “A” in the “C” component
The reference address HADR (A) of the above is (Formula 44-1)
Can be calculated from In addition, the reference address HADR (H) of the lattice point “H” in the “C” component is, as is clear from the above (Equation 44-2), the above (Equation 44-
It can be obtained by incrementing 1). Furthermore, the grid points “A, H” at “M, Y, K”
Can also be obtained by sequentially incrementing (Equation 44-1).

In addition, the grid point "X" in the "C" component
1 ”is the reference address HADR (X1),
-1) can be calculated from. Also, the reference address HA of the grid point "X2" in the other "C" component
DR (X2) can be obtained by incrementing (Equation 46-1), as is clear from (Equation 46-2). Furthermore, the grid points “X1, X2” at “M, Y, K” can also be obtained by incrementing the above (formula 46-1).

Next, the interpolated value I of the "C, M, Y, K" components
A method of calculating PL (N) (however, N = C, M, Y, K) will be described. Color image data output from the reading unit 14, that is, each lower 4 [bi of RGB signals
t] is (RL, GL, BL), and (RL, GL,
Assuming that “LL, LM, LS” are arranged in descending order of the values of (BL), the interpolated value IPL (N) of the “C, M, Y, K” components (where N = C, M, Y, K) is Similar to the cubic interpolation in the first embodiment, it can be shown as follows. IPL (N) = {(16-LL) * f (A) + (LL-LM) * f (X1) + (LM-LS) * f (X2) + LS * f (H)} / 16 ...... ... (Formula 47)

"X1, X2" shown in (Equation 47)
Are two of the grid points “B to G” determined by the magnitude relation of the values of (RL, GL, BL). In addition, (Equation 4
F (A), f (X1), f (X2) and f shown in 7)
(H) is the above (formula 44-1) and the above (formula 46-1).
Based on, "C, M,
It is the grid point data of the grid points "A, X1, X2, H" in the "Y, K" component.

Next, in the triangular pyramid interpolation, the case where the conversion point P corresponds to exceptional processing will be described. Similar to the case of the cubic interpolation in the first embodiment, the case corresponding to the exceptional processing in the triangular pyramid interpolation means that the upper 4 [bit] of each RGB signal is (RH, GH) in the cube shown in FIG. , BH), the (RH, GH, B
This is the case where any of the values in (H) is "15".

Therefore, when the conversion point P corresponds to the exceptional processing, there are the following "7 ways" as in the case of the cubic interpolation in the first embodiment. [1] RH = 15, GH <15, BH <15 (E1 shown in FIG. 8) [2] RH <15, GH = 15, BH <15 (E2 shown in FIG. 8) [3] RH <15 , GH <15, BH = 15 (E3 shown in FIG. 8) [4] RH = 15, GH = 15, BH <15 (E4 shown in FIG. 8) [5] RH <15, GH = 15, BH = 15 (E5 shown in FIG. 8) [6] RH = 15, GH <15, BH = 15 (E6 shown in FIG. 8) [7] RH = 15, GH = 15, BH = 15 (FIG. 8 above) E7)

However, in the triangular pyramid interpolation, the cube is further divided into six to obtain the triangular pyramid. In other words, as shown in FIG.
If [bit] is (RL, GL, BL), then (R
L, GL, BL), the cube is further divided into 6 to obtain a triangular pyramid according to the magnitude relation of each value. That is, (R
L, GL, BL) have the following relationship. [A] BL>GL> RL [b] GL>RL> BL [c] RL>BL> GL [d] GL ≧ BL ≧ RL (where RL ≠ GL) [e] When RL ≧ GL ≧ BL [f] When BL ≧ RL ≧ GL (where GL ≠ BL)

As a result, in the case of exceptional processing in the triangular pyramid interpolation, "7 types" are added depending on the possible values of each value of (RH, GH, BH), and in addition (RL, GL, BH).
There are “6 ways” depending on the magnitude relationship of each value of L). Therefore, in the triangular pyramid interpolation, when the conversion point P corresponds to exceptional processing, “7 ways” * “6 ways” = “42 ways”.

Therefore, first, [1] RH = 15, GH <1
5, BH <15, and [e] RL ≧ G
The weighting coefficient W (w1 to w4) when L ≧ BL is calculated. In FIG. 39, RH = 15, GH <15, BH <
15 shows a triangular pyramid in the case of 15 and RL ≧ GL ≧ BL.

Therefore, as shown in FIGS. 40A to 40D, the triangular pyramid shown in FIG. 39 is further divided into four triangular pyramids on the basis of the conversion point P, and the volume of each triangular pyramid is divided. V
Calculate a to Vd. First, the volume Va of the triangular pyramid shown in FIG. 40 (a) can be shown as follows. Va = 1/3 * (16 * 16) / 2 * (15-RL) = 1/6 * 16 * 15 * (16-16 * RL / 15) ……… (Equation 48-1)

Next, the triangular pyramid shown in FIG.
1 (a). When viewed from the right, the triangular pyramid shown in FIG. 41 (a) can be shown as shown in FIG. 41 (b). Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 15 * SQRT (2)) / 2 * (GL-BL) / SQRT (2) = 1/6 * 16 * 15 * (GL-BL) ……… (Equation 48− 2)

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 15) / 2 * BL = 1/6 * 16 * 15 * BL ……… (Equation 48-3)

As described above, from (Equation 48-1) to (Equation 48-3), the volume Vb of the triangular pyramid shown in FIG. 40 (b) can be expressed as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 16 * {16 * 15-16 * (15-RL) -15 * (GL-BL) -15 * BL} = 1/6 * 16 * 15 * (16 * RL / 15-GL) ……… (Equation 48-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 40 (a) to 40 (d) can be shown as follows. Va: Vb: Vc: Vd = 16-16 * RL / 15: 16 * RL / 15-GL: GL-BL: BL (Equation 48-5) Therefore, the lower 4 bits of the RGB signal, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained by the relation of RL ≧ GL ≧ BL and the above (formula 30-1) to (formula 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-16 * RL / 15: 16 * RL / 15-GL: GL-BL: BL ……… (Equation 48-6)

[5] RH <15, GH = 15, B
When H = 15 and [c] RL>BL> G
The weighting coefficient W (w1 to w4) in the case of L is calculated. In FIG. 42, RH <15, GH = 15, BH = 15
And the triangular pyramid in the case of RL>BL> GL.

Therefore, as shown in FIGS. 43 (a) to 43 (d), the triangular pyramid shown in FIG. 42 is further divided into four triangular pyramids based on the conversion point P, and the volume of each triangular pyramid is divided. V
Calculate a to Vd. First, the volume Va of the triangular pyramid shown in FIG. 43 (a) can be shown as follows. Va = 1/3 * (15 * 15) / 2 * (16-RL) = 1/6 * 15 * 15 * (16-RL) ……… (Equation 49-1)

Next, the triangular pyramid shown in FIG.
4 (a). When viewed from the right, the triangular pyramid shown in FIG. 44 (a) can be shown as shown in FIG. 44 (b). Therefore, the volume Vc of the triangular pyramid shown in FIG.
Can be shown as: Vc = 1/3 * (16 * 15 * SQRT (2)) / 2 * (BL-GL) / SQRT (2) = 1/6 * 15 * 15 * (16 * BL / 15-16 * GL / 15 ) ……… (Equation 49-2)

Next, the volume V of the triangular pyramid shown in FIG.
d can be shown as follows. Vd = 1/3 * (16 * 15) / 2 * GL = 1/6 * 15 * 15 * (16 * GL / 15) ……… (Equation 49-3) Above, (Equation 49-1) to ( From Equation 49-3), FIG.
The volume Vb of the triangular pyramid shown in (b) can be shown as follows. Vb = Total volume- (Va + Vc + Vd) = 1/6 * 15 * {16 * 15-15 * (16-RL) -16 * (BL-GL) -16 * GL} = 1/6 * 15 * 15 * (RL-16 * BL / 15) ……… (Equation 49-4)

As a result, the ratio of the volumes Va to Vd of the triangular pyramid shown in FIGS. 43 (a) to 43 (d) can be shown as follows. Va: Vb: Vc: Vd = 16-RL: RL-16 * BL / 15: 16 * BL / 15-16 * GL / 15: 16 * GL / 15 ……… (Equation 49-5) Therefore, RGB signal Lower 4 bits of, that is, (R
L, GL, BL) volume ratio Va: Vb: Vc:
Vd becomes the weighting factor W in the triangular pyramid.

Therefore, the weighting factor W is obtained from the relation of RL>BL> GL and the above (formula 30-1) to (formula 30-4).
(W1 to w4) can be shown as follows. w1: w2: w3: w4 = 16-RL: RL-16 * BL / 15: 16 * BL / 15-16 * GL / 15: 16 * GL / 15 ……… (Equation 49-6)

Here, based on the above two examples, the weighting factor W (w1 to w4) when not corresponding to the exception processing and the weighting factor W (w1 to w4) when corresponding to the exception processing are compared and examined. To do.

Specifically, the weighting factor W (w1 to w in the case of RL ≧ GL ≧ BL in the case where exception processing is not applicable)
4) and the weighting coefficient W (w in the case of RL ≧ GL ≧ BL in the case of exception processing).
1-w4) and (Equation 48-6) are compared and examined, (Equation 48-6) is obtained by substituting “16 * RL / 15” into “RL” of (Equation 39-6). I know there is.

[0229] Also, the above (Equation 37-6) showing the weighting coefficient W (w1 to w4) in the case of RL>BL> GL in the case of non-exception processing, and RL>BL> in the case of exceptional processing. Weighting coefficient W (w1 to w for GL)
4), the above (formula 49-6) is compared and examined.
(Equation 49-6) has "16" added to "GL" of (Equation 37-6).
It can be seen that “* GL / 15” is a substitution of “16 * BL / 15” into “BL”.

As described above, (Equation 39-6) and (Equation 48-6) and (Equation 37-6) and (Equation 49-6) shown in two examples.
As can be inferred from, in the exceptional processing of the triangular pyramid interpolation, the weighting factor W (w1
~ W4) in the above (Formulas 35-40-6), (RH,
It can be seen that it can be obtained by substituting the following relational expressions based on the possible values of each value of (GH, BH).

[0231]   [1] When RH = 15, GH <15, BH <15   RL ← 16 * RL / 15 ……… (Equation 50)   [2] When RH <15, GH = 15, BH <15   GL ← 16 * GL / 15 ……… (Equation 51)   [3] In the case of RH <15, GH <15, BH = 15   BL ← 16 * BL / 15 ……… (Equation 52)   [4] When RH = 15, GH = 15, BH <15   RL ← 16 * RL / 15 ……… (Equation 53-1)   GL ← 16 * GL / 15 ……… (Equation 53-2)

[0232]   [5] When RH <15, GH = 15, BH = 15   GL ← 16 * GL / 15 ……… (Equation 54-1)   BL ← 16 * BL / 15 ……… (Equation 54-2)   [6] When RH = 15, GH <15, BH = 15   RL ← 16 * RL / 15 ……… (Equation 55-1)   BL ← 16 * BL / 15 ……… (Equation 55-2)   [7] When RH = 15, GH = 15, BH = 15   RL ← 16 * RL / 15 ……… (Equation 56-1)   GL ← 16 * GL / 15 ……… (Equation 56-2)   BL ← 16 * BL / 15 ……… (Equation 56-3)

Next, a method of determining whether or not the conversion point P corresponds to exceptional processing will be examined. When the conversion point P corresponds to an exceptional process, as in the case of the cubic interpolation in the first embodiment, each upper 4 [bit] of the RGB signal at the conversion point P, that is, (RH, GH, BH).
It is possible to judge whether or not any one of the values is “15”.

As a result, the judgment can be made in the same manner as in the case of the cubic interpolation in the first embodiment. Specifically, as shown in FIG. 9, if the value indicating whether or not the exception processing is applied is “Z” and the value of “Z” for (RH, GH, BH) is stored, Based on the value of "Z", it can be determined whether or not the exception processing is applicable. Therefore, "Z" indicating whether or not it corresponds to exception processing
The reference address ZHADR (Z) of can be expressed in the same manner as the above (Formula 15).

Next, the order of storing the weighting factors W (w1 to w4) in the case of exceptional processing will be described with reference to the drawings. As shown in FIG. 45, the value indicating whether or not the exception processing is applied is “Z”, the data of the color image output from the reading unit 14, that is, each lower 4 [bit] of the RGB signal is (RL, GL). , BL) and the weighting factor is “W”, and corresponds to each Z (= 0 to 7) (RL, G
L, BL) are stored. Also, (RL, GL, BL)
The weighting factor W corresponding to is continuously stored in the order of w1 to w4.

Specifically, the weighting coefficient W (w1 to w4) is stored for (0, 0, 0) when Z = “0”. Then, the weighting factor W (w1 to w4) is stored for (1, 0, 0). And finally, Z
The weighting factor W (w1 to w4) is stored for (15, 15, 15) in the case of “7”.

Therefore, the reference address ELADR (w1) of the weighting coefficient W1 including the exception process can be shown as follows. ELADR (w1) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 4 = (RL + GL * 16 + BL * 256) * 8 + Z * 4096 * 4 ...... (Equation 57-1)

Also, the weighting factors w2 to w4 including the exception handling.
Reference address ELADR (N) (where N = w2 to w
4) can be shown as follows. ELADR (w2) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 4 + 1 = ELADR (w1) +1 ……… (Equation 57-2) ELADR (w3) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 4 + 2 = ELADR (w1) +2 ……… (Equation 57-3) ELADR (w4) = (RL + GL * 16 + BL * 16∧2) * 8 + Z * 16∧3 * 4 + 3 = ELADR (w1) +3 ……… (Equation 57-4)

Therefore, if stored in this way, if the reference address ELADR (w1) of the weighting factor w1 is known, then 3
All weighting factors W (w
1 to w4) can be sequentially obtained.

Next, in the exception processing, "C, M, Y,
Interpolated value IPL (N) of the “K” component (where N = C, M,
A method of calculating Y, K) will be described. That is,
In the exception processing, there are “7 ways” depending on the possible values of (RH, GH, BH) and “6 ways” depending on the magnitude relationship of the values of (RL, GL, BL). Therefore, if it corresponds to the exceptional process, (RH, G) is added to the (Formula 47) indicating the interpolation value IPL (N) when the exceptional process is not applied.
H, BH) can be obtained by substituting the relational expressions of (Expression 50) to (Expression 56-3) determined based on the possible values of each value of (H, BH).

Next, in the facsimile apparatus 1, 1 when the RGB signal of the conversion point P is output from the reading unit 14.
The operation of color conversion processing for pixels will be described with reference to the flowchart shown in FIG. 11 in the first embodiment. It should be noted that this operation is executed under the control of the MPU 11 based on the program stored in the ROM 12.

In step S1 shown in FIG. 11, each lower 4 bits, that is, (RL, GL, BL) is extracted from the RGB signal at the conversion point P. In step S2, (RL,
The reference address LADR (w1) of the weighting factor (w1) is calculated based on GL, BL). Specifically, the reference address LADR (w1) of the weighting factor w1 is calculated based on the (formula 41-1).

In step S3, the weighting factors W (w1 to w4) in the triangular pyramid surrounding the conversion point P are stored in the ROM.
It is sequentially read from 12. Specifically, the reference address LADR (w1) calculated in step S2
Are sequentially incremented from, and all weighting factors W
(W1 to w4) is the weighting factor W (w1 shown in FIG. 35.
To w4) are sequentially read from the LUT.

At step S4, the RG of the conversion point P
Each upper 4 bits from the B signal, that is, (RH, G
H, BH) is extracted. In step S5, (RH, GH, B extracted in step S4 is extracted.
Based on H), the reference address HADR (A) of the grid point “A” in the “C” component is calculated. In particular,
First, the offset “X” determined by the magnitude relation of the respective values of (RL, GL, BL) shown in FIG. 36 is read. Then, the reference address HADR (A) of the grid point “A” in the “C” component is calculated based on the above (formula 43).

In step S6, grid points "A, X1, X2, H" in the triangular pyramid surrounding the conversion point P (where X1 and X2 are two of the grid points "BG", in other words, For example, two of the grid points "B to G" determined by the offset "X" in step S5) are stored in the ROM1.
It is sequentially read from 2. Specifically, the step S
Lattice point "A" in "C" component calculated in 5
Are sequentially incremented from the reference address HADR (A) of the grid point “A, X1, X2,
The grid point data of "H" is sequentially read from the LUT of the grid point data shown in FIG.

In step S7, interpolation calculation at the conversion point P is performed. Specifically, the step S3
Based on the weighting factor W (w1 to w4) read in step S6 and the grid point data of the grid points “A, X1, X2, H” in the “C” component read in step S6, The interpolation calculation at the conversion point P is performed by the expression 47).

In step S8, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component. If there is a next component, the process proceeds to step S9. On the other hand, when there is no next component, that is, for all "C, M,
When the interpolation calculation is completed for the “Y, K” components,
This process ends.

In step S9, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 37, since the grid point data of the grid points “A, X1, X2, H” for the “C, M, Y, K” components is stored, The reference address HADR (A) of the lattice point "A" in the "C" component calculated in step S5 is simply incremented. Then, steps S6 to S9 are repeatedly executed, and the interpolation calculation of "M, Y, K" at the conversion point P is performed. That is, steps S6 to S9 are repeatedly executed, and “C, M, Y,
The K ”component is interpolated, and the RGB signal at the conversion point P is converted into a CMYK signal. Note that generally, since there are many pixels to be converted, this processing is repeatedly executed for all the read pixels.

As described above in detail, according to the fourth embodiment, the following operation and effect can be obtained in addition to the operation and effect of the first embodiment. (1) In the fourth embodiment, the LUT of grid point data shown in FIG. 37 is used instead of the LUT of grid point data shown in FIG. 38. It is the reference address HAD of the grid point "A"
This is because if R (A) is calculated, the grid point data at the other four grid points forming the triangular pyramid can be obtained only by sequentially incrementing it. In other words, when the LUT of the grid point data shown in FIG.
The reference address HADR (A) of the grid point "A" and the reference address HADR (X1) of the grid point "X1" must be calculated based on -1) and (Equation 45). Then, these reference addresses HADR (A) and HADR
It is necessary to increment from (X1) to obtain grid point data of grid points “H” and “X2”. Therefore, the LUT of the grid point data shown in FIG. 37 is intentionally used. Therefore, even in the case of triangular pyramid interpolation,
Color conversion processing can be performed at high speed using software.

(2) In the fourth embodiment, since the triangular pyramid interpolation is used, the data required for the interpolation calculation processing is four each for the grid point data and the weighting coefficient W. That is, the grid point data of the grid points “A, X1, X2, H” and the weighting coefficient W (w1 to w4). Therefore, compared with the case of the cubic interpolation in the first to third embodiments, the accuracy of the interpolation calculation result of the conversion point P is inferior, but the amount of data to be read becomes small, so that the color conversion processing is performed at higher speed. You can

(3) As is clear from (Equation 47), in the calculation processing of the triangular pyramid interpolation, the product-sum calculation is 4
[Times] Therefore, the color conversion processing can be performed at a higher speed than in the case of the cubic interpolation in the first to third embodiments.

(4) Even in the case of triangular pyramid interpolation, FIG. 11 used in the cubic interpolation in the first embodiment.
The processing of the flowchart shown in can be applied.
For this reason, the process of the flowchart shown in FIG. 11 can be easily applied to various interpolation calculation methods.

[Fifth Embodiment] ... Triangular pyramid interpolation (four-point interpolation) In this fifth embodiment, the same structure as that of the first embodiment is the same as that of the first embodiment. And the description thereof will be omitted. The fifth embodiment considers the case corresponding to the exception processing in the fourth embodiment. Therefore, in the fifth embodiment, the ROM 12 is an LUT for determining whether or not the conversion point P shown in FIG. 9 corresponds to exception processing, and each value of (RL, GL, BL) shown in FIG. 36. Of the offset determined by the magnitude relationship of L, and the LUT of the grid point data shown in FIG.
And the LUT of the weighting factors W (w1 to w4) including the case corresponding to the exceptional processing shown in FIG. 45 is stored.

Next, in the facsimile apparatus 1, 1 when the RGB signal of the conversion point P is output from the reading unit 14.
Regarding the operation of color conversion processing for pixels, FIG. 12 and FIG.
This will be described with reference to the flowchart shown in. Note that this operation is based on a program stored in the ROM 12,
It is executed under the control of the PU 11.

In step S11 shown in FIG. 12,
From the RGB signal at the conversion point P, each upper 4 [bit], that is, (RH, GH, BH) is extracted. Step S12
In step S11, it is determined whether the conversion point P corresponds to an exceptional process, based on (RH, GH, BH) extracted in step S11. Specifically, the reference address ZHADR (Z) of “Z” indicating whether or not the exception processing is performed is calculated based on the above (Equation 15).
Then, the value of "Z" is extracted based on the reference address ZHADR (Z). Specifically, it is determined whether or not the conversion point P corresponds to the exceptional process based on the LUT for determining whether or not the conversion point P shown in FIG. 9 corresponds to the exceptional process. That is, when the value of “Z” indicating whether or not the exception processing is applied is “1 to 7”, it is determined that the exception processing is applied, and the process proceeds to step S31 shown in FIG. On the other hand, when the value of “Z” indicating whether or not the exception processing is applied is “0”, it is determined that the exception processing is not applied,
Control goes to step S13. In addition, the above (formula 15-1)
It may be possible to determine whether or not it corresponds to the exception processing based on.

At step S13, R of the conversion point P
Each lower 4 bits from the GB signal, that is, (RL, G
L, BL) are extracted. In step S14,
(RL, GL, extracted in step S13)
The reference address ELADR (w1) of the weighting coefficient (w1) including the case corresponding to the exceptional process is calculated based on BL). Specifically, based on the above (formula 57-1),
The reference address ELADR (w1) of the weighting factor w1 is calculated.

In step S15, the weighting coefficient W (w1 to w4) in the triangular pyramid surrounding the conversion point P is RO.
The data is sequentially read from M12. Specifically, the reference address ELADR (w calculated in step S4 is
1) are sequentially incremented, and all the weighting factors W (w1 to w4) are read from the LUT of the weighting factors W (w1 to w4) including the case corresponding to the exceptional processing shown in FIG. Specifically, the reference address ELADR of the weighting factor w1 calculated in step S14.
The LUT of the weighting factors W (w1 to w4) including the case where all the weighting factors W (w1 to w4) are sequentially incremented from (w1) and correspond to the exceptional processing shown in FIG. 45.
Are sequentially read. More specifically, the weighting factor W (w1−w1) including the case corresponding to the exceptional processing shown in FIG.
w4) LUT, step S shown in FIG.
The weighting coefficient W (w1 to w) corresponding to the value of “Z (= 0)” calculated based on the above (Equation 15) in the processing of 12
4) are sequentially read.

It should be noted that, instead of the processing of steps S14 and S15, (R
Based on L, GL, BL), the weighting factor W (w1 to w
4) may be calculated. Specifically, (RL, GL, B
Based on the magnitude relation of each value of L),
The weighting factor W (w1 to w4) may be calculated by -6).

In step S16, R of the conversion point P
Each upper 4 bits from the GB signal, that is, (RH, G
H, BH) is extracted. In addition, this upper 4 [bit]
Since the extraction processing of step S11 is the same as that of step S11, the top 4 [b
It] may be used as it is.

In step S17, the reference address HADR (A) of the grid point "A" in the "C" component is calculated based on (RH, GH, BH) extracted in step S16. Specifically, first, FIG.
The offset "X" determined by the magnitude relation of the respective values of (RL, GL, BL) shown in 6 is read. Then, the reference address HADR (A) of the grid point “A” in the “C” component is calculated based on the above (formula 43).

The reference address HADR (A) of the lattice point "A" in the "C" component is the reference address ZHADR of "Z" calculated in step S12.
(Z), that is, the reference address ZHADR of "Z" indicating whether or not the exception processing shown in (Equation 15) is satisfied.
The reference address HADR (A) of the grid point “A” in the “C” component may be calculated by multiplying (Z) by “96” and then adding the value of “16 * X”.

In step S18, grid points "A, X1, X2, H" in the triangular pyramid surrounding the conversion point P are shown.
(However, X1 and X2 are 2 of the lattice points “BG”)
In other words, two of the grid points “B to G” determined by the offset “X” in step S17) are sequentially read from the ROM 12. Specifically, it is sequentially incremented from the reference address HADR (A) of the grid point “A” in the “C” component calculated in step S17, and the grid point “A, X in the“ C ”component is sequentially incremented.
1, X2, H ”grid point data are sequentially read from the LUT of grid point data shown in FIG.

In step S19, interpolation calculation at the conversion point P is performed. Specifically, the step S
Weighting coefficient W (w1 to w4) read in 15
And the grid point data of the grid points “A, X1, X2, H” in the “C” component read in step S18, the interpolation calculation at the conversion point P is performed by the above (formula 47). Be seen.

In step S20, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component.
If there is a next component, the process proceeds to step S21. On the other hand, when there is no next component, that is, for all "C,
When the interpolation calculation is completed for the “M, Y, K” components, this process is completed.

In step S21, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 37, for the "C, M, Y, K" components, the grid points "A, B, D, H" and the grid points "A, C, G, H", respectively. , Grid point “A, E, F, H”, grid point “A, C, D, H”, grid point “A, E, G, H”, grid point “A, B, F, H” Since the grid point data is stored, "C" calculated in step S17
Reference address HADR of grid point "A" in the component
It simply increments sequentially from (A). Then, steps S18 to S21 are repeatedly executed, and the interpolation calculation of "M, Y, K" at the conversion point P is performed. That is, steps S18 to S21 are repeatedly executed, and “C, M, Y,
The K ”component is interpolated, and the RGB signal at the conversion point P is converted into a CMYK signal.

In step S31 shown in FIG. 13,
Each lower 4 bits, that is, (RL, GL, BL) is extracted from the RGB signal at the conversion point P. Step S32
In step S31, the reference address ELADR of the weighting factor (w1) including the case corresponding to the exceptional processing is included based on (RL, GL, BL) extracted in step S31.
(W1) is calculated. Specifically, (Formula 57-
1) Based on 1) the reference address ELAD of the weighting factor w1
R (w1) is calculated.

In step S33, the weighting coefficient W (w1 to w4) in the triangular pyramid surrounding the conversion point P is RO.
The data is sequentially read from M12. Specifically, the weighting coefficient w1 calculated in step S32 is sequentially incremented from the reference address ELADR (w1),
The weighting factors W (w1 to w4) including the case where all the weighting factors W (w1 to w4) correspond to the exceptional processing shown in FIG.
It is sequentially read from the LUT of 4). More specifically,
In the LUT of the weighting factors W (w1 to w4) including the case corresponding to the exceptional process shown in FIG. 45, “Z (= 1 to 7) ”, the weighting factors W (w1 to w4) are sequentially read.

Note that, instead of the processing of steps S32 and S33, the weighting factor W (depending on the value "Z" indicating whether or not the exceptional processing shown in the processing of step S2 shown in FIG. You may calculate w1-w4). Specifically, depending on the case where the value “Z” indicating whether or not the exception processing is applicable is “1 to 7,” in other words, (RH, GH,
BH) depending on the above (formula 48-6) and (formula 49-
6), the weighting factor W (w1
~ W4) may be calculated.

In step S34, R of the conversion point P
Each upper 4 bits from the GB signal, that is, (RH, G
H, BH) is extracted. In addition, this upper 4 [bit]
Since the extraction processing of step S11 is the same as the step S11 shown in FIG. 12, the upper 4 bits extracted in step S11 shown in FIG. 12 may be used as they are.

In step S35, the reference address HADR (A) of the grid point "A" in the "C" component is calculated based on (RH, GH, BH) extracted in step S34. Specifically, first, FIG.
The offset "X" determined by the magnitude relation of the respective values of (RL, GL, BL) shown in 6 is read. Then, the reference address HADR (A) of the grid point “A” in the “C” component is calculated based on the above (formula 43).

The reference address HADR (A) of the grid point "A" in the "C" component is the reference address ZHADR of "Z" calculated in step S12.
(Z), that is, the reference address ZHADR of "Z" indicating whether or not the exception processing shown in (Equation 15) is satisfied.
The reference address HADR (A) of the grid point “A” in the “C” component may be calculated by multiplying (Z) by “96” and then adding the value of “16 * X”.

In step S36, grid points "A, X1, X2, H" in the triangular pyramid surrounding the conversion point P are shown.
(However, X1 and X2 are 2 of the lattice points “BG”)
In other words, two of the grid points “B to G” determined by the offset “X” in step S35) are sequentially read from the ROM 12. Specifically, the reference address HADR (A) of the grid point “A” in the “C” component calculated in step S35 is sequentially incremented to obtain the grid points “A, X in the“ C ”component.
1, X2, H ”grid point data are sequentially read from the LUT of grid point data shown in FIG.

In step S37, interpolation calculation at the conversion point P is performed. Specifically, the step S
Weighting coefficient W (w1 to w4) read in 33
And the grid point data of the grid points “A, X1, X2, H” in the “C” component read in step S36, the interpolation calculation at the conversion point P is performed by the above (formula 47). Be seen.

In step S38, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component.
If there is a next component, the process proceeds to step S39. On the other hand, when there is no next component, that is, for all "C,
When the interpolation calculation is completed for the “M, Y, K” components, this process is completed.

In step S39, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 37, for the "C, M, Y, K" components, the grid points "A, B, D, H" and the grid points "A, C, G, H", respectively. , Grid point “A, E, F, H”, grid point “A, C, D, H”, grid point “A, E, G, H”, grid point “A, B, F, H” Since the grid point data is stored, "C" calculated in step S35 is stored.
Reference address HADR of grid point "A" in the component
It simply increments sequentially from (A). Then, steps S36 to S39 are repeatedly executed, and the interpolation calculation of "M, Y, K" at the conversion point P is performed. That is, steps S36 to S39 are repeatedly executed, and "C, M, Y,
The K ”component is interpolated, and the RGB signal at the conversion point P is converted into a CMYK signal. Note that generally, since there are many pixels to be converted, this processing is repeatedly executed for all the read pixels.

As described above in detail, according to the fifth embodiment, the following operation and effect can be obtained in addition to the operation and effect of the fourth embodiment. (1) In the fifth embodiment, as shown in FIG. 45, the weighting factor W including the case corresponding to the exceptional process is w1.
They are stored in the ROM 12 in the order of to w4. Therefore, based on (Equation 57-1), the reference address ELADR of the weighting factor w1 in the case of exception processing.
If you know (w1), you can simply increment it,
Other weighting factors w2-w4 can also be obtained. Therefore,
Even if the conversion point P corresponds to exceptional processing, RGB
When performing color conversion processing from signals to CMYK signals,
High-speed color conversion processing can be performed using software. Moreover, even if the conversion point P corresponds to an exceptional process, the color conversion process can be executed without any error.

[Sixth Embodiment] ... Triangular pyramid interpolation (four-point interpolation) In this sixth embodiment, the same structure as that of the first embodiment is the same as that of the first embodiment. And the description thereof will be omitted. The sixth embodiment is similar to the fifth embodiment in consideration of the exception processing in the fourth embodiment. In addition, the sixth embodiment is different from the fifth embodiment in that the color conversion process is executed without making a judgment as to whether or not the process corresponds to the exceptional process.

In the sixth embodiment, the ROM 12
Is an LUT for determining whether or not the conversion point P shown in FIG. 9 corresponds to exception processing, and (RL, GL, B shown in FIG. 36.
L) The offset LU determined by the magnitude relationship of each value
T, the LUT of the grid point data shown in FIG. 37, and the weighting factor W (w1 to w including the case corresponding to the exceptional processing shown in FIG. 45).
4) Store the LUT.

Next, in the facsimile apparatus 1, 1 when the RGB signal of the conversion point P is output from the reading unit 14.
The operation of color conversion processing for pixels will be described with reference to the flowchart shown in FIG. In addition, this operation is RO
It is executed under the control of the MPU 11 based on the program stored in M12.

In step S41 shown in FIG. 14,
From the RGB signal at the conversion point P, each upper 4 [bit], that is, (RH, GH, BH) is extracted. Step S42
In, the reference address ZHADR (Z) of “Z” indicating whether or not the exception processing is performed is calculated based on (RH, GH, BH) extracted in step S41. Specifically, based on the above (Equation 15), the reference address Z of "Z" indicating whether or not it corresponds to exception processing.
HADR (Z) is calculated.

At step S43, the reference address ZHA of "Z" calculated at step S42.
The value of “Z” is extracted based on DR (Z). Specifically, the value of “Z” is extracted based on the LUT for determining whether or not the conversion point P shown in FIG. 9 corresponds to exceptional processing. Therefore, when the value of "Z" is "0",
It can be determined that the conversion point P does not correspond to exception processing. When the value of "Z" is "1 to 7," it can be determined that the conversion point P corresponds to exceptional processing, and
It is also possible to determine which of the seven exceptions [1] to [7] corresponds to the exception handling.

At step S44, R of the conversion point P
Each lower 4 bits from the GB signal, that is, (RL, G
L, BL) are extracted. In step S45,
Extracted in step S44 (RL, GL,
BL) based on the reference address E of the weighting factor (w1)
LADR (w1) is calculated. Specifically, the reference address E of the weighting factor w1 is calculated based on the above (formula 57-1).
LADR (w1) is calculated.

At step S46, the weighting coefficient W (w1 to w4) in the triangular pyramid surrounding the conversion point P is RO.
The data is sequentially read from M12. Specifically, the weighting coefficient w1 calculated in step S45 is sequentially incremented from the reference address ELADR (w1),
The weighting factors W (w1 to w4) including the case where all the weighting factors W (w1 to w4) correspond to the exceptional processing shown in FIG.
It is sequentially read from the LUT of 4). More specifically,
In the weighting factor W (w1 to w4) including the case corresponding to the exceptional process shown in FIG. 45, the weighting factor W corresponding to “Z (= 0 to 7)” extracted in step S43.
(W1 to w4) is read.

At step S47, R of the conversion point P
Each upper 4 bits from the GB signal, that is, (RH, G
H, BH) is extracted. In addition, this upper 4 [bit]
Since the extraction processing of step S41 is similar to that of step S41, the top 4 [b
It] may be used as it is.

In step S48, the reference address HADR (A) of the grid point "A" in the "C" component is calculated. Specifically, first, as shown in FIG. 36 (RL, G
The offset “X” determined by the magnitude relationship between the respective values of (L, BL) is read. Then, the reference address HADR (A) of the grid point “A” in the “C” component is calculated based on the above (formula 43).

The reference address HADR (A) of the lattice point "A" in the "C" component is the reference address ZHADR of "Z" calculated in step S42.
(Z), that is, the reference address ZHADR of "Z" indicating whether or not the exception processing shown in (Equation 15) is satisfied.
The reference address HADR (A) of the grid point “A” in the “C” component may be calculated by multiplying (Z) by “96” and then adding the value of “16 * X”.

At step S49, grid points "A, X1, X2, H" in the triangular pyramid surrounding the conversion point P are formed.
(However, X1 and X2 are 2 of the lattice points “BG”)
In other words, two of the grid points "B to G" determined by the offset "X" in step S48) are sequentially read from the ROM 12. Specifically, it is sequentially incremented from the reference address HADR (A) of the grid point “A” in the “C” component calculated in step S48 to calculate the grid point “A, X in the“ C ”component.
1, X2, H ”grid point data are sequentially read from the LUT of grid point data shown in FIG.

In step S50, the interpolation calculation at the conversion point P is performed. Specifically, the step S
Weighting coefficient W (w1 to w4) read in 46
And the grid point data of the grid points “A, X1, X2, H” in the “C” component read in step S49, the interpolation calculation at the conversion point P is performed by the (Equation 47). Be seen.

In step S51, it is determined whether or not there is a next component. Specifically, when the "C" component is interpolated, the "M" component is interpolated, and when the "M" component is interpolated, the "Y" component and the "Y" component are interpolated. If it does, it is determined whether or not there is a “K” component.
If there is the next component, the process proceeds to step S52. On the other hand, when there is no next component, that is, for all "C,
When the interpolation calculation is completed for the “M, Y, K” components, this process is completed.

In step S52, the reference address of the lattice point "A" in the next component is calculated. Specifically, as shown in FIG. 37, for the "C, M, Y, K" components, the grid points "A, B, D, H" and the grid points "A, C, G, H", respectively. , Grid point “A, E, F, H”, grid point “A, C, D, H”, grid point “A, E, G, H”, grid point “A, B, F, H” Since the grid point data is stored, "C" calculated in step S48
Reference address HADR of grid point "A" in the component
It simply increments sequentially from (A). Then, steps S49 to S52 are repeatedly executed, and the interpolation calculation of "M, Y, K" at the conversion point P is performed. That is, steps S49 to S52 are repeatedly executed, and “C, M, Y,
The K ”component is interpolated, and the RGB signal at the conversion point P is converted into a CMYK signal. Note that generally, since there are many pixels to be converted, this processing is repeatedly executed for all the read pixels.

As described above in detail, according to the sixth embodiment, the following operation and effect can be obtained in addition to the operation and effect of the fifth embodiment. (1) Even in the triangular pyramid interpolation, the interpolation calculation can be performed without determining whether the conversion point P corresponds to exceptional processing. Therefore, unlike the second embodiment (flowcharts shown in FIGS. 12 and 13), the same processing (flowchart shown in FIG. 14) is performed regardless of whether the conversion point P corresponds to exception processing. Weighting coefficient W (w1 to w
8) and the grid point data can be read. Therefore, even if the color conversion processing is processed by software,
The processing can be performed at a higher speed. Moreover,
Even if you configure the color conversion process with hardware,
A simpler configuration can be achieved.

The above-described respective embodiments can be modified and embodied as follows. -Various data required for the interpolation calculation are stored in the ROM 12, but in order to read them at high speed, the RAM 1 is used before the conversion process.
You may make it transfer to 3 etc. and perform a process.・ When data is transferred to the RAM 13 for processing, R
The various data stored in the OM 12 may have the data structure shown in FIG. 3 and may be transferred to the RAM 13 so as to have the data structure shown in FIG. Note that the weighting coefficient may be calculated and stored in the RAM 13 instead of being stored in the ROM 12.

In step S3 shown in FIG. 11, the weighting coefficient is read at the time of the process of step S7 in the case of the MPU having the instruction to repeatedly perform the product-sum calculation.
That is, in the process of step S7, the step is performed based on the reference address of the weighting factor calculated in step S2, the reference address of the grid point “A” in the “C” component calculated in step S5, and the number of product-sum operations. S2
The weight coefficient is read from the reference address of the weight coefficient calculated in step S5, the grid point data is read from the reference address of the grid point "A" in the "C" component calculated in step S5, multiplication is performed, and each reference address is sequentially incremented. Then, the product-sum calculation is repeated. The same applies to the flowcharts shown in FIGS. 12 and 13 and the flowchart shown in FIG.

In each of the above-described embodiments, the description has been made based on the grid point "A", but other grid points "B to H".
May be In addition, in the fourth to sixth embodiments, since only the grid points “A, H” are grid points included in the six triangular pyramids, when grids other than the grid point “A” are used as a reference, the grid points Point "H" is optimal.

In each embodiment, "C, M, Y,
Although the lattice point data is stored in the order of the “K” component, any order may be used regardless of this order. After the RGB signal is converted into the CMYK signal, the binarization process or the dither process may be continuously executed and the recording unit 15 may record the recording paper. As shown in FIG. 46, even in the case of triangular prism interpolation (6-point interpolation) using a triangular prism that divides a cube into two in the R = G plane,
The above embodiment may be applied.

As shown in FIG. 47, even in the case of quadrangular pyramid interpolation (5-point interpolation) using quadrangular pyramids in which arbitrary five points are selected from eight vertices of a cube, each of the above embodiments is also applied. good. It goes without saying that the color space conversion is not limited to the above-mentioned respective embodiments because there are several conversions such as Lab conversion and XYZ conversion other than conversion from RGB signals to CMYK signals.

Further, technical ideas other than the claims grasped from the above-described respective embodiments will be described below together with their effects. [1] In an image processing device for converting a color space, storage means for continuously storing grid point data and weighting factors of grid points surrounding the conversion point, and grid point data of reference grid points. And the reference address of the reference weighting coefficient, and by sequentially incrementing from those reference addresses, the grid point data and weighting coefficient necessary for the interpolation calculation of the conversion point are read, and the interpolation calculation of the conversion point is performed. And an image processing apparatus including means. With this configuration, color conversion processing can be performed at high speed using software.

[2] In the image processing apparatus according to any one of [1] to [3] and [1], the control means performs a binarization process or a dither process after converting the color space. An image processing device that is continuously executed. According to this structure, the value obtained by converting the color space can be recorded on the recording paper by the recording unit without temporarily storing it in a memo or the like. Therefore, a memory for storing the converted value is unnecessary, and the reading operation and the recording operation can be collectively processed.

[3] The image processing device according to claim 1 or [1], wherein the storage means stores a set of grid point data of grid points surrounding the conversion point as a set. With this configuration, it is possible to reduce the stored grid point data. Therefore, the storage capacity of the storage means can be reduced.

[4] The image processing device according to claim 2, wherein the storage means also stores which exceptional process is applicable. According to this structure, it is possible to determine which exception process is applicable.

[5] A recording medium provided in an image processing device for converting a color space, in which grid point data and weighting factors of grid points surrounding a conversion point are continuously stored in the order necessary for calculation, The grid point data of the grid point and the reference address of the standard weighting coefficient are calculated, and the reference points are sequentially incremented to read out the grid point data and the weighting coefficient necessary for the interpolation calculation of the conversion point, and the conversion is performed. A recording medium recorded as a program for causing a computer to execute a process of interpolating points. According to this structure, it is possible to provide a program capable of performing color conversion processing at high speed.

In the present specification, the term "recording medium" refers to any readable semiconductor memory device, magnetic recording device recording medium, magneto-optical recording device recording medium, etc. as long as it can record a computer program. It may be one. Specifically, it includes a semiconductor ROM, a floppy disk, a hard disk, an optical disk, a magneto-optical disk, a phase change disk, a magnetic tape and the like.

[0303]

Since the present invention is constructed as described above, it has the following effects. According to the first aspect of the invention, color conversion processing can be performed at high speed even with software.

According to the invention of claim 2, claim 1
In addition to the effect of the invention described in (1), the color conversion processing can be performed accurately without any error. According to the invention of claim 3, in addition to the effect of the invention of claim 1,
It is possible to further increase the speed.

[Brief description of drawings]

FIG. 1 is a circuit configuration diagram of a facsimile apparatus according to first to sixth embodiments.

FIG. 2 is an explanatory diagram for explaining grid points in RGB space.

FIG. 3 is an explanatory diagram for explaining an LUT of grid point data in cubic interpolation.

FIG. 4 is an explanatory diagram for explaining a cube surrounding a conversion point P in RGB space.

FIG. 5 is an explanatory diagram for explaining an LUT of grid point data in cubic interpolation.

FIG. 6A is an explanatory diagram for explaining the principle of interpolation in cubic interpolation. (B) Explanatory drawing for demonstrating the weighting factor in cubic interpolation.

FIG. 7 is an explanatory diagram for explaining an LUT of weighting factors in cubic interpolation.

FIG. 8 is an explanatory diagram for explaining a case where a conversion point P in RGB space corresponds to exceptional processing.

FIG. 9 is an explanatory diagram illustrating an LUT for determining whether or not a conversion point P corresponds to exceptional processing.

FIG. 10 is a weighting coefficient L including a case corresponding to exceptional processing.
Explanatory drawing for demonstrating UT.

FIG. 11 is a flowchart showing an operation when performing color conversion processing in the first and fourth embodiments.

FIG. 12 is a flowchart showing an operation when performing color conversion processing in the second and fifth embodiments.

FIG. 13 is a flowchart showing an operation when performing color conversion processing in the second and fifth embodiments.

FIG. 14 is a flowchart showing an operation when performing color conversion processing in the third and sixth embodiments.

FIG. 15A is an explanatory diagram for explaining a method of dividing a cube into six parts. (B) Explanatory drawing for demonstrating the triangular pyramid obtained by dividing into 6 parts.

FIG. 16A is an explanatory diagram for explaining a triangular pyramid in the case of BL>GL> RL. (B) Explanatory drawing for demonstrating the triangular pyramid in case of GL>RL> BL. (C) Explanatory drawing for demonstrating a triangular pyramid in case of RL>BL> GL. (D) Explanatory drawing for explaining a triangular pyramid in the case of GL ≧ BL ≧ RL (RL ≠ GL). (E) Explanatory drawing for explaining a triangular pyramid in the case of RL ≧ GL ≧ BL. (F) Explanatory drawing for demonstrating the triangular pyramid in case of BL> = RL> = GL (GL <= BL).

FIG. 17 is an explanatory diagram for explaining a triangular pyramid in the case of BL>GL> RL.

FIG. 18A is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of BL>GL> RL. (B) Explanatory drawing for demonstrating the volume of the triangular pyramid Vb in case of BL>GL> RL. (C) Explanatory drawing for demonstrating the volume of the triangular pyramid Vc in case of BL>GL> RL. (D) Explanatory drawing for demonstrating the volume of the triangular pyramid Vd in case of BL>GL> RL.

FIG. 19 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Vc in the case of BL>GL> RL. (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 20 is an explanatory diagram for explaining a triangular pyramid in the case of GL>RL> BL.

FIG. 21 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of GL>RL> BL. (B) Explanatory drawing for demonstrating the volume of the triangular pyramid Vb in case of GL>RL> BL. (C) Explanatory drawing for demonstrating the volume of the triangular pyramid Vc in case of GL>RL> BL. (D) Explanatory drawing for demonstrating the volume of the triangular pyramid Vd in case of GL>RL> BL.

FIG. 22 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Vc in the case of GL>RL> BL. (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 23 is an explanatory diagram for explaining a triangular pyramid in the case of RL>BL> GL.

FIG. 24A is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of RL>BL> GL. (B) Explanatory drawing for demonstrating the volume of the triangular pyramid Vb in case of RL>BL> GL. (C) Explanatory drawing for demonstrating the volume of the triangular pyramid Vc in case of RL>BL> GL. (D) Explanatory drawing for demonstrating the volume of the triangular pyramid Vd in the case of RL>BL> GL.

FIG. 25 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Vc in the case of RL>BL> GL. (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 26 is an explanatory diagram for explaining a triangular pyramid in the case of GL ≧ BL ≧ RL (RL ≠ GL).

FIG. 27 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of GL ≧ BL ≧ RL (RL ≠ GL). (B) Explanatory drawing for demonstrating the volume of the triangular pyramid Vb in case of GL> = BL> = RL (RL <= GL). (C) An explanatory view for explaining the volume of the triangular pyramid Vc in the case of GL ≧ BL ≧ RL (RL ≠ GL). (D) An explanatory diagram for explaining the volume of the triangular pyramid Vd in the case of GL ≧ BL ≧ RL (RL ≠ GL).

FIG. 28 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Vc in the case of GL ≧ BL ≧ RL (RL ≠ GL). (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 29 is an explanatory diagram for explaining a triangular pyramid when RL ≧ GL ≧ BL.

FIG. 30 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of RL ≧ GL ≧ BL. (B) Explanatory drawing for demonstrating the volume of the triangular pyramid Vb in case of RL> = GL> = BL. (C) Explanatory drawing for explaining the volume of the triangular pyramid Vc in the case of RL ≧ GL ≧ BL. (D) Explanatory drawing for demonstrating the volume of the triangular pyramid Vd when RL ≧ GL ≧ BL.

FIG. 31A is an explanatory diagram for explaining the volume of the triangular pyramid Vc when RL ≧ GL ≧ BL. (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 32 is an explanatory diagram for explaining a triangular pyramid in the case of BL ≧ RL ≧ GL (GL ≠ BL).

FIG. 33 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of BL ≧ RL ≧ GL (GL ≠ BL). (B) An explanatory diagram for explaining the volume of the triangular pyramid Vb in the case of BL ≧ RL ≧ GL (GL ≠ BL). (C) An explanatory view for explaining the volume of the triangular pyramid Vc in the case of BL ≧ RL ≧ GL (GL ≠ BL). (D) Explanatory diagram for explaining the volume of the triangular pyramid Vd in the case of BL ≧ RL ≧ GL (GL ≠ BL).

FIG. 34 (a) is an explanatory diagram for explaining the volume of the triangular pyramid Vc when BL ≧ RL ≧ GL (GL ≠ BL). (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 35 is an explanatory diagram illustrating an LUT of weighting factors in triangular pyramid interpolation.

FIG. 36 is an explanatory diagram for explaining an LUT that stores an offset determined by the magnitude relationship of each value of (RL, GL, BL).

FIG. 37 is an explanatory diagram illustrating an LUT of grid point data in triangular pyramid interpolation.

FIG. 38 is an explanatory diagram illustrating an LUT of grid point data in triangular pyramid interpolation.

FIG. 39 is an explanatory diagram for explaining a triangular pyramid in the case of RH = 15, GH <15, BH <15 and in the case of RL ≧ GL ≧ BL.

FIG. 40 (a) RH = 15, GH <15, BH <15
2 is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of RL ≧ GL ≧ BL. (B) Triangle pyramid V in the case of RH = 15, GH <15, BH <15, and in the case of RL ≧ GL ≧ BL
Explanatory drawing for demonstrating the volume of b. (C) Triangular pyramid V when RH = 15, GH <15, BH <15 and when RL ≧ GL ≧ BL
Explanatory drawing for demonstrating the volume of c. (D) Triangular pyramid V in the case of RH = 15, GH <15, BH <15, and in the case of RL ≧ GL ≧ BL
Explanatory drawing for demonstrating the volume of d.

FIG. 41 (a) RH = 15, GH <15, BH <15
2 is an explanatory diagram for explaining the volume of the triangular pyramid Vc in the case of RL ≧ GL ≧ BL. (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 42 is an explanatory diagram for explaining a triangular pyramid in the case of RH <15, GH = 15, BH = 15 and in the case of RL>BL> GL.

FIG. 43 (a) RH <15, GH = 15, BH = 15
FIG. 6 is an explanatory diagram for explaining the volume of the triangular pyramid Va in the case of RL>BL> GL. (B) Triangle pyramid V in the case of RH <15, GH = 15, BH = 15 and in the case of RL>BL> GL
Explanatory drawing for demonstrating the volume of b. (C) Triangular pyramid V when RH <15, GH = 15, BH = 15 and when RL>BL> GL
Explanatory drawing for demonstrating the volume of c. (D) Triangular pyramid V in the case of RH <15, GH = 15, BH = 15 and in the case of RL>BL> GL
Explanatory drawing for demonstrating the volume of d.

FIG. 44 (a) RH <15, GH = 15, BH = 15
2 is an explanatory diagram for explaining the volume of the triangular pyramid Vc in the case of RL>BL> GL. (B) The side view which looked at the triangular pyramid Vc shown in (a) from the arrow direction.

FIG. 45 is a weighting coefficient L including a case corresponding to exceptional processing;
Explanatory drawing for demonstrating UT.

FIG. 46 is an explanatory diagram showing an example of triangular prism interpolation (6-point interpolation) as another embodiment.

FIG. 47 is an explanatory diagram showing an example of quadrangular pyramid interpolation (5-point interpolation) as another embodiment.

[Explanation of symbols]

DESCRIPTION OF SYMBOLS 1 ... Facsimile machine as an image processing apparatus, 11 ... MPU which comprises a control means, 12 ... ROM which comprises a memory means and a control means, 13 ... RAM which comprises a control means.

─────────────────────────────────────────────────── ─── Continued Front Page (58) Fields surveyed (Int.Cl. ⁷ , DB name) G06T 1/00 510 G06T 11/60 120 G09G 5/06 H04N 1/46 H04N 1/60 H04N 9/64 JISST File (JOIS)

Claims (3)

(57) [Claims] 1. In an image processing device for converting a color space, a storage means for continuously storing at least grid point data of grid points surrounding a conversion point in an order necessary for calculation, and grid point data at a reference grid point. The image processing apparatus includes a control unit that calculates the reference address of, and sequentially increments from the reference address to read grid point data necessary for the interpolation calculation of the conversion point, and perform the interpolation calculation of the conversion point. 2. The image processing apparatus according to claim 1, wherein the storage means stores whether or not the exceptional processing is applied,
The control means is an image processing device for performing interpolation calculation of conversion points depending on whether the exception processing is applied or not. 3. The image processing apparatus according to claim 1, wherein the control means performs the interpolation calculation of the conversion points without determining whether or not the exception processing is applicable.