JPH0366288A - Picture processing unit - Google Patents

Abstract

PURPOSE:To improve interpolation arithmetic accuracy and to reduce the cost of the unit by storing a data being the result of color correction being combination of high-order n-bit smaller than m-bit input data into a table memory and generating an interpolation data with the interpolation method according to an algorithm. CONSTITUTION:Low-order 2 bits of input data C, M, Y and SEL1, SEL2 signals from a control circuit 61 are inputted to a ROM 11. At first high-order 6 bits of the input data and outputs from a ROM 11 are added by adders 21, 22, 23 based on an address of a 1st selection table when both the SEL1, SEL2 signals are at a low level, and inputted to a table memory 31. A data read from a table memory 31 is latched in a latch 41 by using an LCK1 from the control circuit 61. An adder 51 calculates a means value of output data from latches 41-44 and outputs an obtained data C'. Signals M', Y' are obtained by the similar procedure to above.

Description

DETAILED DESCRIPTION OF THE INVENTION [Industrial Field of Application] The present invention is expressed in color printers, color copying machines, etc., and the quadratic masking operation is expressed by equation (2). Input data (C, M, Y) is corrected to (C', M', Y') by masking operation.

As a method of realizing color correction, the calculation result of equation (1) or (2), or the output data that is directly correlated to the input data, is a combination of all the gradations of the input data (C1M, Y). There is a way to store it in table memory. However, with this method, C, M, and Y each have 8
If the input data is bits, a large amount of memory of 224 x 3 bytes is required, which significantly increases the cost of the device. As a method to improve this point, Tokuko Sho 5216
No. 403 describes a method in which the capacity of the table memory is reduced and combinations of input data that cannot be stored in the table memory are interpolated.

[Problem to be Solved by the Invention 1] In the above interpolation method, when calculating output data C' for input data (C, M, Y), interpolation is performed by focusing only on the value of C. The accuracy of calculation becomes very poor. If the table corresponding to the input data (C1M, Y) is (C,, Mr, Y,), and the table one level above the input data is (cr, +, Mr+1, Yr-1), then the interpolation calculation is It is obtained from the relationship between C, Cr, and Cr. Therefore, the input data M and Y are Mr≦M≦M r
+ + + Y If the range is r≦Y≦Yr++, all values will be the same. The same applies when determining M' and Y'.

Further, a multiplier is required to calculate the ratio of C to Cr and Cr + 1, making the device expensive.

[Means for Solving the Problems] The present invention has been made to solve the above-mentioned problems. Therefore, only the correction results obtained by combining the upper n bits of input data of three colors, which are smaller than the number of bits, are stored in a table memory. , for combinations of input data that are not stored in the table memory, 2f'''-n1 types of data are read out from the table memory according to a certain algorithm, and the average value is calculated to generate interpolated data.

[Operation 1] In the present invention, a table of combinations of input data of the upper n bits of each color is created by calculating interpolated data by calculating several types of data read from the table memory, which are combinations of input data that are not stored in the table memory. Even if only a memory is provided, masking processing can be performed with accuracy comparable to the case where a table memory is provided for all combinations of m-bit input data for each color.

[Embodiment] FIG. 1 is a block diagram for implementing the present invention. In this embodiment, when m=8 and n=6, each of C, M, and Y is 8
This is a case where the color correction result of a combination of bit (0 to 255) input data in which the lower two bits are ○O is stored in the table memory. First, the interpolation algorithm of the present invention will be explained. FIG. 2 is a spatial coordinate diagram of combinations of input data stored in a table memory.

The 8 bits of input data (0 to 255) are converted to the upper 6 bits (O
~63) and the lower 2 bits (0~3), the input data is: C=4CI + i fc+ :Q~63.1.0~3
)M "4M+" J (IL, 0-63, j:
0-3) Y = 4Yl + k (y, : 0-63,
k: 0 to 3). Here C,,M,,
Y is the upper 6 bits, and i and j are the lower 2 bits.

If 1=j=and=Q, the data stored in the table memory can be read. For other combinations of i, j, data is generated using the compensation method described below.

Figure 3 is derived from the spatial coordinates of Figure 2 (C, M, Y) = (
80,40,12) and (C,M,Y) = (84,4
4, 16) and illustrates the following specific example.

When i, j, l<2, C=2(CI+1
), M=2 (M,+1), Y=2 (Y+1) and a table where 1=j= and=Q are selected, and the intermediate value of the two table decks is calculated. For example, if (C, M, Y) = (82, 42, 16), table (80, 40, 16) is changed to table (84, 44,
16) is selected, data is read from the two tables, and an intermediate value of the read data is calculated.

Next, when i, j, and k are 1, find the calculation result when i, j, and k are 2 using the method described above when i, j, and k are 2, and combine the calculation result and 1= Calculate the intermediate value of the table data when =j=kO. Also, i, j, k are 3
When , C=2 (C+ +1) and M= (
The intermediate value of the table value with M, +1) and Y=2 (Y, +1) and the calculation result when i, j, and k are set to 2 is calculated.

For example, if the input data is (C, M, Y) = (83, 42,
13), (4X20+3.4XlO+2.4X3
+1). First, using the above method, i, j,
When k is processed from 1 to 3, the value of (4x20+2.4X 10+2.4x3)■ and (4X21, 4X10+2, 4X3
+2)...The intermediate value of the value of ■ may be calculated and shifted. Next, the above 12, j2, k
When processing is performed when 2 is 2, the value of ■ is shown in the table (80
, 40, 12) and table (84, 44, 12), and the value of
It can be determined from the table data. Furthermore, by calculating the intermediate value between ■ and ■, the input data (83, 4
2, 13) color correction results can be obtained.

In the case of m - n = 2, when i, j, and k are a combination of 1 or 3, three calculations are required to obtain the intermediate value, and a maximum of 4
I need to read data from two tables. If the data read from the four tables are A, B, C, and D, the three-color correction data E to be obtained is = -(A + B + C + D), which means that it is sufficient to calculate the average value of the four data. .

In this embodiment, the device was realized by reading data from the table four times for all combinations of i, j, and calculating the average value of the four data. 1=j
Even when = and =o, the same table is read and calculated four times. FIG. 4 shows four types of tables read from the table memory for combinations of ll, j, and .

For example, when (i, j, k) = (0, 3, 1),
(4C1,4M+, 4Yl), (4CI, 4M
, +4.4Yl +4) and (4C4Ml +
4.4Y1) table is read twice.

In Figure 1, RQMI l is ○ when 4C+, 4M+, or 4Y is selected for the combination of i, J, and l when 4C1+4, 4M, +4 or 4Y1+4 is selected. is memorized.

FIG. 5 is a timing chart. Figures 1 and 5 below
The operation of obtaining C' from input data (C, M, Y) will be explained with reference to the figure.

The lower two bits of each of the input data C, M, and Y and the 5ELL and 5EL2 signals from the control circuit 61 are input to the ROMII, and a 3-bit data is outputted to determine the selection table. First, when both the 5ELL and 5EL2 signals are LOW, the address value of the first selection table is the upper 6 bits of the input data and the output from the ROMII is added to the adder 21.2.
2.23 and input into the table memory 31.

Data read from the table memory 31 is latched into the latch 41 by the LCK 1 signal from the control circuit 61. The data read out from the table memory 31 in a similar manner is transmitted from the control circuit 61 to LCK2, LCK3,
LCK4 is latched by latch 42, latch 43, and latch 44, respectively. Adder 51 includes latches 41 to 4
The average value of the output data of 4 is calculated and the desired data C' is output. M' and Y' can also be found using the same procedure as above.

Furthermore, if the interpolation method of the present invention is used, color correction results for combinations where the lower 3 bits are OOO or lower 4 bits are oooo out of each 8-bit C, M, and Y input data can be stored in the table memory using the same algorithm. By storing the selection table information in the ROM, each color has 8
Color correction processing can be performed with accuracy comparable to when all color correction results of bit combinations are stored in a table memory.

[Effects of the Invention] By using the above-described interpolation method, the present invention can significantly reduce the capacity of the table memory, and still achieve the same accuracy as when the table memory is provided for all the combinations of input data of m bits for each color. You can perform color correction processing in real time.

Further, according to the present invention, since the device can be constructed without using a multiplier, the cost of the device can be kept low.

[Brief explanation of drawings]

Figure 1 is a block diagram of an embodiment according to the present invention, Figure 2 is an explanatory diagram showing spatial coordinates, Figure 3 is an explanatory diagram showing spatial coordinates of a specific example, and Figures 4 (a) and (b) are selections. An explanatory diagram showing a table, FIG. 5 is a timing chart according to the present invention. l 1 ・ ・ ・ ・ ・ ・ ・ ・ ・ ・ ・
ROM21.22.23... Adder 31... Table ROM1 2 41.42.43.44... Latch circuit 51...
・・・・・・Adder 61・・・・・・・・・・・・More than control circuit

Claims (1)

[Claims] The first storage unit stores color-corrected output data C', M', Y' of combinations of upper n (m>n) bits for m bits of color separation input data C, M, and Y for each color of a color image.
a second table memory storing information on a memory table to be selected for a combination of lower (m−n) bits of each color input data of C, M, and Y;
2^(^m^-^) read from the first table memory
n^) comprises means for calculating the average value of the types of data;
An image processing device characterized by generating interpolated data by the above means.