JP3503711B2 - Raster / block conversion method and apparatus for implementing the method - Google Patents

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a raster / block conversion method that can be used for image compression processing and the like, and an apparatus for implementing this method.

[0002]

2. Description of the Related Art In conventional image processing, it has been a mainstream to perform data processing for each pixel by using image data as a raster (image data in a horizontal direction arranged in line order).
However, with recent improvements in image compression technology and the like, it is becoming more common to perform image processing on blocks (horizontal × vertical block-shaped image data groups) than on the raster. . In particular, for color still images, the JPEG system has been standardized to perform image processing in blocks, so that the importance of raster / block conversion devices is increasing.

[0003]

By the way, in the conventional raster / block conversion device, when the conversion process is executed by software, the image data for one screen is stored in the frame memory to generate a necessary address. Therefore, when executing by hardware, the address is specified by the screen size.

FIG. 14 is a flow chart showing the processing contents when performing block-to-raster conversion using a frame memory which stores image data for one screen.

In this flow chart, A is an address and B is a value of the number of horizontal pixels / 8 (assuming 8 × 8 blocks). First, initialization is executed (step 71), and then it is determined whether or not to count up based on the count UP factor (see the embodiment of the present invention) (step 72). When counting up, it is determined whether the value of C matches 8 (step 73). C
If the value of is equal to 8, D = D + 1, A = A +
(8 × D), and processing of C = 0 is executed (step 7
4). On the other hand, if they do not match, A = A + B, and C
= C + 1 is executed (step 75).

However, in the above raster / block conversion, it is necessary to provide a frame memory for one screen, so the cost becomes high. Further, in hardware, the address is designated by the screen size, and the larger the screen size, the larger the memory required and the higher the cost. Further, it is necessary to generate different addresses if the screen size is different, but this cannot be dealt with.

In view of the above circumstances, the present invention can perform raster / block conversion without providing a memory having a large memory capacity, and can arbitrarily change the arithmetic coefficient that changes depending on the screen size. It is an object of the present invention to provide a raster / block conversion method capable of automatically generating a signal and a device for implementing this method.

[0008]

According to the raster / block conversion method of the present invention, A is the current address coordinate, A'is the next address coordinate, B is the addition value, and C is the number of horizontal pixels. When χ is the number of pixels in the horizontal direction of one block, X is a value obtained by the calculation of C / χ, and Y is the number of conversions, X ^Y > (C-1) is satisfied, then X ^Y Let B be the remainder in the formula of / (C-1), and X ^Y
When the relation of> (C-1) is not established, the calculation result of X ^Y is set to B, and the following address is calculated based on the above-obtained B by the formula of A '= A + B, and the raster data is converted into block data. On the other hand, when the relation of χ ^Y > (C-1) is established, the remainder in the formula of χ ^Y / (C-1) is B
And if the relationship of χ ^Y > (C-1) does not hold, χ
^Let B be the calculation result of ^Y , and based on B obtained above, A ′ =
It is characterized in that the next address is obtained by the formula A + B and the block data is converted into raster data.

The value of X ^Y or χ in the above method
^Y is obtained by using a value obtained by sequentially adding the same value until the number of counts reaches X or χ, and at the time of the next conversion process after the conversion process for all pixels in the memory determined by the block size is completed, The values that are sequentially added may be set to the same value.

Further, the raster / block conversion method of the present invention may be characterized in that the variables are appropriately changed in accordance with the size of the block.

Further, the raster / block conversion device of the present invention, in order to carry out the raster / block conversion method, adds means for sequentially adding the same value until the count reaches X or χ, and the size of the block. And latch means for setting the sequentially added values to the same value in the next conversion process after the conversion process for all the pixels in the memory determined by.

[0012]

According to the first configuration, raster / block conversion can be performed if there is a memory capacity of the number of horizontal pixels × the number of pixels on one side of the block. Further, it is possible to automatically generate a memory address for raster / block conversion in an arbitrary screen size set externally. Moreover, the raster → block conversion and the block → raster conversion can be performed by the same arithmetic expression (only the coefficients are different).

According to the second and fourth configurations, it is possible to generate an address for conversion by addition processing without performing multiplication processing.

According to the third configuration, the coefficient is arbitrarily changed according to the size of the block.
Not only blocks of 8 pixels, but for example 16x16
Blocks of pixels can be easily accommodated.

[0015]

【Example】

(Embodiment 1) An embodiment of the present invention will be described below with reference to the drawings. In this embodiment, software performs the raster / block conversion.

(Raster-to-block conversion) First, the conversion process from raster to block (8 × 8) will be described. The raster-to-block conversion process is performed by generating a read / write address for the memory as follows.

The read / write address A'is a current address A and an added value B given by the following calculation.
It is obtained by adding and. The added value B is X ^Y
When the relationship of> (C-1) is established, X ^Y / (C-
1) The remainder value (which can be represented by ^XY % (C-1) as described later), and is the ^XY value when the relationship of ^XY > (C-1) is not established. . The X is calculated by calculating C / χ. C is the number of pixels in the horizontal direction, and χ
Is the block size (number of horizontal pixels in the block)
And Y is the number of conversions. In this embodiment, since 8 × 8 blocks are used, χ = 8.

The above relationship and each variable are summarized as follows.

[0019]

## EQU1 ## When the relationship of ^XY > (C-1) is established, B = ^XY % (C-1) When the relationship of ^XY > (C-1) is not established, B = ^XY A '= A + B A = current address coordinate A ′ = next address coordinate B = addition value C = horizontal pixel number χ = block size (χ = 8 in this embodiment) X = C / χ Y = number of conversions

FIG. 1 is a flowchart showing a raster-to-block conversion process. First, in step 1, each variable is initialized. Next, it is determined whether to count up based on the count-up determination factor (step 2). The factors that determine the count-up mean the following cases. For example, when a block is composed of 8 × 8 and eight addresses in a row in this 8 × 8 block are generated, these eight addresses are “000 to 000” in the lower 3 bits of the address signal.
111 ", and it is necessary to carry to the 4th bit in order to generate the next address, but the point at which this carry is necessary is the count-up factor.

When it is judged in step 2 that the count-up should be executed, B = X ^Y is calculated (step 3). Next, it is determined whether B> (C-1) (step 4). If YES in step 4, X ^Y >
Since the relationship (C-1) is established, B = B%
The calculation of (C-1) is executed (step 5). Since B = X ^Y is calculated in step 3, the calculation in step 5 means that the remainder of X ^Y / (C-1) is B. Then, using the B thus calculated, the next address A'is obtained by the calculation process of A = A + B and the process of A '= A (step 6).

On the other hand, if NO in step 4,
Since it is a case where the relationship of ^XY > (C-1) is not established,
Using B (B = ^XY ) obtained in step 3, the calculation of A = A + B and A '= A are executed as they are (step 6).

Reading / writing from / to the memory is executed by the next address A'obtained as described above.

After the processing of step 6 above, it is determined whether or not the value of A'is equal to the value of (C-1) (step 7). If they do not match, the process proceeds to step 2, while if they match, the process of A ′ = 0 and A = 0 is executed, and the process of incrementing Y is executed (step 8), and the process proceeds to step 2.

By executing the above processing, the process shown in FIG.
As shown in FIG. 4, data is stored in the memory. It should be noted that this figure shows the contents of the 8 × 56 line memory, and the numbers in the 8 × 56 frame indicate the numbers of the pixels lined up in line order. Also,
2 shows the memory contents at the 0th time (Y = 0), FIG. 3 shows the memory contents at the 1st time (Y = 1), and FIG. 4 shows the memory contents at the 2nd time (Y = 2). Shows.

Further, in this example, the number of pixels in the horizontal direction is 56, so that C = 56, and since the block is 8 × 8, X = 7 (56/8). In this example,
Although an example using a line memory is shown in this way, a single screen memory can also be used in this method.

Next, the contents of the processing will be concretely described with reference to the flowchart of FIG. First, in the calculation performed in the first step 3 after initialization, since Y = 0, X ^Y = 7
⁰ = 1. Therefore, X ^Y > (C-1) does not hold,
In step 4, NO is determined, and the process proceeds to step 6. Since B = 1 has been obtained by the process of step 3, the next address A'is calculated by the process of A = A + 1 and the process of A '= A in step 6. The above processing is repeatedly executed until A = (56-1).

Here, a total of 8 addresses from 0 to 7 can be performed by sequentially incrementing the lower 3 bits of the address signal from 000 to 111.
In the next eight addresses in total, 8 to 15, the fourth bit is "1" and the 3-bit signal is 000 to 111.
It will be sufficient to increment in order. For the subsequent generation of eight addresses, similarly, carry may be sequentially executed at the fourth bit and above. The process of sequentially carrying the digits in the fourth bit and higher is executed by making YES in step 2 each time a count-up factor occurs and going through step 6.

As described above, A'obtained by A = A + 1 and A '= A corresponds to the value of the fourth bit or more, and the lower 3 bits of this A'are 000. ~
Since the value increases by 1 every time 111 is repeated, the data is stored in the memory address order as shown in FIG. Specifically, when the value of A is "0", the address is 0 to 7, and when it is "1", the address is 8 to 15, and the data is stored. Note that, in the state shown in FIG. 2, since no data is initially stored in the memory, only the writing of data is performed by the address generated as described above.

When A '= (56-1) in step 7, the process proceeds to step 8. The above "56" is "111000" when expressed in binary. As described above, A'corresponds to the value of the fourth bit and above. "1110
Since 00 "is the value of the 4th to 9th bits as A ', the decimal value is" 448 ". That is, the address when A' = (56-1) is the memory. The last address is 447, which means that the processing of the entire memory has been completed.

Then, in step 8, A and A'are cleared and Y is incremented, and step 2 is executed again.
Will proceed to.

When proceeding to step 2 through step 8 described above, since Y = 1, X ^Y = 7 ¹ = 7.
Also in this case, in steps 3 and 4, X ^Y > (C−
1) does not hold and B = ^XY = 7. Therefore, in step 6, the next address is calculated by calculating A = A + 7 and processing A '= A.

As described above, A'obtained by A = A + 7 and A '= A corresponds to the value of the fourth bit or more, and the lower 3 bits of this A'are 000. ~
Since it increases by 7 every time 111 is repeated, as shown in FIG. 3, every 56 addresses are generated in the memory, and the next data is stored at that address.

Specifically, when the value of A is "0", 0
At address 7, 0 to 7 in the next data, when it is "7", at addresses 56 to 63, 8 to 15 in the next data,
Data is stored at the generated address. That is, 0 to 7, 56 to 63, 112 in FIG.
.. 119, ..., 392 to 399 data are read out, and at the same time, as shown in FIG.
0-7, 8-15, 16-23, ... in the following data
.., 56 to 63, that is, eight pieces of data are written.

Here, for a total of eight addresses obtained by adding 0 to 7 to a certain address, the fourth and higher bits correspond to the certain address, and the lower 3 bits are from 000 to 111. It can be generated by sequentially incrementing. In the case of FIG. 3,
The data of 0 to 7 are at addresses 0 to 7, the data of 8 to 15 are at addresses of 56 to 63, and the certain address is "0" or "5" in this case.
6 ”,“ 112 ”, ...,“ 392 ”.

As described above, A'obtained by the calculation of A = A + 7 and the process of A '= A corresponds to the fourth and higher bits. When A ′ = 7, this 7
When expressed in binary, it is "111".
Is applied to the 4th to 6th bits, this "111" means "56" in decimal. Also, A '=
When it is 14, if this 14 is expressed in a binary number, it is “111
It is 0, and if this "1110" is applied to the 4th to 7th bits, this "1110" means "112" in decimal.

That is, the above calculation of A = A + 7 and A '= A
A'obtained by the above process represents the above-mentioned certain addresses "0", "56", "112", ..., "392". Such an address (start address)
And the addition of the lower 3 bits added thereto, the 8 × 8 block data is read as shown in FIG.

Next, in step 7, A '= (56-
When 1) is reached, the process proceeds to step 8. The above "56" is
When expressed in binary, it is “111000”. As described above, A'is a value of the fourth bit or higher. "11100
Since 0 "is the value of the 4th to 9th bits as A ', its decimal value is" 448 ". That is, the address when A' = (56-1) is the memory. The last address is 447, which means that the processing of the entire memory has been completed.

Then, in step 8, A and A'are cleared and Y is incremented, and the process proceeds to step 2 again.

When proceeding to step 2 through step 8 described above, since Y = 2, X ^Y = 7 ² = 49. Also in this case, in steps 3 and 4, X ^Y > (C
-1) does not hold, and B = ^XY = 49. Therefore, in step 6, the next address is calculated by calculating A = A + 49 and processing A '= A.

As described above, A'obtained by A = A + 49 and A '= A corresponds to the value of the fourth bit or more, and the lower 3 bits of this A'are 000.
Each time ~ 111 is repeated, it increases by 49, so as shown in FIG. 4, memory addresses are generated in 392 jumps.
The next data will be stored at that address.

Specifically, when the value of A is "0", it is 0
The data is stored in the generated addresses such as 0 to 7 in the next data at address 7, 8 to 15 in the next data at addresses 56 to 63 when "49". That is, 0-7, 56-63, 112- in FIG.
At the same time that the data 119, ..., 392 to 399 are read out, as shown in FIG. 4, 0 to 7, 8 to 15, 16 to 2 in the next data are stored in the read address.
The data of 3, ..., 56-63 will be written.

Next, when proceeding to step 2 through step 8, since Y = 3, X ^Y = 7 ³ = 343. In this case, in steps 3 and 4, ^XY >
Since (C-1) is established, B = ^XY / (C-1) = 1
It becomes 3. Therefore, in step 6, A = A + 13
And the processing of A ′ = A, the next address is calculated.

After that, the same processing is executed, and as described above, the conversion processing from raster to block is executed for the data of one screen by using the line memory.

(Block → Raster Conversion) Next, the conversion processing from block (8 × 8) to raster will be described. The block-to-raster conversion process is performed by generating a read / write address for the memory as follows.

The read / write address to be calculated is A '.
It is calculated by adding the current address A and the addition value B given by the following calculation. The added value B is the remainder of χ ^Y / (C-1) when the relation of χ ^Y > (C-1) holds, and χ ^Y > (C-1)
If the relation of does not hold, then χ ^Y. C is the number of pixels in the horizontal direction, χ is the size of the block, and Y is the number of conversions. In this embodiment, since 8 × 8 blocks are used, χ = 8.

The above relationship and each variable can be summarized and shown as follows. The definition of the variable is the same as in the case of equation 1.

[0048]

[Formula 2] When the relationship of χ ^Y > (C-1) is established, B = χ ^Y % (C-1) When the relationship of χ ^Y > (C-1) is not established, B = χ ^Y A ′ = A + B

FIG. 5 is a flow chart showing the conversion processing from block to raster. First, in step 11, initialization processing of each variable is performed. Next, it is determined whether to count up (step 12). The factors that determine this count-up are the same as in the case of the flowchart of FIG.

When it is judged in step 12 that the count-up should be executed, B = χ ^Y is calculated (step 13). Next, it is determined whether B> (C-1) (step 14). If YES in step 14, it means that the relation of χ ^Y > (C-1) is established, and therefore the calculation of B = B% (C-1) is executed (step 1
5). In this calculation, since B = X ^Y is calculated in step 13, it means that the remainder of χ ^Y / (C−1) is B. Then, using the B thus calculated, the next address A'is obtained by the calculation process of A = A + B and the process of A '= A (step 16).

On the other hand, if NO in step 14, it means that the relationship of χ ^Y > (C-1) is not established. Therefore, B (B = χ ^Y ) obtained in step 13 is used as it is. The calculation of A = A + B is executed (step 16).

Reading / writing from / to the memory is executed by the next address A'obtained as described above.

After the processing of step 6 above, it is judged whether or not the value of A'is equal to the value of (C-1) (step 7). If they do not match, the process proceeds to step 2. On the other hand, if they do match, the process of A '= 0 and A = A' is executed, and the process of incrementing Y is executed (step 8), and the process proceeds to step 2.

By executing the above processing, FIG.
As shown in FIG. 8, data is stored in the memory.

FIG. 6 shows the memory contents at the 0th time (Y = 0). At the 0th time, A = A + 1 and A ′ =
Since the next address A ′ is obtained by A, the same address generation as the 0th time in the raster → block conversion is executed.

FIG. 7 shows the memory contents at the first time (Y = 1). At the first time, the next address A'is obtained by A = A + 8 and A '= A, so the raster →
With respect to the address generation by the process of A = A + 7 and A ′ = A in the block conversion, an address increased by 8 is generated.

FIG. 8 shows the memory contents at the time of the second time (Y = 2). At the first time, the next address A'is obtained by A = A + 9 and A '= A, so the raster →
With respect to the address generation by the processing of A = A + 7 and A ′ = A in the block conversion, the address generation increased by 16 is performed.

There is no correspondence between the memory contents shown in FIGS. 6 to 8 and the memory contents shown in FIGS.

With the above configuration, the number of horizontal pixels ×
Raster data can be converted into block data and block data can be converted into raster data with the memory capacity of the number of pixels on one side of the block. Further, as in the above embodiment, not only the 8 × 8 blocks but also blocks of any size and vice versa can be converted. For example, the value of X in the expression of Equation 1 is set to the number of horizontal pixels / 7,
By setting χ in the expression of Equation 2 to 7, conversion into a 7 × 7 block and its inverse conversion can be performed. In addition,
When χ is set to 7, the horizontal pixel number C is a multiple of 7 such as 49. Furthermore, 7 × 8
It can also be applied to such blocks.

(Embodiment 2) An embodiment of the present invention will be described below with reference to the drawings. In this embodiment, a case where the raster / block conversion is executed by hardware will be described.

FIG. 9 shows that the raster
It is the flowchart which showed the processing content in the case of performing block conversion. In the above conversion processing by hardware, it is configured by an adder / subtractor without using a multiplier. The processing content of FIG. 9 shows a case where the above conversion processing is performed using an adder / subtractor.

FIG. 10 is a flow chart showing the processing contents when the block → raster conversion is executed by hardware. In the above conversion processing by hardware, it is configured by an adder / subtractor without using a multiplier. Figure 1
The processing content of 0 indicates a case where the above conversion processing is performed using an adder / subtractor.

FIG. 11 is a circuit diagram showing hardware for executing the processing contents of FIGS. 9 and 10.

Note that, although the respective codes in FIG. 9, FIG. 10, and FIG. 11 have a corresponding relationship, they do not have a corresponding relationship with the codes used in the first embodiment. In the present embodiment, the code is defined as follows.

A: Number of pixels in horizontal direction / size of block B: Number of pixels in horizontal direction C: Count value D: Corresponding to B in the first embodiment (initial value is set to 1 because X ⁰ = 1) E: Remainder Value F: Address

(Raster → Block Conversion) First, the raster → block conversion will be described with reference to the flowchart of FIG. The concept of conversion itself is the same as the conversion process based on the expression of Formula 1 in the first embodiment. The flow chart and the circuit configuration described above are configurations suitable for performing the conversion process based on the equation (1) by hardware.

The flowchart of FIG. 9 will be described in relation to the flowchart of FIG. Steps 31 and 32 are
Processing similar to steps 1 and 2 of FIG. 1 is executed. next,
According to the determination processing of step 33, YES is obtained while C takes a value of 1 to 6, and in step 34, addition processing of C = C + 1 and E = E + D is performed, and E> (B-
It is determined whether or not 1) (step 35). E is 0 to 55
While the value of is taken, NO is determined, and the process proceeds to step 37. However, when E takes a value of 56 or more, the process of E = E- (B-1) is executed (step 36), and then step 37
Proceed to. If C = 7 in step 33, the determination result is NO and the process directly proceeds to step 37.

In step 37, the process of F = F + D is executed. This process corresponds to step 6 in FIG. That is, this is a process of calculating an address.

Here, when F increases like 1, 2, 3, ..., That is, in step 38, F> (B-
If the answer is NO in step 40, and the answer is NO in step 40, ie, if F == (B-1), the process returns to step 37. Similar to the first time (Y = 0), an address is generated and data is stored as shown in FIG.

By the processing of steps 38 and 39, when the generated address F exceeds the maximum address of the line memory, the address F is corrected to an appropriate value. Further, the step 40 executes the processing corresponding to the step 7 of FIG. The case of NO in step 40 is a case where all the addresses of the line memory are generated, and in this case, the processing of F = 0, D = E, and C = 1 is executed (step 41). F = 0 is a process for initializing the address. The process of D = E corresponds to the process of incrementing Y and obtaining B from this Y in the process of FIG. C = 1 is the initialization of C.

After the processing of step 41, the processing after step 32 is the first time (Y =
This corresponds to the process 1). That is, the processing of D = E results in D = 7, and the address F at this time is
Will be represented by F = F + 7. This means that C = 0 to 6 in step 33, and E = E in step 34.
This is because the processing of + D is repeated, so that E = 7, and D = 7 leads to D = 7.

In step 38, F> (B-
If it is determined to be 1) or not, and if it is determined to be NO in step 40 and if it is determined to be NO in step 40 to reach step 37 again, by the processing of F = F + D, F will increase like 7, 14, 21 ,. Therefore, similarly to the first time (Y = 1) shown in the first embodiment, as shown in FIG. 2, an address is generated and data is stored.

Further, as described above, after the same processing as the first time (Y = 1) shown in the first embodiment is performed, step 41 is performed.
After step 32, the process after step 32 is repeated again, but at this time, step 3 after the process of step 41 is performed.
The second and subsequent processes correspond to the second (Y = 2) process shown in the first embodiment. That is, by the processing of D = E, D = 4
9, and the address F at this time is F = F + 4
It will be represented by 9. This is E = 49 because the processing of E = E + D was repeated in step 34 while C was 1 to 6 in step 33, and thus D =
This is because E = D = 49.

In step 38, F> (B-
If it is determined to be 1) or not, and if it is determined to be NO in step 40 and if it is determined to be NO in step 40 to reach step 37 again, by the processing of F = F + D, F will increase like 49,98, .... Therefore, similarly to the second time (Y = 2) shown in the first embodiment, as shown in FIG. 4, the address is generated and the data is stored.

(Block → Raster Conversion) First, referring to FIG.
The block → raster conversion will be described with reference to the flowchart of FIG. The conversion process itself is the same as the conversion process based on Formula 2 of the first embodiment. This flowchart is similar to the relationship of the flowchart of FIG. 5 with respect to the flowchart of FIG. 1, and in the process of the flowchart of FIG. 9, the number of pixels on one side forming a block as a comparison target of C is applied.

As a result, the same address generation as that shown in FIGS. 6 to 8 of the first embodiment is executed.

Next, the circuit diagram of FIG. 11 will be described in correspondence with the processing contents of FIGS.

The subtractor 1 executes the processing of B-1, and the subtraction output is input to the comparators 7, 12, 13 and the subtracters 8, 14. The above-mentioned comparators 7, 12, 13 are the same as those in step 3 of the flowchart.
5, 55, steps 40 and 60, and steps 38 and 58, respectively.
6, 56 and steps 39, 59, respectively.

The 3-bit shifter 2 generates a value that is 1/8 of B, that is, A, and inputs this to the multiplexer 3.

The multiplexer 3 adopts the value of A in raster-to-block conversion and the value of 8 in block-to-raster conversion, and outputs the adopted value to the comparator 4. The value of A and the value of 8 are switched by a conversion switching signal.

The comparator 4 compares the input value from the multiplexer 3 with the count value (C). This process corresponds to steps 33 and 53. The comparator 4 outputs the comparison result to the C counter 5 and the adder 6. When the clock is input to the C counter 5, the comparison result is C <A.
When (or8) is indicated, the count UP is executed. C
The counter and adder 6 performs the processing of C = C + 1 and E = E + D, and corresponds to steps 34 and 54.
The output E of the adder 6 is input to the comparator 7, the subtractor 8 and the multiplexer 9.

Relatively 7 compares the output E of the adder 6 with the B-1 output of the subtracter 1. This process corresponds to steps 35 and 55. The subtracter 8 outputs the output E of the adder 6.
To subtract the B-1 output of the subtractor 1. This process
Corresponding to steps 36 and 56. The multiplexer 9 is
When the output result of the comparator 7 does not show E> (B-1), the output E of the adder 6 is adopted, and when E> (B-1) is shown, the output E of the subtractor 8 is used. (E = E- (B-1)) is adopted. When a clock is input to the E register 10,
The output E of the multiplexer 9 is latched in the E register.

The adder 11 adds a value obtained by adding the output F from the F register 16 and the output D from the D register 17 to F.
Output as. This process corresponds to steps 37 and 57. The output F of the adder 11 is input to the comparators 12 and 13 and the subtractor 14.

Relatively 13 compares the output F of the adder 11 with the B-1 output of the subtracter 1 and outputs the comparison result to the subtractor 14 and the comparator 12. This process corresponds to steps 38 and 58. The subtractor 14 performs the subtraction process of F- (B-1) when the output of the comparator 13 indicates F> (B-1), but when it does not indicate F> (B-1). Outputs F as it is. This process is step 38.
Corresponds to the processing of step 39 or step 58 and step 59. The comparator 12 determines whether the output F of the adder 11 and the output B-1 of the subtractor 1 match, and the determination result is C.
Counter 5, multiplexer 15, and D register 17
Output to.

The multiplexer 15 selects the output F of the subtracter 14 when the comparison result of the comparator 12 does not indicate a match, and selects 0 when it indicates a match and selects the F register 1
Output to 6. This processing corresponds to the processing of steps 40 and 60.

When the clock is input, the F register 16 latches the value selected by the multiplexer 15.
The process in which the F register 16 latches 0 is step 4
This corresponds to the processing of F = 0 of 1,61. The output of the F register 16 becomes an address output.

The D register 17 holds the value of D when the comparison result of the comparator 12 does not show a match, and holds the output E of the E register 10 when the comparison result shows a match. The process for indicating a match corresponds to the process of D = E in steps 41 and 61.

Further, the C counter 5 holds the value of D when the comparison result of the comparator 12 does not show a match, and holds the output E of the E register 10 when it shows a match. The process for indicating a match corresponds to the process of D = E in steps 41 and 61.

With the above configuration, the raster / block conversion is executed by the hardware configuration (IC or a device equivalent thereto) that does not use the multiplier.

Further, a line memory / frame memory interface 18 is provided in the above circuit, and pixel data can be input. By using a frame memory interface, it can also support a frame memory.

12 and 13 are configured so that the color conversion and the raster / block conversion can be realized at the same time by using the raster / block conversion device (including the one configured by software) configured as described above. It is a thing.
Of course, the block data defined in the JEPEG system may be input / output.

In the color conversion IC 27 shown in FIG. 12, the raster data input from the input circuit 21 is input to the multiplier 22.
The color conversion is performed on the basis of the coefficient data 26.
The raster / block conversion device 23 executes raster-to-block conversion processing using the raster / block RAM 25, and the output circuit 24 outputs the blocked data.

In the color conversion IC 34 shown in FIG. 13, the block data input from the input circuit 28 is raster /
It is output to the block conversion device 29. The raster / block conversion device 29 uses the raster / block RAM 33 to execute raster-to-block conversion processing.
The block data from the raster / block conversion device 29 is color-converted by the multiplier 30 based on the coefficient data 32, and is output through the output circuit 31.

With the ICs shown in FIGS. 12 and 13, color conversion can be performed without any special software or hardware externally provided.

[0095]

As described above, according to the present invention, the line memory can be used to reduce the cost, and the memory address for conversion can be automatically set in any screen size. It has various excellent effects that can be generated.

[Brief description of drawings]

FIG. 1 is a flowchart showing the contents of raster → block conversion processing in the present invention.

FIG. 2 is an explanatory diagram showing the contents of a line memory at the 0th time (Y = 0) in the processing of FIG.

FIG. 3 is an explanatory diagram showing the contents of the line memory at the first time (Y = 1) in the processing of FIG.

FIG. 4 is an explanatory diagram showing the contents of the line memory at the second time (Y = 2) in the processing of FIG.

FIG. 5 is a flowchart showing the contents of block → raster conversion processing in the present invention.

FIG. 6 is an explanatory diagram showing the contents of the line memory at the 0th time (Y = 0) in the processing of FIG. 5;

FIG. 7 is an explanatory diagram showing the contents of the line memory at the first time (Y = 1) in the processing of FIG.

8 is an explanatory diagram showing the contents of the line memory at the second time (Y = 2) in the processing of FIG.

FIG. 9 is a flowchart showing the processing contents when the raster → block conversion in the present invention is executed by addition processing.

FIG. 10 is a flowchart showing the processing contents when the block → raster conversion in the present invention is executed by addition processing.

FIG. 11 is a circuit diagram showing a circuit that implements the processing of FIGS. 9 and 10;

FIG. 12 is a circuit diagram showing a circuit that incorporates the raster / block conversion device of the present invention, has a color conversion circuit and the like, and freely performs color conversion and raster → block conversion at an arbitrary screen size.

FIG. 13 is a circuit diagram showing a circuit that incorporates the raster / block conversion device of the present invention, has a color conversion circuit, etc., and freely performs color conversion and block → raster conversion at an arbitrary screen size.

FIG. 14 is a flowchart showing processing contents that can be considered when performing block → raster conversion using a single screen memory.

[Explanation of symbols]

1 subtractor 2 Left 3 bit shifter 3 multiplexer 4 comparator 5 C counter 6 adder 7 comparator 8 subtractor 9 multiplexer 10 E register 11 adder 12 Comparator 13 Comparator 14 Subtractor 15 multiplexer 16 F register 17 D register 18 memory interface

Claims (4)

(57) [Claims] 1. A is the current address coordinate, A'is the next address coordinate, B is the added value, C is the horizontal pixel number, χ is the horizontal pixel number of one block, and X is C. / Y, where Y is the number of conversions, and when the relationship X ^Y > (C-1) holds, the remainder in the formula X ^Y / (C-1) is B, When the relation of ^XY > (C-1) is not established, the calculation result of ^XY is set to B, and based on the above-obtained B, the next address is obtained by the formula of A '= A + B and the raster data is made into block data. On the other hand, when the relation of χ ^Y > (C-1) is established, the remainder in the formula of χ ^Y / (C-1) is set as B, and the relation of χ ^Y > (C-1) is not established. If the calculation result of the chi ^Y is is B, on the basis of the obtained B, following the equation a '= a + B Raster / Block conversion method and converting the block data seeking address into raster data. 2. The value of X ^Y or χ ^Y in the method of claim 1 is obtained by using a value obtained by sequentially adding the same value until the number of counts reaches X or χ, and is determined by the block size. 2. The raster / block conversion method according to claim 1, wherein the sequentially added values are set to the same value in the next conversion processing after the conversion processing is completed for all the pixels in the memory. 3. The raster / block conversion method according to claim 1, wherein the variable is appropriately changed in accordance with the size of the block. 4. In order to implement the raster / block conversion method according to claim 2, addition means for sequentially adding the same value until the number of counts reaches X or χ, and all pixels in the memory determined by the size of the block. A raster / block converter, which is provided with latch means for making the sequentially added values the same value in the next conversion processing after the conversion processing for