JP2839942B2 - Image data restoration method and apparatus - Google Patents

Description

DETAILED DESCRIPTION OF THE INVENTION [Overview] For each block obtained by dividing an original image into a plurality of blocks each consisting of a plurality of pixels (N × N), a gradation value of a plurality of pixels in the block is calculated. A method and apparatus for quantizing transform coefficients obtained by two-dimensional discrete cosine variable (ADCT) and restoring an image from code data obtained by encoding the obtained quantized coefficients, which speeds up image restoration by inverse DCT transform. The purpose is to use the fact that it is not necessary to operate on the decoded zero DCT coefficient, determine whether the decoded DCT coefficient is zero, and determine the zero coefficient and the non-zero coefficient. Is set, and the number of DCT coefficients for performing the inverse DCT operation is controlled according to the set value of the flag.

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data restoring method and apparatus for restoring a data compressed image, and in particular, divides a multi-valued image into blocks each including a plurality of pixels and orthogonally intersects the pixels in the blocks. The present invention relates to an image data restoration method and apparatus for restoring an image from orthogonally-transformed encoded data of a multi-valued image that has been encoded after conversion.

Accumulates image data whose information amount is orders of magnitude larger than numerical data, especially data of halftone images and color images.
Alternatively, in order to transmit at high speed and with high quality, it is necessary to efficiently encode the gradation value for each pixel.

Conventionally, as an efficient compression method for image data, for example, there is an adaptive discrete cosine transform coding method.

Adaptive Discret Cosine Transform Coding (Adaptive Discret
e Cosine Transform (hereinafter abbreviated as “ADCT”) will be described below.

The ADCT divides an image into blocks of 8 × 8 pixels, converts the image signal of each block into coefficients of a spatial frequency distribution by a two-dimensional discrete cosine transform (hereinafter, referred to as “DCT”), and adapts it to vision. Quantization is performed using a threshold, and the obtained quantization coefficient is encoded by a Huffman table that is statistically obtained.

The encoding operation will be described in detail according to the basic configuration diagram of the ADCT shown in FIG.

First, the image is divided into blocks of 8 × 8 pixels shown in FIG.
In the two-dimensional DCT transform unit 24, the input image signal is orthogonally transformed by DCT, converted into spatial frequency distribution coefficients shown in FIG. 12, and output to the linear quantization unit 25.

Specifically, as shown in FIG. 8, the image signal input from the terminal 23 is subjected to one-dimensional DCT conversion by the one-dimensional DCT conversion unit 30,
The transposition unit 31 transposes (transposes) the rows and columns of the coefficients in the block and outputs the result to the one-dimensional DCT transformation unit 32. The one-dimensional DCT conversion unit 32 performs one-dimensional DCT conversion in the same manner as the one-dimensional DCT conversion unit 30 and outputs the result to the transposition unit 33. The transposition unit 33 performs the same transposition processing as the transposition unit 31, and outputs the result to the terminal.

By performing such processing for all blocks of the image data, it is converted into DCT coefficients.

Referring again to FIG. 7, the linear quantization unit 25 linearly quantizes the input DCT coefficients by using a quantization matrix 29 composed of the threshold values shown in FIG. 13 and determined by a visual experiment. Are obtained. First
As shown in Fig. 4, the quantized DCT coefficient has a value smaller than the threshold.
The DCT coefficient becomes 0, and a quantized DCT coefficient in which only the DC component and a few AC components have values is generated.

The two-dimensionally arranged quantized DCT coefficients are one-dimensionally converted according to a scanning order called a zigzag scan shown in FIG. The variable length coding unit 26 calculates the DC component at the head of each block and the DC component of the previous block.
The difference from the component is variable-length coded. For the AC component, the value of the effective coefficient (coefficient whose value is not 0) (hereinafter, referred to as “index”) and the invalid coefficient (coefficient whose value is 0) up to that value
(Hereinafter, referred to as “run”) is subjected to variable-length coding for each block. The DC and AC components are encoded using a code table 27 composed of a Huffman table created based on statistics for each image, and the obtained code data is sequentially output from a terminal 28.

On the other hand, the code data is restored to an image by the following method.

FIG. 9 shows a configuration diagram of the ADCT restoration circuit, and FIG.
FIG. 2 shows a configuration diagram of a dimensional inverse DCT transform unit.

In FIG. 9, code data input from a terminal 40 is input to a variable length decoding unit 41. In the variable length decoding unit 41, the seventh
The decoding unit 42 decodes the input code data into fixed-length data of an index and a run by using a decoding table 42 formed by a table opposite to the Huffman table of the code table 27 shown in FIG. The inverse quantization unit 43 multiplies each of the quantization matrices 29 to inversely quantize the input quantization coefficient to restore the DCT coefficient, and outputs the DCT coefficient to the two-dimensional inverse DCT transformation unit 44.

The two-dimensional inverse DCT transformer 44 converts the input DCT coefficient into an inverse DCT
The orthogonal transform is performed by the transform, and the coefficient of the spatial frequency distribution is converted into an image signal. Specifically, as shown in FIG.
The input DCT coefficient is subjected to one-dimensional inverse DCT transform by a one-dimensional inverse DCT transform unit 51 and output to a transpose unit 52. The transposition unit 52
One-dimensional inverse DCT by exchanging the rows and columns of coefficients in one block
Output to the converter 53. The one-dimensional inverse DCT transformer 53 performs one-dimensional inverse DCT on the input transposed coefficients again, and
Output to The transposition unit 54 exchanges the rows and columns of the coefficients in one block again as in the transposition unit 52, and
Outputting from 45 restores the image.

[Prior Art] Conventional one-dimensional inverse DCT transform units 51 and 53 convert one block into 8 ×
In the case of an eight-pixel configuration, an image in one column is restored by performing an inverse DCT operation on the eight pixels in one column by a matrix operation shown in (Equation 1) to (Equation 9). Note that in (Equation 1) to (Equation 9), [X11] to [X81] are DCT of one row
The coefficients [Y11] to [Y81] are the restored image signals in one column, [A1
[1] to [A88] are conversion constants.

Y11 = A11 * X11 + A12 * X21 + A13 * X31 + A14 * X41 + A15 * X51 + A16 * X61 + A17 * X71 + A18 * X81 = F11 + F12 + F13 + F14 + F15 + F16 + F17 + F18 (Equation 2) Y21 = A21 * X11 + A22 * X21 + A23 * X31 + A24 * X41 + A25 * X51 + A26 * X61 + A27 * X71 + A28 * X81 = F21 + F22 + F23 + F24 + F25 + F26 + F27 + F28 (formula 3) Y31 = A31 * X11 + A32 * X21 + A33 * X31 + A34 * X41 + A35 * X51 + A36 * X61 + A37 * X71 + A38 * X81 = F31 + F32 + F33 + F34 + F35 + F36 + F37 + F38 (equation 4) Y41 = A41 * X11 + A42 * X21 + A43 * X31 + A44 * X41 + A45 * X51 + A46 * X61 + A47 * X71 + A48 * X81 = F41 + F42 + F43 + F44 + F45 + F46 + F47 + F48 (Equation 5) Y51 = A51 * X11 + A52 * X21 + A53 * X31 + A54 * X41 + A55 * X51 + A56 * X61 + A57 * X71 + A58 * X81 = F51 + F52 + F53 + F54 + F55 + F56 + F57 + F58 (Equation 6) = F61 + F62 + F63 + F64 + F65 + F66 + F67 + F68 (Equation 7) Y71 = A71 * X11 + A72 X21 + A73 * X31 + A74 * X41 + A75 * X51 + A76 * X61 + A77 * X71 + A78 * X81 = F71 + F72 + F73 + F74 + F75 + F76 + F77 + F78 (Equation 8) Y81 = A81 * X11 + A82 * X21 + A83 * X31 + A84 * X41 + A85 * X51 + A86 * X61 + A87 * X71 + A88 * X81 = F81 + F82 + F83 + F84 + F85 + F86 + F87 + F88 (Equation 9) [invention moth resolution However, in such a conventional image data restoring method and apparatus, when restoring DCT coefficients into an image, the DCT coefficients of the pixels of all blocks are subjected to inverse DCT transformation.
There was a problem that it took time to restore the image.

That is, when one block is composed of 8 × 8 pixels, the inverse DCT transform is an 8 × 8 matrix operation, and the transform of one pixel is 8 × 8.
Since one multiplication and eight additions are performed, the conversion of 64 pixels in one block requires 512 multiplications and 512 additions. For this reason, when the pixels of all the blocks of one screen are subjected to the inverse DCT transform, there is a problem that it is difficult to speed up image restoration.

The present invention has been made in view of such a conventional problem, and an object of the present invention is to provide a method and apparatus for restoring image data, which can speed up image restoration by inverse DCT.

[Means for Solving the Problems] FIG. 1 is a principle diagram of the present invention, taking an image data restoring device as an example.

First, according to the present invention, an original image is divided into a plurality of pixels (N ×
N; where N is a positive integer), and for each block obtained by dividing into a plurality of blocks consisting of: a conversion coefficient obtained by performing a two-dimensional discrete cosine transform on the tone values of a plurality of pixels in the block. The present invention is directed to an image data restoration apparatus that restores an image from code data obtained by quantizing the obtained quantized coefficients.

According to the present invention, there is provided a DCT coefficient holding unit 12 for holding a decoded DCT coefficient; and a zero detecting unit 1 for determining whether or not the decoded DCT coefficient is zero.
3; a flag setting means 14 for setting a flag indicating a distribution of a coefficient having a value of zero and a non-zero coefficient among the N × N DCT coefficients based on the determination result of the zero detecting means 13; Conversion constant holding means 16 for holding a conversion constant for performing the inverse DCT conversion of the obtained DCT coefficient; and inverse DCT conversion means 17 for performing an inverse DCT conversion of the DCT coefficient read from the DCT coefficient holding means 12 and restoring the image signal. And reading control means 15 for controlling reading of signals from the DCT coefficient holding means 12 and the conversion constant holding means 16.

Here, the zero detecting means 13 operates in parallel with the DCT coefficient holding means 12.

Further, in order to operate the DCT coefficient holding means 12 and the inverse DCT transform means 17 in parallel, the DCT coefficient holding means 12 is used for two blocks of D.
A DCT coefficient holding area is provided, and one block of DCT coefficients is written in one DCT coefficient holding area, and at the same time, a written DCT coefficient is read from the other DCT coefficient holding area.

Further, the reading control means 15 operates according to the setting result of the flag setting means 14, and the flag setting means 14 generates a first flag indicating whether or not all the N DCT coefficients in one column are zero. One flag creation means and N / 2 DTCs in the latter half of one column
The apparatus further includes a second flag creation unit that creates a second flag indicating whether or not all coefficients are zero, and a third flag creation unit that creates a third flag using the first flag and the second flag.

As a result, when the N DCT coefficients are all zero, the reading control means 15 skips reading of the N DCT coefficients in one column held in the DCT coefficient holding means 12, and reads N / 2 DCT coefficients. When the coefficients are all zero, the reading of the N / 2 DCT coefficients in one column held in the DCT coefficient holding means 12 is skipped, and the reading of the conversion constant held in the conversion constant holding means 16 is also performed simultaneously. Control. On the other hand, the image data restoring method according to the present invention is configured such that the processing operation of the above-described apparatus configuration is performed as a temporal processing step, and the original image is composed of a plurality of pixels (N
.Times.N), for each block obtained by dividing the block into a plurality of blocks, the transform coefficients obtained by two-dimensional discrete cosine transform of the tone values of the plurality of pixels in the block are quantized. In an image data restoring method for restoring an image from encoded data obtained by encoding the quantized coefficients, wherein a first step of holding the decoded N × N DCT coefficients; To set flag information indicating the distribution of a coefficient having a value of zero and a non-zero coefficient being a non-zero value.
And a third step of performing an inverse DCT transform on the input DCT coefficient to restore the image signal to an image signal, wherein the DCT coefficient stored in the first step is stored in accordance with the information set in the second step. Is controlled.

Here, a fourth step of determining whether or not the DCT coefficient is zero is provided in parallel with the holding of the decoded DCT coefficient in the first step, and the second step is based on the determination result of the fourth step, and is based on N × N DCT
It is based on setting flag information indicating the distribution of a coefficient having a value of zero and a non-zero coefficient among the coefficients, and performing the inverse DCT transform of the third process by the readout control based on the flag information.

[Operation] According to the image data restoration method and apparatus of the present invention having such a configuration, when the value of the restored DCT coefficient is zero, the inverse DCT operation result of this coefficient is zero, and the coefficient of zero Using the fact that there is no need to operate on
By determining whether or not the T coefficient is zero, setting a flag indicating the distribution of the zero coefficient and the non-zero coefficient, and controlling the number of DCT coefficients for performing the inverse DCT operation according to the set value of the flag. In addition, the number of operations of the inverse DCT transform can be significantly reduced, and the image restoration speed can be improved.

Embodiment FIG. 2 is a block diagram showing an embodiment of an inverse DCT conversion circuit according to the present invention.

That is, the image data restoration apparatus of the present invention is constituted by the restoration circuit shown in FIGS. 9 and 10 for restoring the code data obtained by the ADCT type encoding apparatus shown in FIGS. The two-dimensional inverse DCT transform unit 44 shown has the configuration shown in FIG.
Note that one block restored in this embodiment has an 8 × 8 pixel configuration.

First, the variable length decoding unit 41 shown in FIG. 9 decodes the data and then inversely quantizes it in the inverse quantization unit 43, as shown in FIG.
The DCT coefficients for one block of the × 8 pixel configuration are sequentially input from the terminal 61 to the DCT coefficient holding unit 62 and held. The DCT coefficient from the terminal 61 is also input to the zero detector 63 at the same time. It detects whether or not the DCT coefficient input to the zero detector 63 is zero, and outputs the detection result to the flag setter 64.

The flag setting unit 64 prepares two types of control flags A and B shown in FIG. 3 as flags indicating the distribution of zero DCT coefficients.

First, the control flag A (first flag) in FIG.
Indicates whether all eight DCT coefficients in the n-th column in one block are zero or not. Here, n = 1 to 8 are column numbers. That is, flag data of 1 bit is provided for flag Nos. 1 to 8 indicating a column, and the flag data is bit 0; DCT coefficients of one column are all zero; bit 1; DCT coefficients of one column are all zero Is not.

The control flag B (second flag) in FIG.
Indicates whether or not the latter four of the eight data in the n-th column in one block are all zero. That is, 1-bit flag data is provided for the flag Nos. 1 to 8 indicating the columns, and the flag data is as follows: bit 0; all four in the second half of the column are not zero; All zeros. The flag setting operation by the flag setting unit 64 is performed as follows in accordance with the operation flow of FIG.

First, in step S1 (hereinafter "step" is omitted), the control flags A = 0 and B = 1 indicating that the DCT coefficients in one column are all zero are initialized.

Next, proceed to S2, and read the first DCT in the first column read first.
The coefficient is held in the DCT coefficient holding unit 62, and in the next S3, the zero detector 63
It is determined whether the detection result is zero. If the DCT coefficient is zero, the process proceeds to S4, where it is determined whether or not the counter i indicating the number of DCT coefficients in one column has reached the last eighth, and if not,
In S5, the counter i is incremented by one, and in S2, the next i is incremented.
The second DCT coefficient where = 1 is read, and the same processing is repeated as long as the DCT coefficient is zero. In this case, the control flags A,
In B, the initial values A = 0 and B = 1 are set as they are.

On the other hand, if the DCT coefficient is not zero in S3, the process proceeds to S6, where the control flag A is changed to A = 1 indicating that the DCT coefficients are not all zero, and the counter i indicates the latter four in S7. 3
Is determined. If the counter i is 3 or less, the control returns to S2 via S4 and S5 without changing the control flag B because one of the first four is not zero, and reads the next DCT coefficient. If the counter i exceeds 3 and is one of the latter four at S7, the control flag B, which has been set to B = 1 indicating that the latter four are all zero in the initial state, is set to S8. To change B = 0.

By repeating the above process for one column (eight columns) for one block, a flag for one block (eight columns) is set, and the setting result of the flag is output to the read control unit 65.

That is, the flag setting results for each column output to the read control unit 65 are as follows: A = 0, B = 1; all columns are zero A = 1, B = 1; Any other than B = 0;

For example, if the DCT coefficient for one block held in the DCT coefficient holding unit 62 is in the state shown in FIG.
The control flags A and B in the first to eighth columns shown in FIG. 6 are set by the flag setting operation shown in FIG.

The read control unit 65, according to the set value of the input flag,
Read the corresponding DCT coefficient from the DCT coefficient holding unit 62 and
In addition to the output to the T calculation unit 67, the conversion constant required for the calculation is output from the conversion constant holding unit 66 to the inverse DCT calculation unit 67.

Here, as shown in FIG. 10, the inverse DCT calculation unit 67
It is composed of a dimensional inverse DCT transform unit 51, a transpose unit 52, a one-dimensional inverse DCT transform unit 53, and a transpose unit 54.

The DCT coefficient holding unit 62 has a storage area for two blocks. The DCT coefficient for one block is stored in one area at the same time as the DCT coefficient for one block previously stored is read from the other area. The conversion is performed by the conversion unit 67 in parallel with the operation.

More specifically, based on the control flags A and B in FIG. 6, the read control unit 65 first sets the column where the control flag (AB) = (10) with respect to the DCT coefficient in FIG. , The first column and the seventh column are read out, and are calculated by the inverse DCT calculator. Further, for the column in which the control flag is (AB) = (11), that is, for the second and sixth columns, 1
Only the first four elements of the eight elements in the column are read and operated. Of course, reading is performed in the order of the first to eighth columns.

As a result, in the example of FIG. 5, one block of 64 DCTs
Of the coefficients, eight in the first column, four in the third column, four in the sixth column
If only 24 DCT coefficients, ie, 8 in the seventh column and a total of 24 DCT coefficients, are read, operated and output from the terminal 68, an image signal for one block is restored.

By repeating the above process for each block and for one screen, an image for one screen is restored.

[Effect of the Invention] As described above, according to the present invention, the decoded DCT
Zero detection is performed in parallel with the holding of the coefficient, and a flag indicating the distribution of the coefficient whose value is zero and the non-zero coefficient among the DCT coefficients in one column is set, and the inverse DCT calculation is performed according to the set value of the flag. By controlling the number of DCT coefficients output to the
The number of operations can be greatly reduced, and the speed of image restoration processing can be increased.

[Brief description of the drawings]

FIG. 1 is a diagram illustrating the principle of the present invention; FIG. 2 is a diagram illustrating an embodiment of an inverse DCT transform unit according to the present invention; FIG. 3 is a diagram illustrating a control flag indicating a zero DCT coefficient distribution according to the present invention; FIG. 5 is a flowchart of a flag setting operation of the present invention; FIG. 5 is an explanatory diagram of DCT coefficients for one block processed by the present invention; FIG. 6 is an explanatory diagram of control plugs A and B obtained from FIG. FIG. 8 is a block diagram of an ADCT-type encoding circuit; FIG. 8 is a block diagram of a two-dimensional DCT converter shown in FIG. 7; FIG. 9 is a block diagram of an ADCT-type restoration circuit; FIG. 11 is an explanatory diagram of an original image signal of one block; FIG. 12 is an explanatory diagram of DCT coefficients when the image signal of FIG. 11 is subjected to DCT; FIG. FIG. 14 is an explanatory diagram of the threshold of the adaptive DCT transform; FIG. 14 is an explanatory diagram of the quantized DCT coefficient when the DCT coefficient of FIG. 12 is quantized using the threshold of FIG. 10; Conversion FIG. 4 is an explanatory diagram of a scanning order for performing In the figure, 12: DCT coefficient holding means 13: Zero detection means 14: Flag setting means 15: Read control means 16: Conversion constant holding means 17: Inverse DCT conversion means 62: DCT coefficient holding unit 63: Zero detector 64: Flag Setting unit 65: Read control unit 66: Conversion constant holding unit 67: Inverse DCT calculation unit

──────────────────────────────────────────────────続 き Continuation of the front page (72) Inventor Kimitaka Murashita 1015 Kamikodanaka, Nakahara-ku, Kawasaki City, Kanagawa Prefecture Inside Fujitsu Limited (56) References JP-A-2-51978 (JP, A) JP-A-2 -154571 (JP, A) JP-A-60-247782 (JP, A) JP-B-56-32662 (JP, B2) (58) Fields investigated (Int. Cl. ⁶ , DB name) H04N ^7/ 24- 7/68 H04N 1/41-1/419

Claims (8)

(57) [Claims] 1. For each block obtained by dividing an original image into a plurality of blocks each consisting of a plurality of pixels (N × N), two-dimensional discrete cosine transform is performed on the tone values of the plurality of pixels in the block. In the image data restoring method for quantizing the transform coefficients obtained as a result and restoring an image from code data obtained by encoding the obtained quantized coefficients, a method for holding the decoded N × N DCT coefficients A second step of setting flag information indicating a distribution of a coefficient having a value of zero and a non-zero coefficient among the N × N DCT coefficients; and a DCT read from the first step. A third process of inverse DCT transforming the coefficients into an image signal; and a second process in which a first flag state indicating that all DCT coefficients in one column of N are zero is set. Does not read the DCT coefficients of the one column from the first step. Image data restoration method. 2. An original image is composed of a plurality of pixels (N × N).
For each of the blocks obtained by dividing into a plurality of blocks consisting of: a transform coefficient obtained by performing a two-dimensional discrete cosine transform on the tone values of the plurality of pixels in the block; An image data restoration method for restoring an image from code data obtained by encoding coefficients, a first step of holding decoded N × N DCT coefficients; and a value among the N × N DCT coefficients A second process for setting flag information indicating a distribution of a coefficient where is zero and a non-zero coefficient which is not zero; and a third process for performing an inverse DCT transform of the DCT coefficient read from the first process to restore the image signal. In parallel with the holding of the decoded DCT coefficients of the first course,
A fourth process for determining whether or not the CT coefficient is zero, wherein the second process is performed in one row based on the determination result of the fourth process.
When it is detected that the latter N / 2 of the DCT coefficients are all zero, the DCT coefficient of the one column held in the first process is
An image data restoration method, wherein only the first N / 2 of the T coefficients are read out in the third process. 3. The image data restoration method according to claim 2, wherein the latter half of the DCT coefficients in the second step is detected when all N / 2 of the latter half are zero. An image data restoration method, which is effective when at least one of the DCT coefficients is not zero. 4. An original image is composed of a plurality of pixels (N × N).
For each block obtained by dividing into a plurality of blocks consisting of: quantizing a transform coefficient obtained by performing a two-dimensional discrete cosine transform on the gradation of the plurality of pixels in the block;
In an image data restoration apparatus for restoring an image from code data obtained by encoding the obtained quantized coefficients, DCT coefficient holding means (12) for holding N × N decoded DCT coefficients; Zero detection means for determining whether the DCT coefficient is zero (1
3) and; N × N DCs based on the detection result of the zero detection means (13).
Flag setting means (14) for setting a flag indicating a distribution of a coefficient having a value of zero and a non-zero coefficient among the T coefficients; and holding a conversion constant for performing an inverse DCT conversion of the input DCT coefficient. Transform constant holding means (16); and invert the DCT coefficient read from the DCT coefficient holding means (12).
Inverse DCT transform means for DCT transform and restoration to image signal (17)
The DCT coefficient holding means (12) and the conversion constant holding means (1
Read control means (15) for controlling reading from (6); and the read control means (15) comprises: the flag setting means (14).
When the first flag state indicating that the DCT coefficients of the N columns in one column are all zero is set, the reading of the DCT coefficients in the one column from the DCT coefficient holding means (12) is prohibited. An image data restoration device characterized by the above-mentioned. 5. An original image is composed of a plurality of pixels (N × N).
For each block obtained by dividing into a plurality of blocks consisting of: quantizing a transform coefficient obtained by performing a two-dimensional discrete cosine transform on the gradation of the plurality of pixels in the block;
In an image data restoration apparatus for restoring an image from code data obtained by encoding the obtained quantized coefficients, DCT coefficient holding means (12) for holding N × N decoded DCT coefficients; Zero detection means for determining whether the DCT coefficient is zero (1
3) and; N × N DCs based on the detection result of the zero detection means (13).
Flag setting means (14) for setting a flag indicating a distribution of a coefficient having a value of zero and a non-zero coefficient among the T coefficients; and holding a conversion constant for performing an inverse DCT conversion of the input DCT coefficient. Transform constant holding means (16); and invert the DCT coefficient read from the DCT coefficient holding means (12).
Inverse DCT transform means for DCT transform and restoration to image signal (17)
The DCT coefficient holding means (12) and the conversion constant holding means (1
Reading control means (15) for controlling reading from 6); and the flag setting means (14) detects that the latter N / 2 of the DCT coefficients in one column are all zero. The DCT
An image data restoration apparatus characterized in that only N / 2 first half DCT coefficients of said one column held in coefficient holding means (12) are read out to said inverse DCT transform means (17). 6. The image data restoration apparatus according to claim 5, wherein the detection of the DCT coefficient in the flag setting means is N.
An image data restoration apparatus, which is effective when at least one of the DCT coefficients in one column is not zero. 7. The image data restoring device according to claim 4, wherein said zero detecting means (13) comprises:
An image data restoration device which operates in parallel with the DCT coefficient holding means (12). 8. The image data restoration apparatus according to claim 7, wherein said DCT coefficient holding means has a DCT coefficient holding area for two blocks, and one DCT coefficient holding area in one DCT coefficient holding area. An image characterized in that the DCT coefficient holding means (12) and the inverse DCT transform means (17) are operated in parallel by reading the written DCT coefficients from the other DCT coefficient holding area while writing the coefficients. Data recovery device.