JPH0818970A - Encoder and decoder for image data - Google Patents

Abstract

PURPOSE:To improve picture quality by generating boundary signal level difference addition information used in the reduction processing of block distortion in an image data decoder which performs decoding based on an adjacent boundary signal level differential value. CONSTITUTION:A correction information generating part 32 provided on an image data encoder side generates information used in block distortion reduction processing when encoded data whose data quantity is compressed and encoded by such compression is decoded again to two-dimensional image data on an image data decoder side. The information is the boundary signal level difference addition information, and it is generated based on the adjacent boundary signal level differential value. A block distortion correction part 34 provided on the decoder side performs the block distortion reduction processing as changing the degree of block distortion reduction corresponding to the boundary signal level difference addition information transmitted via a certain transmission line 50b. The correction part 34 suppresses the degree of block distortion reduction corresponding to the boundary signal level difference addition information.

Description

Detailed Description of the Invention

[0001]

BACKGROUND OF THE INVENTION The present invention converts two-dimensional image data into two.
A coded image in which the data amount of the two-dimensional image data is compressed by dividing the two-dimensional image data into a plurality of blocks in a three-dimensional space and subjecting the divided two-dimensional image data to the two-dimensional image data. BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image data encoding device that generates data, or an image data decoding device that decodes the encoded data by inverse quantization processing and inverse orthogonal transform processing. Block distortion called image quality disparity caused by discontinuity of the strength of low frequency components between the blocks, which occurs in the two-dimensional image after the processing, thereby further improving the image quality of the decoded two-dimensional image. The present invention relates to an image data encoding device and an image data decoding device capable of performing the above.

[0002]

2. Description of the Related Art Usually, image information has a large amount of data, and especially natural images and moving images have a very large amount of data. Therefore, various data compression techniques for efficiently transmitting or storing such image information have been disclosed.

As one of data compression techniques, DCT (di
There is a technique of quantizing and further encoding after orthogonal transformation such as screte cosine transform). Hereinafter, such a technique is referred to as a transform-coded data compression technique. Further, such data compression is hereinafter referred to as transform-coded data compression.

Further, as such a transform-coded data compression technique, there is a technique in which a pixel image to be data-compressed is first divided into blocks and the transform-coded data is compressed for each divided block. After that, such technology
This is called a block division conversion coding data compression technique.

FIG. 5 is a block diagram of a general image coding apparatus.

The image coding apparatus shown in FIG.
The block-divided pixel image is compressed by the transform coded data. Further, the image coding apparatus of FIG. 5 performs coding with data compression on a pixel image signal X to be transmitted or stored, and generates a coded image signal Z. .

The image coding apparatus shown in FIG. 5 mainly includes an orthogonal transformation means 40, a quantization means 42, a quantization table 44, a variable length coding means 46, and a variable length coding table 48. It is composed by.

The orthogonal transformation means 40 uses the pixel image signal X
Is subjected to orthogonal transformation to generate a frequency component signal F. Examples of the orthogonal transformation include two-dimensional Fourier transformation, Walsh-Hadamard transformation, Haar transformation, and slant.
Transform (gradient transform), Discrete Cosine Transform (DCT), K
There are Arhunen-Loeve conversion (hereinafter also referred to as KLT) and the like.

The quantizing means 42 generates a coded image signal Y using the quantized coefficient Q obtained from the quantizing table 44. The encoded image signal Y has been subjected to data compression and may be used for transmission or storage, but the variable length encoding means 46 may further perform data compression.

The variable length coding means 46 generates a coded image signal Z from the coded image signal Y by using the variable length coding table data T obtained from the variable length coding table 48. The encoded image signal Z is used to transmit or store the pixel image signal X.
The encoded image signal Z is obtained by compressing the pixel image signal X in data, and is further compressed in data than the encoded image signal Y.

FIG. 6 is a block diagram of a general image decoding apparatus.

The image decoding apparatus shown in FIG.
The pixel image signal X2 corresponding to the pixel image signal X is generated (decoded) from the coded image signal Z generated by the image encoding device of FIG.

The image decoding apparatus of FIG. 6 mainly includes a variable length decoding means 12, a variable length coding table 14, and
The inverse quantization means 16, the quantization table 18, and the inverse orthogonal transformation means 20 are included.

The variable length coding means 12 decodes the coded image signal Y2 from the coded image signal Z using the variable length decoding table data T2 obtained from the variable length coding table 14.

The inverse quantization means 16 uses the inverse quantization coefficient Q2 obtained from the quantization table 18 to generate a frequency component signal F2 from the encoded image signal Y2.

The inverse orthogonal transform means 20 performs an inverse orthogonal transform on the frequency component data F2 to generate a pixel image signal X2. The pixel image signal X2 corresponds to the pixel image signal X in FIG.

Data compression techniques include data compression by lossless encoding and data compression by lossy encoding. In the image encoding device and image decoding device described above, data compression is performed by the lossy encoding. Such data compression (encoding) involves a certain amount of reduction in the amount of information such as distortion, but significant data compression can be achieved. Further, when the image to be data-compressed is intended for visual observation, by setting the numerical values (data) of the quantization tables 18 and 44 described above in consideration of visual psychological characteristics and the like, the adverse effect of the distortion or the like is exerted. Can be reduced to a range where there is no problem.

However, when the data amount is compressed by performing the orthogonal transformation process or the quantization process as described above, block distortion occurs in the two-dimensional image after being decoded into the two-dimensional image data again. There was a problem.

This block distortion is the strength of the frequency component between the blocks divided at the time of coding, which occurs when the coded image data is coded or decoded as described above. This is due to the image quality gap due to the discontinuity of. Such block distortion is because the two-dimensional image data is divided into a plurality of blocks in the two-dimensional space at the time of encoding for compressing the data amount of the two-dimensional image data.

Further, such block distortion becomes larger as the degree of compression of the amount of data made in encoding becomes larger. For example, in the quantization processing when the two-dimensional image data is encoded into the encoded image data, such a block distortion becomes more remarkable as the degree of compression of the data amount increases. Is.

In order to suppress the deterioration of the image quality of a two-dimensional image after decoding due to such block distortion, conventionally, when decoding the coded image data in which the data amount is compressed and coded, such Correction of block distortion is performed.

For example, in Japanese Patent Laid-Open No. 62-196990,
When decoding the encoded image data by the inverse quantization process and the inverse orthogonal transform process, after the inverse quantization process and the inverse orthogonal transform process, each pixel of a pixel row extending over two blocks Using multiple image data for
From the image data in one block, the image data relating to the block boundary pixels of the pixel row in the other block are extrapolated and predicted. Further, the image data of the pixels near the boundary of the pixel column in each block is corrected so as to approach the image data value predicted by extrapolation as the block boundary is approached.

Further, in Japanese Patent Laid-Open No. 5-176176, in a picture decoding apparatus for decoding a block-divided two-dimensional image after orthogonal transformation, quantized and further coded into a two-dimensional image again, an inverse orthogonal transformation is performed. From the previous image signal, D
A DC component extracting means for extracting the C (direct current) component is provided. Further, a correction value calculating means is provided for calculating a correction value used for reducing the discontinuity of the magnitude of the DC component between the blocks of the image signal after the inverse orthogonal transform from the DC component. Further, the correction value is used to correct the image signal after the inverse orthogonal transform, and a block distortion correction means for reducing block distortion is provided.

The above-mentioned Japanese Patent Laid-Open No. 62-19699
0 and the above-mentioned Japanese Patent Laid-Open No. 5-176176, it is possible to reduce the block distortion as described above, and improve the image quality of a two-dimensional image after decoding, especially when the target two-dimensional image is a visual target. It is possible to further improve the image quality in terms of visual psychological characteristics.

[0025]

However, the above-mentioned JP-A-62-196990 and JP-A-5-176176 are known.
Although the conventional techniques for reducing block distortion can reduce block distortion as described above, on the other hand, there is a problem that image deterioration occurs due to such processing for reducing block distortion.

The conventional techniques for reducing block distortion, such as Japanese Patent Laid-Open No. 62-196990 and Japanese Patent Laid-Open No. 5-176176, are based on such smoothing processing and filtering processing. For this reason, the two-dimensional image after decoding has a blurred impression as a whole. For example, the contours of various images in the decoded two-dimensional image, the contrast of the edges of the boundary, and the like are reduced, and the image is deteriorated.

The present invention has been made to solve the above-mentioned conventional problems, and divides two-dimensional image data into a plurality of blocks in a two-dimensional space, and divides the divided two-dimensional image data into blocks. By performing the orthogonal transformation process and the quantization process, the data amount of the two-dimensional image data is compressed, and the encoded image data encoded by the two-dimensional image data is decoded again into the two-dimensional image data. An image that can reduce the block distortion that occurs in the same time and at the same time suppress the deterioration of the image due to the process of reducing the block distortion, thereby improving the image quality of the two-dimensional image after decoding. An object is to provide a data encoding device and an image data decoding device.

[0028]

The image data encoding apparatus of the first invention of the present application divides two-dimensional image data into a plurality of blocks in a two-dimensional space, and divides the divided two-dimensional image data into blocks. In an image data coding apparatus for generating coded image data in which the data amount of the two-dimensional image data is compressed by performing orthogonal transform processing and quantization processing by the vertical transform, among the plurality of blocks, a vertical direction or a horizontal direction. For two adjacent blocks that are adjacent to each other in the direction, the signal level of the pixel in each of the adjacent blocks that approach each other via the adjacent boundary line is determined as a block peripheral signal level for each of the adjacent blocks. A difference value between the signal level calculation unit and the adjacent block of the block peripheral signal level is defined as an adjacent boundary signal level difference value. Boundary signal level difference additional information used in block distortion reduction processing in an image data decoding device that decodes the coded image data based on the obtained signal level difference calculation unit and the adjacent boundary signal level difference value By providing the additional information generating unit, the image data encoding device that can achieve the above object is provided.

On the other hand, the image data decoding apparatus of the second invention of the present application divides the two-dimensional image data into a plurality of blocks in the two-dimensional space, and performs orthogonal transformation processing on the divided two-dimensional image data in block units. And by performing quantization processing,
In an image data decoding device for decoding coded image data in which the data amount of the two-dimensional image data is compressed by an inverse quantization process and an inverse orthogonal transform process, a vertical direction or a horizontal direction among a plurality of the blocks. For two adjacent blocks that are adjacent to each other, as the signal level of the pixels in each of the adjacent blocks that approach each other through the adjacent boundary line, the encoded image to be decoded for each of the adjacent blocks. Regarding the block peripheral signal level obtained on the image data encoding device side for generating data, and as the difference value between the adjacent blocks of the block peripheral signal level, it was obtained on the image data encoding device side. For the adjacent boundary signal level difference value,
A boundary used in the block distortion reduction process on the image data decoding device side for decoding the encoded image data, which is obtained on the image data encoding device side based on the adjacent boundary signal level difference value. Regarding the signal level difference additional information, as the absolute value of the adjacent boundary signal level difference value increases according to the boundary signal level difference additional information, block distortion at the adjacent boundary line related to the adjacent boundary signal level difference value By providing a block distortion correction unit configured to suppress the degree of block distortion reduction in the reduction processing, the present invention provides an image data decoding device that can achieve the above-mentioned object.

[0030]

The conventional processing for reducing block distortion, which has been performed when decoding coded image data having a compressed data amount, is to carry out the smoothing processing and the filtering processing as described above.

Therefore, the discontinuity or change in the signal level of the image near the boundary between the blocks is uniformly corrected even if the signal level originally exists in the original image. For this reason, the discontinuity of the signal level originally present in the original image is smoothed, or the degree of change of the signal level originally present in the original image becomes gentle.

Therefore, in the present invention, the processing for reducing the block distortion as described above is not performed on the boundaries between all the blocks, but the boundary signal level generated on the image data encoding device side. It is configured to be selectively performed based on the difference addition information. In this manner, the boundary data level difference additional information is generated on the image data encoding device side by
In the image after decoding, when there is a signal level difference near the boundary between blocks, whether the signal level difference is due to the discontinuity or change of the signal level originally present in the original image, or such a signal level This is because it is not possible to determine whether the difference is due to block distortion.

Therefore, such a boundary signal level difference addition information is provided on the side of the image data encoding device on the side on which the encoded image data is encoded, for use in the determination on the side of the image data decoding device. I am trying to generate. On the image data decoding device side, the processing content of the block distortion reduction processing is changed according to such boundary signal level difference additional information.

Here, it is necessary to suppress the total amount of data sent from the image data encoding device side to the image data decoding device side, including the encoded image data and the like. Therefore, it is also necessary to suppress the data amount of the boundary signal level difference additional information. Considering these points,
The boundary signal level difference additional information is obtained as follows.

FIG. 1 is a line showing the gist of the present invention, which shows block division of two-dimensional image data, which is targeted by the image data encoding device of the first invention and the image data decoding device of the second invention. It is a figure.

In FIG. 1, the blocks B (iL-1), B (iL), B (i-L + 1) are centered around the block Bi.
), B (i-1), B (i + 1), B (i + L-1), B (i + L
) And B (i + L + 1) are shown. Where i is
Although the two-dimensional image data is divided into a plurality of blocks in the two-dimensional space, the block number from the beginning of the two-dimensional image in the raster scan image is shown. Further, L is the number of blocks per row of a plurality of blocks two-dimensionally arranged in such a raster scan image.

Here, the block Bi is the block of interest during the orthogonal transformation process, the quantization process and the like in the encoding by the image data encoding device. Alternatively, it is set as a block of interest that has been subjected to the inverse quantization process and the inverse orthogonal transform process during decoding in the image data decoding device.

For the target block Bi,
, The blocks B (iL-1), B (iL).
..A total of 8 blocks such as B (i + L + 1) are adjacent.

Here, for the target block Bi,
The block B (iL) is an upper adjacent block. The block B (i + L) is a lower adjacent block. The block B (i-1) is the left adjacent block. The block B (i + 1) is the right adjacent block.

Further, line segments W1-W2, W3- are provided for these four adjacent blocks with respect to the noted block Bi.
W4, W1-W3, and W2-W4 are adjacent boundary lines, respectively. Here, with respect to the noted block Bi, the block distortion as described above is caused by these adjacent boundary lines W1-W2,
There is a problem in each of W3-W4, W1-W3, W2-W4, or in the vicinity thereof.

FIG. 2 is a diagram showing the outline of the first invention and the second invention regarding the adjacent boundary line around the target block.

In FIG. 2, as an example, a block in which one block has a total of (8 × 8 = 64) pixels is shown. Particularly, the block B of interest shown in FIG.
i is composed of pixels P1 to P64.

The block of interest Bi shown in FIG.
With regard to the above, the operation of the first invention and the second invention will be described with the adjacent boundary lines W1-W2 as an example.

As described above, the first invention and the second invention
According to the invention, whether the signal level difference near the adjacent boundary lines such as the adjacent boundary lines W1-W2 after the inverse quantization process and the inverse orthogonal transform process at the time of decoding originally existed in the original image. , Or block distortion.

Therefore, in the first aspect of the invention, when a two-dimensional image is encoded, it is determined whether or not there is a signal level difference in the original image at the adjacent boundary lines such as the adjacent boundary lines W1-W2. The image data encoding device side makes a decision, and information based on the decision result is generated as the boundary signal level difference additional information.

Therefore, in order to generate such boundary signal level difference additional information, among the plurality of blocks,
For two adjacent blocks that are adjacent to each other in the up-down direction or the left-right direction, as the signal level of the pixel in each adjacent block approaching each other via the adjacent boundary line,
The block peripheral signal level is obtained for each of the adjacent blocks.

For example, the adjacent boundary lines W1-W in FIG.
In FIG. 2, the signal level of the pixels P1 to P8 or the pixels P9 to P16 is targeted for the noted block Bi side. Further, in the adjoining boundary line W1-W2, on the adjoining block B (iL) side, pixels U57 to
The signal level of U64, and further, the pixels U49 to U56 and the like are targeted.

In the first invention, the block peripheral signal level is not particularly limited. As for the block peripheral signal level, at least one of two adjacent blocks adjacent to each other via one adjacent boundary line is obtained for each adjacent block.

For example, with respect to the adjacent boundary lines W1-W2, the average value of the signal levels of the pixels P1 to P8 may be used as one of the block peripheral signal levels on the side of the target block Bi. On the other hand, the pixels U57 to U64
The average value of the signal levels of the above may be used as the block peripheral signal level on the side of the adjacent block B (iL).

Alternatively, a plurality of the block peripheral signal levels may be obtained for each of the two adjacent blocks adjacent to each other via the adjacent boundary line. For example, with respect to the adjacent boundary line W1-W2, the signal level of each of the eight pixels P1 to P8 may be the total of eight block peripheral signal levels on the target block Bi side. On the other hand, the signal level of each of the eight pixels U57 to U64 in total may be the total of eight block peripheral signal levels.

Alternatively, for example, the adjacent boundary lines W1-W
2, the signal level of one pixel of the pixel P4 at the center of the adjacent boundary line W1-W2 is set to the target block B.
It may be the block peripheral signal level of i. On the other hand, the one-pixel signal level of the pixel U60, which is opposed to this and is substantially at the center of the adjacent boundary lines W1-W2, may be the block peripheral signal level.

Alternatively, with respect to the adjacent boundary lines W1-W2, the average value of the signal levels of the 16 pixels P1 to P16 in total may be set as the block peripheral signal level on the side of the noted block Bi. On the other hand, the average value of the signal levels of the 16 pixels U49 to U64 in total may be used as the block peripheral signal level of the adjacent block B (iL). That is, the target pixels are not only those that are in direct contact with the adjacent boundary line,
It may be in the vicinity of the target adjacent boundary line.

In the first invention and the second invention, the difference value of the block peripheral signal level thus obtained between the adjacent blocks is obtained as the adjacent boundary signal level difference value. That is, the difference value between one and the other of the block peripheral signal levels of two adjacent blocks that are adjacent to each other via the adjacent boundary line is set as the adjacent boundary signal level difference value.

For example, with respect to the adjacent boundary lines W1-W2, the difference value between the block peripheral signal level on the side of the target block Bi and the block peripheral signal level of the adjacent block B (iL) is determined by the adjacent boundary line W1-W2. This is the signal level difference value.

The adjacent boundary signal level difference value is not limited to the one obtained for each of the adjacent boundary lines. For example, in one adjacent boundary line, when a plurality of the block peripheral signal levels are obtained for each of two adjacent blocks that approach each other via the adjacent boundary line, the difference value between the corresponding blocks through the adjacent boundary line is used. Alternatively, a plurality of the adjacent boundary signal level difference values may be obtained.

For example, with respect to the adjacent boundary lines W1-W2, the pixels P1 to P8 are provided on the side of the block of interest Bi.
This is the case, for example, when a total of eight independent block peripheral signal levels are obtained for each signal level. On the other hand, for the adjacent block B (iL) side, a total of 8
For each of the pixels U57 to U64, a total of 8 independent pixels
This is a case where the number of the peripheral signal level of each block is obtained.
In such a case, a plurality of adjacent boundary signal level difference values may be obtained. For example, in such a case, the difference value between the signal level of the pixel P1 and the signal level of the pixel U57 is set as one of the adjacent boundary signal level difference values. The difference value between the signal level of the pixel P2 and the signal level of the pixel U58 is defined as one of the adjacent boundary signal level difference values. In this way, the difference values of the signal levels of the corresponding pixels may be sequentially obtained, and a total of eight adjacent boundary signal level difference values may be obtained.

In this way, when the adjacent boundary signal level difference value is obtained, in the first aspect of the invention, image data for decoding the encoded image data based on the obtained adjacent boundary signal level difference value. Boundary signal level difference additional information used in block distortion reduction processing on the decoding device side is generated.

As described above, on the image data decoding device side of the second aspect of the present invention, as the absolute value of the adjacent boundary signal level difference value increases in accordance with the additional boundary signal level difference information, the adjacent boundary signal level difference value increases. The degree of block distortion reduction of the block distortion reduction processing on the adjacent boundary line related to the signal level difference value is suppressed. This is because, when the absolute value of the adjacent boundary signal level difference value is large, it is assumed that the original image originally has a signal level discontinuity or change, and such an original image originally present signal level discontinuity or change. However, this is to prevent the contrast and the like from being reduced by the block distortion reduction processing on the image data decoding device side.
Therefore, in the first invention, the boundary signal level difference additional information used in the second invention is thus generated.

In the first invention and the second invention, the boundary signal level difference additional information is not particularly limited. That is, the boundary signal level difference additional information is obtained based on the adjacent boundary signal level difference value, and the tendency of whether or not the absolute value of the adjacent boundary signal level difference value is large on the image data decoding device side. You just need to know.

For example, the boundary signal level difference additional information is
The adjacent boundary signal level difference value may be used as it is. That is, when there is one adjacent boundary signal level difference value for one adjacent boundary line, this may be used as it is as the boundary signal level difference additional information of the adjacent boundary line. Alternatively, even when a plurality of the adjacent boundary signal level difference values are obtained for one of the adjacent boundary lines, all such a plurality of the adjacent boundary signal level difference values are set to the boundary signal level. It may be difference addition information.

Alternatively, the boundary signal level difference additional information may be a numerical value obtained based on some calculation formula using one or a plurality of the adjacent boundary signal level difference values.

For example, if a predetermined threshold value is used for the adjacent boundary signal level difference value and the adjacent boundary signal level difference value is equal to or more than the threshold value, the boundary signal level difference additional information of "1" is generated. You may do it. on the other hand,
If the adjacent boundary signal level difference value is less than or equal to the threshold value,
The boundary signal level difference additional information of "0" may be generated. In this case, if the boundary signal level difference additional information is “1”, the degree of block distortion reduction of the block distortion reduction processing in the image data decoding device is suppressed,
On the other hand, if the boundary signal level difference additional information is "0", the degree of block distortion reduction of the block distortion reduction processing in the image data decoding device is increased. Alternatively,
If the boundary signal level difference additional information is "1", the block distortion reduction processing in the image data decoding device is stopped, while if the boundary signal level difference additional information is "0", the image data decoding is performed. The block distortion reduction processing in the device may be made effective.

Alternatively, in the image data decoding apparatus of the second invention, for example, if the absolute value of the adjacent boundary signal level difference value recognized by the boundary signal level difference additional information is not less than a predetermined value k, the block If the absolute value of the adjacent boundary signal level difference value is less than or equal to the predetermined value k, the distortion reduction process is stopped, and the block distortion reduction is performed according to the magnitude of the absolute value of the adjacent boundary signal level difference value. It is also conceivable to carry out the block distortion reduction processing whose degree is controlled. In this case, for example, in the image data encoding device according to the first aspect of the present invention, if the adjacent boundary signal level difference value is equal to or more than the predetermined value k, the generation of the boundary signal level difference additional information is stopped, and the amount of data is increased by that amount. May be reduced.

The calculation of the boundary signal level difference additional information for the adjacent boundary line is basically performed for all four adjacent boundary lines of each block. For example, for the block of interest Bi in FIG. 2, four adjacent boundary lines W1-W2 and W3-W are provided.
4, W1-W3, W2-W4.

However, depending on the position of the target block Bi in the two-dimensional image, the target block Bi
There may be only three blocks or two blocks adjacent to. For example, for the block of interest Bi at the upper left of the entire two-dimensional image, there are only two adjacent blocks. In this case, the boundary signal level difference additional information for the two adjacent boundary lines W3-W4 and W2-W4 may be obtained respectively.

It should be noted that with respect to one of the adjacent border lines,
Two blocks correspond. For example, with respect to the adjacent boundary line W1-W2, the target block Bi and the adjacent block B (iL) correspond to each other. Therefore, 1
The boundary signal level difference additional information may be obtained only once for each of the two adjacent boundary lines. For example, the boundary signal level difference additional information of the adjacent boundary lines W2-W4 regarding the target block Bi in FIG. 2 is common to the adjacent block B (i + 1).
Further, the adjacent boundary line W3 for the target block Bi
The boundary signal level difference additional information of -W4 can be shared with the adjacent block B (i + L). Therefore,
Regarding the block of interest Bi, the adjacent boundary line W1-
Only the boundary signal level difference additional information regarding W2 and the boundary signal level difference additional information regarding the adjacent boundary lines W1-W3 may be obtained. In this case, the boundary signal level difference additional information related to the adjacent boundary lines W2-W4 and the adjacent boundary lines W3-W4 is obtained for the adjacent blocks B (i + 1) and B (i + L), respectively. The one that is used is used.

[0067]

Embodiments of the present invention will be described in detail below with reference to the drawings.

FIG. 3 is a block diagram showing a configuration of an image data encoding / decoding device to which the first invention and the second invention are applied.

As shown in FIG. 3, the image data encoding / decoding apparatus of the present embodiment, as the image data encoding apparatus, is similar to the conventional apparatus of FIG. The encoding unit 42, the quantization table 44, the variable length encoding unit 46, and the variable length encoding table 48 are provided. Further, the image data encoding / decoding device is the same as the conventional image data decoding device shown in FIG.
The variable length decoding means 12, the inverse quantization means 16 and the inverse orthogonal means 20 are provided. The variable length decoding means 12
Then, the variable length coding table 48 shared with the variable length decoding means 46 is used. Further, the inverse quantization means 16 uses the quantization table 18 shared with the quantization means 42.

Further, in the image data encoding / decoding apparatus of the present embodiment, the correction information generating section 32 to which the first invention is applied and the block distortion correction to which the second invention is applied are applied. And a portion 34.

First, the correction information generator 32 is provided on the image data encoding device side. The correction information generation unit 32 is used for block distortion reduction processing when the coded data whose data amount is compressed and coded by this is decoded again on the image data decoding device side into two-dimensional image data. Information is generated on the image data encoding device side. Such information is the boundary signal level difference additional information, and is generated based on the adjacent boundary signal level difference value.

On the other hand, the block distortion correction section 34 is provided on the image data decoding device side.
The block distortion correction unit 34 changes the degree of block distortion reduction according to the boundary signal level difference additional information generated on the image data encoding device side and transmitted through some transmission path 50b, Block distortion reduction processing is performed. The block distortion correction unit 34, in accordance with the boundary signal level difference additional information, increases the absolute value of the adjacent boundary signal level difference value. The degree of block distortion reduction of the block distortion reduction processing is suppressed.

This is because when the absolute value of the adjacent boundary signal level difference value is large, the difference in image quality between the blocks at the adjacent boundary lines related to the adjacent boundary signal level difference value is the signal originally present in the original image. This is because it is assumed to be due to discontinuity or change in level, etc. In this way, with respect to the discontinuity or change in the signal level originally present in the original image, if the leveling process or the filtering process is performed as the block distortion reduction process, the two-dimensional image obtained again becomes unclear. This is because it ends up.

FIG. 4 is a block diagram showing the arrangement of the correction information generating section used in this embodiment.

As described above, two blocks correspond to one adjacent boundary line. For example, in FIG. 2, for the adjacent boundary lines W1-W2, the block of interest Bi and the adjacent block B (i-
L) corresponds to. Therefore, in the correction information generation unit 32 of the present embodiment, regarding the target block Bi being processed, the boundary signal level difference additional information related to the adjacent boundary line W1-W2 and the adjacent boundary line W1-W3 are added. Only the boundary signal level difference additional information is obtained. In addition, the adjacent boundary lines W2-W4 and the adjacent boundary lines W
Regarding the boundary signal level difference additional information related to 3-W4, the adjacent blocks B (i + 1) and B (i + L) are respectively included.
) Is used.

As shown in FIG. 4, the correction information generator 32 shown in FIG. 3 has a peripheral signal level calculator 32a, a signal level difference calculator 32b, and an additional information generator 32c. , Is configured on the side of the image data encoding device.

First, the peripheral signal level calculator 32a
Among the plurality of blocks divided in the orthogonal transformation process, two adjacent blocks that are adjacent to each other in the up-down direction or the left-right direction are included in each of the adjacent blocks that approach each other through the adjacent boundary line. The signal level of the pixel is obtained as a block peripheral signal level for each of the adjacent blocks.

First, the adjacent boundary lines W1-W in FIG.
2 to obtain the boundary signal level difference additional information according to
The peripheral signal level calculation unit 32a includes the orthogonal transform means 4
Of the four adjacent boundary lines of the block after the orthogonal transformation by 0, the block peripheral signal level of the lower one in particular is stored. Specifically, when the orthogonal transform means 40 is currently performing the orthogonal transform of the i-th block, the blocks below the respective blocks of a total of L blocks from the adjacent blocks B (iL) to Bi. Memorize the peripheral signal level. For example, for the adjacent block B (iL) shown in FIG.
U64 is stored. Further, similarly, in order to obtain the boundary signal level difference additional information related to the adjacent boundary lines W1-W2 of FIG. 2, the block of interest B shown in FIG.
The signal levels of the pixels P1 to P8 of i are respectively extracted.

Next, the peripheral signal level calculator 32a
2 is processed by the orthogonal transform means 40 in order to obtain the boundary signal level difference additional information relating to the adjacent boundary lines W1-W3 in FIG.
The pixels R8, R16, R24, R32, R40, R48, R of the adjacent block B (i-1) on the right, which is one before i.
The signal level of each pixel of 56 and R64 is stored.
Similarly, regarding the adjacent boundary lines W1-W3, as shown in FIG.
The pixels P1, P9, P of the block of interest Bi shown in FIG.
The signal levels of 17, P25, P33, P41, P49 and P57 are extracted.

Further, the signal level difference calculating section 32b is
A difference value between the adjacent blocks of the block peripheral signal level obtained by the peripheral signal level calculating section 32a is obtained as an adjacent boundary signal level difference value.

The signal level difference calculating section 32b first determines the first adjacent value from the difference value between the block peripheral signal level above the target block Bi and the block peripheral signal level below the adjacent block B (iL). The boundary signal level difference value is calculated. This is to subtract the signal level of the corresponding one of the pixels P1 to P8 from the corresponding one of the signal levels of the pixels U57 to U64 stored in advance in FIG. The signal level of the pixel P1 is subtracted from the signal level of U57, and the pixel P1 is subtracted from the signal level of the pixel U58.
The signal level of 2 is subtracted.

Further, the signal level difference calculating section 32b is
A second adjacent boundary signal level difference value is obtained from the difference value between the block peripheral signal level on the left side of the target block Bi and the block peripheral signal level on the right side of the adjacent block B (i-1). This is because, in FIG. 2, the signal level of the pixel P1 is subtracted from the signal level of the pixel R8, the signal level of the pixel P9 is subtracted from the pixel R16, and the pixel P17 is subtracted from the signal level of the pixel R24.
The signal level of is subtracted.

Therefore, in the signal level difference calculating section 32b, for example, in FIG. 2, a total of eight adjacent boundary signal level difference values are obtained for the adjacent boundary lines W1-W2 of the target block Bi, and the adjacent boundary lines are obtained. For W1-W3, a total of eight adjacent boundary signal level difference values are obtained.

Further, the additional information generating section 32c is on the image data decoding device side for decoding the encoded image data based on the adjacent boundary signal level difference value obtained by the signal level difference calculating section 32b. The boundary signal level difference additional information used in the block distortion reduction process is generated.

The additional information generator 32c obtains the boundary signal level difference additional information on the adjacent boundary lines W1-W3 of the block of interest Bi. This is a total of eight adjacent boundary lines W1 obtained by the signal level difference calculation unit 32b.
-This is based on the average value of the adjacent boundary signal level difference values related to W2. Further, the additional information generation unit 32c further
The boundary signal level difference additional information concerning the adjacent boundary lines W1-W3 of the block of interest Bi is obtained. this is,
This is based on the average value of the adjacent boundary signal level difference values related to the total eight adjacent boundary lines W1-W3 obtained by the signal level difference calculation unit 32b.

Regarding the two pieces of the boundary signal level difference additional information obtained by the additional information generating section 32c, those relating to the adjacent boundary lines W1-W2 of the target block Bi are the adjacent blocks B (iL). It is also used as a thing. The boundary signal level difference additional information related to the adjacent boundary lines W1 to W3 of the block of interest Bi is also used as that of the adjacent block B (i-1). Further, the boundary signal level difference additional information thus obtained is transmitted to the image data decoding device side via the transmission line 50b.

As described above, in this embodiment, the peripheral signal level calculating section 32a obtains the block peripheral signal level, the signal level difference calculating section 32b obtains the adjacent boundary signal level difference value, and finally, The boundary information level difference additional information is obtained by the additional information generating unit 32c.

As described above, according to the present embodiment, when the coded image data is decoded on the side of the image data decoding device and further the block distortion reduction processing is performed by the block distortion correction unit 34, The boundary signal level difference additional information generated by the correction information generation unit 32 on the image data encoding device side can be used. That is, in the block distortion correction unit 34 on the side of the image data decoding device, in the process of reducing the block distortion as described above, according to the boundary signal level difference additional information, the boundary signal level difference additional information is used. As the absolute value of the adjacent boundary signal level difference value increases, the degree of block distortion reduction on the adjacent boundary line related to the adjacent boundary signal level difference value can be suppressed. As a result, it is possible to prevent the discontinuity or change of the signal level originally present in the original image from being unnecessarily corrected, and it is possible to further improve the image quality of the two-dimensional image after decoding.

Further, since the processing centered on the correction information generating section 32 and the block distortion correcting section 34 can be constructed by a relatively simple circuit, it is possible to configure the circuit when it is constructed by hardware. It is possible to suppress the increase.

In this embodiment, a total of eight adjacent boundary signal level differences are obtained for the adjacent boundary lines W1-W2 of the target block Bi, and the adjacent boundary lines W1-W2 are calculated based on the average value. Of the boundary signal level difference additional information is obtained. Further, for the adjacent boundary lines W1-W3 of the block of interest Bi, a total of eight adjacent boundary signal level differences are calculated, and the average value thereof is used to obtain the boundary signal level difference additional information of the adjacent boundary lines W1-W3. ing.

However, for this, only one adjacent boundary signal level difference value for each pixel at the center position in the length direction of each adjacent boundary line is obtained, and based on this, the boundary signal level difference additional information is obtained. If the above is obtained, the circuit configuration can be simplified. For example, regarding the block of interest Bi in FIG. 2, only the adjacent boundary signal level difference between the signal level of the pixel U60 and the signal level of the pixel P4 is obtained for the adjacent boundary lines W1-W2, and the adjacent boundary signal level difference is obtained. Based on the above, the boundary signal level difference additional information is obtained. For the adjacent boundary lines W1-W3, the adjacent boundary signal level difference between the signal level of the pixel R32 and the signal level of the pixel P25 is obtained, and the additional information of the boundary signal level difference is obtained from the adjacent boundary signal level difference. It is to ask.

In this way, when only one adjacent boundary signal level difference is obtained for each adjacent boundary line and the boundary signal level difference additional information is generated based on this, the average as in the above-mentioned embodiment is obtained. The calculation of values and the like are unnecessary, and the circuit configuration can be simplified.

[0093]

As described above, according to the present invention, 2
Dimensional image data is divided into multiple blocks in a two-dimensional space,
Encoded image data encoded by the data amount of the two-dimensional image is compressed by subjecting the two-dimensional image data to the orthogonal transformation process and the quantization process in block units after division. While reducing the block distortion called the image quality disparity due to the discontinuity of the strength of the low frequency component between the blocks, which occurs in the two-dimensional image after the two-dimensional image is decoded into data again, at the same time, It is possible to obtain an excellent effect that it is possible to suppress the deterioration of the image due to the processing of reducing the block distortion and improve the image quality of the two-dimensional image after decoding.

[Brief description of drawings]

FIG. 1 is a diagram showing block division according to a first invention and a second invention of the present application.

FIG. 2 is a diagram showing adjacent boundary lines or block peripheral signal levels according to the first invention and the second invention.

FIG. 3 is a block diagram showing a configuration of an image data encoding / decoding device to which the first invention and the second invention are applied.

FIG. 4 is a block diagram showing a configuration of a correction information generation unit used in the embodiment.

FIG. 5 is a block diagram showing a configuration of a conventional image data encoding device.

FIG. 6 is a block diagram showing a configuration of a conventional image data decoding device.

[Explanation of symbols]

3 ... Original image data memory 5 ... Decoded image data memory 12 ... Variable length decoding means 14 ... Variable length decoding table 16 ... Inverse quantization means 18 ... Inverse quantization table 20 ... Inverse orthogonal transformation means 32 ... Correction information generation unit 32a ... Peripheral signal level calculation unit 32b ... Signal level difference calculation unit 32c ... Additional information generation unit 34 ... Block distortion correction unit 40 ... Orthogonal transformation unit 42 ... Quantization unit 44 ... Quantization table 46 ... Variable length coding unit 48 ... Variable-length coding table 50a, 50b ... Transmission path Bi ... Block of interest B (iL), B (i-1), B (i + 1), B (i + L) ... Adjacent block P1 to P64 ... Pixel

Claims (2)

[Claims] 1. Two-dimensional image data is obtained by dividing two-dimensional image data into a plurality of blocks in a two-dimensional space, and subjecting the divided two-dimensional image data to orthogonal transformation processing and quantization processing in block units. In an image data encoding device that generates encoded image data in which the data amount of data is compressed, in the plurality of blocks, two adjacent blocks that are adjacent to each other in the up-down direction or the left-right direction have their adjacent boundary lines defined. A peripheral signal level calculation unit that obtains a signal level of a pixel in each of the adjacent blocks that approach each other as a block peripheral signal level for each of the adjacent blocks, and between the adjacent blocks of the block peripheral signal level. A signal level difference calculation unit that obtains the difference value at the adjacent boundary signal level difference value; An image data decoding apparatus for decoding the coded image data based on a fractional value, and an additional information generation unit for generating boundary signal level difference additional information used in block distortion reduction processing. Image data encoding device. 2. The two-dimensional image data is obtained by dividing the two-dimensional image data into a plurality of blocks in a two-dimensional space, and subjecting the divided two-dimensional image data to orthogonal transformation processing and quantization processing in block units. In an image data decoding device that decodes encoded image data in which the data amount of data is compressed by an inverse quantization process and an inverse orthogonal transform process, a plurality of blocks are adjacent to each other in the vertical direction or the horizontal direction. For the two adjacent blocks, the coded image data to be decoded is generated for each of the adjacent blocks as a signal level of a pixel in each of the adjacent blocks that approach each other via the adjacent boundary line. Regarding the block peripheral signal level obtained on the image data encoding device side, and between the adjacent blocks of the block peripheral signal level. As a difference value of the adjacent boundary signal level difference value obtained on the side of the image data encoding device, and based on the adjacent boundary signal level difference value, obtained on the side of the image data encoding device, On the image data decoding device side for decoding encoded image data, regarding the boundary signal level difference additional information used in the block distortion reduction processing, according to the boundary signal level difference additional information, the adjacent boundary signal level difference value As the absolute value increases, a block distortion correction unit is provided so as to suppress the degree of block distortion reduction of the block distortion reduction processing on the adjacent boundary line related to the adjacent boundary signal level difference value. Image data decoding device.