JPH03167962A - Block distortion improving system - Google Patents

Abstract

PURPOSE:To minimize th. deterioration in the picture quality caused following the block distortion improvement processing by estimating the part of occurrence of block distortion at a receiver side in a 2-dimension DCT transformation coding system, applying picture processing to the part and improving the block distortion. CONSTITUTION:A sender side of the DCT transformation coding system takes a difference between picture information and predicted picture for each prescribed block with respect to the picture information. Then a DCT transformation coefficient series obtained by applying the 2-dimension discrete cosine transformation(DCT) is coded and the result is sent to a receiver side. The receiver side estimates whether or not block distortion is generated in the block from the DCT transformation coefficient for each sent block and when the block distortion generation is estimated, the picture processing to improve the block distortion is implemented as to the block. In this case, the picture processing applies, e.g. low pass filter to eliminate high frequency component. Thus, the block distortion generating part as to the receiver side is estimated and the picture processing is applied to the part to improve the block distortion, then the block distortion improving processing by the receiver side processing only is attained.

Description

[Detailed Description of the Invention] [Summary] Regarding an image quality improvement method in an image encoding system, image quality deterioration due to block distortion improvement is minimized, and processing for block distortion improvement is performed only by processing on the receiving side. The purpose of this method is to provide a block distortion improvement method that allows image information to be encoded/decoded on the transmitting side for each certain block, and does not require uniformity of encoding/decoding algorithms on the transmitting and receiving sides. A series of DCT transform coefficients obtained by performing dimensional discrete cosine transform (DCT) is encoded and transmitted, and on the receiving side, the DCT transform coefficients obtained by decoding are subjected to inverse discrete cosine transform to create an image. On the receiving side of a DCT transform coding system for restoring information, a block distortion occurrence estimator that estimates the occurrence of block distortion in the DCT transform coefficients of each received block and generates an output of the estimation result; and a block distortion occurrence estimation unit that generates an output of the estimation result; This is achieved by including an image processing section that performs image processing to improve block distortion with respect to the block.

[Industrial application field]

The present invention relates to an image quality improvement method in an image coding system, and more particularly to a block distortion improvement method for improving block distortion in an image coding system using two-dimensional discrete cosine transform (DCT).

In a transform coding system using two-dimensional DCT, on the transmitting side, a two-dimensional block of a fixed size obtained by dividing an original image is treated as one unit, and information compression processing is performed for each block. On the receiving side, these blocks are connected to reproduce the image. Therefore, if blocks with different properties exist adjacent to each other, the joints between the blocks may be noticeable in the received image. This distortion is called block distortion, and when this distortion occurs, the quality of the transmitted image visually deteriorates significantly.

Therefore, such block distortion cannot be seen (visibly<<
) is desired.

[Conventional technology]

Figure 6(a). (b) shows the basic structure of a DCT transform coding system, exemplifying one that performs layered coding processing, (a) shows the transmission (encoding) side, and (
b) indicates the receiving (decoding) side.

In FIG. 6(a), image information consisting of a luminance signal Y, a color difference signal U, and In the preprocessing section, preprocessing is performed for each block consisting of, for example, 8x8 pixels. That is, the image information in the frame memory l1 is L
After being band-limited in the PF 12, it is sampled at a ratio of 2:1 in the vertical (V) direction and the horizontal (H) direction in the sub-sampling unit 13 to compress the amount of information. This image information is again input to the LPF 12 and band-limited, and then sampled at a ratio of 2:1 in the V direction and the H direction in the sub-sampling section 13 to compress the amount of information. By repeating such processing, the amount of information is compressed to, for example, 1/16, and then subtraction is performed block by block between it and the predicted screen in the subtractor l4, thereby generating information on a difference image.

The information of the difference image is processed by two-dimensional discrete cosine transformation (2D-D
In the CT) unit 15, a two-dimensional discrete cosine transform is performed for each 8×8 pixel block, converting a time domain signal into a frequency domain signal, and generating a series of transform coefficients representing information of the difference image. ．． The variable-length entropy encoding unit 16 performs encoding to assign the shortest code word to the information of this transform coefficient sequence, and generates an encoded signal made of a variable-length code. This encoded signal is
This is the information of the difference image whose information amount has been compressed to 1/16 as described above.

On the other hand, the two-dimensional inverse discrete cosine transform (2D-IDCT) section 17 performs two-dimensional inverse discrete cosine transform on the information of the transform coefficient series of the 2D-DCT section 15, thereby
Restore the difference image. The adder 18 adds the difference image information and the predicted image information block by block to generate restored image information. The linear interpolation unit 19 performs linear interpolation on the information of the restored image, and generates image information with twice the amount of information in each of the V direction and the H direction. The frame memory 20 generates predicted screen information whose information amount is compressed to 1/4 by accumulating the image information of the linear interpolation unit 19 for one screen.

Next, in the preprocessing section, the same processing as the previous time is performed again on the image information whose information amount has been compressed to 1/4, thereby generating an encoded signal for the information of the difference image whose information amount has been compressed to 1/4. At the same time, linear interpolation is performed on the information of the restored image restored from this to generate predicted screen information having the same amount of information as the original image.

Furthermore, in the preprocessing section, the same processing is performed again on the image information whose information amount is not compressed, thereby generating an encoded signal for the information of the difference image with respect to the original image.

In FIG. 6(b), the variable length code decoding section 2l decodes the variable length code sent from the transmitting side and reproduces a series of transform coefficients. The 2D-IDCT unit 22 restores a difference image by performing two-dimensional inverse discrete cosine transform on the information of this transform coefficient series. The adder 23 adds the difference image information and the prediction screen information block by block to generate restored image information, which is stored in the frame memory 2.
4 restores the original image by accumulating this image information for one screen. The programmable linear interpolation unit 25 performs linear interpolation on the information of the restored image at a rate that changes depending on the compression ratio, and generates an image output signal.

On the other hand, the linear interpolation unit 26 performs linear interpolation on the restored image in the frame memory 24 to generate image information with twice the amount of information in the V direction and in the H direction. This image information is used as the above-mentioned predicted screen information.

FIG. 7 is a diagram illustrating the basic operation flow of the DCT transform encoding system.

■ Divide the original image into blocks (subpictures) of a certain size.

(2) Next, perform two-dimensional discrete cosine transformation processing for each block. The two-dimensional discrete cosine transform is expressed by the following formula Y (f+ j
) = A X
is a constant, and Y(i,=+ is a conversion coefficient array.

■ Image information is restructured using the transform coefficients obtained for each block in this way.

■Next, select the conversion coefficients to be transmitted.

This is done by thresholding, subsampling of transform coefficients, etc.

■ Next, perform quantization of the transformation coefficients taking visual characteristics into consideration.

(2) Code words are assigned to the transform coefficients quantized in this way.

When encoding is performed in a hierarchical manner in this way, the first encoded signal with the compressed amount of information has a small amount of information, so the screen definition is low, but the encoding processing time is short, so when switching screens , it is possible to quickly send one screen's worth of information. The amount of information in this image increases each time it is encoded, so the definition gradually increases, and eventually the information of the original screen is sent.

Therefore, according to such a layered encoding method, when still images are intermittently switched and transmitted, it is possible to prevent blanks from appearing on the screen because it takes time to transmit the first image when switching the screen.

In this way, when image information is encoded and transmitted block by block and decoded on the receiving side to reproduce the original image, if there are adjacent blocks with different properties within the bronc, the received image In this case, block distortion occurs where the joints between blocks are noticeable.

In conventional DCT transform coding systems, the following methods are used to improve block distortion.

■ In order to seamlessly connect blocks, the receiving side connects each block and reproduces the received image, then applies a low-pass filter uniformly to the entire screen.

(2) On the transmitting side, which can completely grasp the information of the current image, additional information for detecting the presence or absence of block distortion and correction (improvement) is created and transmitted together with the image information. On the receiving side, image information is corrected based on this information.

[Problem to be solved by the invention]

The block distortion improvement method in the conventional DCT transform coding system has the following problems.

(1) Since the improvement method shown in item (■) in the previous section is a process performed only on the receiving side, it is not possible to completely grasp the information of the original image. Therefore, uniform low-pass filter processing is applied to the entire image, but as a result, processing is applied to areas where block distortion is not visually visible or areas unrelated to block distortion, which adversely affects image quality. It may lead to deterioration.

(2) In the improvement method shown in item (2) in the previous section, the amount of information to be transmitted increases as additional information is added. Furthermore, since the processing required to calculate the additional information is added to the processing on the sending side, it is expected that the processing load and processing time on the sending side will increase. Furthermore, the encoding/decoding algorithms on the transmitting and receiving sides must be completely unified, which impairs the versatility of the encoding system.

The present invention is an attempt to solve the problems of the prior art as described above, and it is possible to minimize the deterioration of image quality due to block distortion improvement, and to improve block distortion only by processing on the receiving side. The present invention aims to provide a block distortion improvement method that does not require uniformity of encoding/decoding algorithms on the transmitting side and the receiving side.

[Means to solve the problem]

As shown in FIG. 1, the present invention has a DC transform obtained by performing a two-dimensional discrete cosine transform (DCT) on image information on a transmitting side 1 for each predetermined block.
Receiving side 2 of a DCT transform encoding system that encodes and transmits a T-transform coefficient sequence, and performs inverse discrete cosine transform on the DCT transform coefficients obtained by decoding it at the receiving side 2 to restore image information. By including a block distortion occurrence estimation section 3 and an image processing section 4, adaptive block distortion improvement can be performed only by processing on the receiving side.

Here, the block distortion occurrence estimation unit 3 estimates the occurrence of block distortion in the received DCT transform coefficients for each block and generates an output of the estimation result, and the image processing unit 4 outputs the estimation result. When this occurs, image processing is performed on that block to improve block distortion.

[Effect]

On the transmitting side of the DCT transform coding system, the difference between the image information and the predicted image is calculated for each fixed block of image information. Then, a two-dimensional discrete cosine transform (DCT) is performed on this difference, and the obtained DCT transform coefficient sequence is encoded and transmitted to the receiving side. On the receiving side, the DCT transform coefficients obtained for each block are decoded and subjected to inverse discrete cosine transform to reproduce a difference image, which is then accumulated to restore the original image information.

In the present invention, on the receiving side, it is estimated from the DCT transform coefficients of each transmitted block whether block distortion will occur in that block. This estimation can be performed, for example, based on whether there is a difference of a certain value or more in the direct current in the DCT transform coefficients between adjacent blocks.

Then, when occurrence of block distortion is estimated, image processing to improve block distortion is performed for that block. Image processing in this case can be performed, for example, by applying a low-pass filter to remove a certain amount of high frequencies.

In this way, the receiving side estimates the location where block distortion occurs and performs image processing on that area to improve professional and mechanical distortion, making it possible to improve block distortion only by processing on the receiving side. It becomes possible.

〔Example〕

Figure 2(a). (b) shows the basic structure of the present invention, and exemplifies an example that performs layered encoding processing in the same way as in the cases of FIGS. 6(a) and (b). ) side, and (b) shows the receiving (decoding) side.
). The same items as in (b) are indicated by the same numbers, 27
28 is a comparison test result storage unit, and 29 is an image processing unit.

In the present invention, block distortion removal processing is performed only on the receiving (decoding) side, and therefore
The configuration of the transmitting (encoding) side 1 shown in Figure (a) is shown in Figure 6 (a).
) is exactly the same as the conventional case shown in .

On the receiving side shown in FIG. 2(b), the variable length code decoding unit 21 decodes the variable length code to reproduce the transform coefficient sequence, and the 2D-IDCT unit 22 converts the information of the transform coefficient sequence into a two-dimensional discrete cosine. Restoration of the difference image by inverse transformation, generation of restored image information by adding the difference image information and predicted screen information in the adder 23, and storage of one screen worth of restored image information in the frame memory 24 to restore the original image. Restoration, programmable linear interpolation unit 25
Generation of an image output signal by linear interpolation with a ratio that changes according to the Ashk ratio to the information of the restored image in ,
The generation of prediction screen information by linear interpolation of the restored image in the frame memory 24 in the linear interpolation unit 26 is performed in the sixth step.
This is done in the same way as in the case shown in FIG.

The DC predetermined comparison test unit 27 calculates the magnitude of the DC component coefficient of the target block and the DC component coefficient of blocks adjacent to the target block.
The difference is calculated by comparing the magnitude of the certain coefficient, and a portion where the magnitude of the difference is equal to or greater than a certain value is estimated as a location where block distortion occurs.

The comparison test result storage unit 28 stores information on the estimation result of the location where block distortion occurs in the DC partial comparison test unit 27 as 1-bit information per block.

The image processing unit 29 performs, for example, low-pass filter processing for each block while referring to the estimation information for each block stored in the comparison test result storage unit 28. At this time, the low-pass filter processing is performed only on the portion where occurrence of block distortion is estimated as a result of the comparison test processing. That is, by applying a low-pass filter only to the portion where block distortion is expected to occur, processing is performed to smoothly connect blocks. As a result of the comparison test processing, no processing is performed on portions in which block distortion is not estimated to occur.

FIG. 3 shows one embodiment of the present invention, showing only the receiving (decoding) side, and the same parts as in FIG. 2(b) are indicated by the same numbers, and 30 is for display. It's memory. In addition, in the DC certain comparison test section 27, 31 is 2D
- DCT section, 32 is a direct current (DC) storage memory, 33
is the DC comparison test section.

In the embodiment shown in FIG. 3, the display memory 30 stores the linearly interpolated image output signal from the programmable linear interpolator 25 for one screen. The image signal for one screen stored in the display memory 30 is used for display on a display (not shown) or the like.

In addition, in the DC certain comparison test section 27, 2D-DCT
The unit 31 performs a two-dimensional discrete cosine transform on the restored image for one screen stored in the frame memory 24, and generates a transform coefficient series representing information on the difference image, which is made up of frequency domain signals. do.

The DC fraction storage memory 32 stores only the direct current (DC) fraction in the frequency domain output signal of the 2D-DCT section 31 for each block.

The DC certain comparison test section 33 compares the DC certain output from the 2D-DCT section 31 with the DC certain output of the adjacent block stored in the DC discrimination storage memory 32, thereby determining block distortion. Visually visible parts are tested and an output signal is generated to estimate the location where block distortion occurs.

The comparison test result storage unit 28 stores the test result in the DC precept comparison test unit 33 as information of 1 bit per block. The image processing unit 29 includes the comparison test result storage unit 2
Based on the information of the storage result in step 8, the image signal for one screen stored in the display memory 30 is subjected to block distortion improvement processing such as a low-pass filter as described above. No processing is performed on portions in which block distortion is not estimated to occur.

Figures 4(a) and 4) exemplify the determination of areas where block distortion is easily visible; (a) shows a target block to be determined and a block to be compared with it; (b) is a flowchart showing the flow of determination.

Block distortion is likely to be visually visible in areas where there is a difference in the tendency of signal level change within two adjacent blocks, and where there is a difference in average signal level. Therefore, the value of the DC coefficient of the judgment target block A, which is the object of judgment of the nature of the image information inside the block, and each of the upper, lower, left, and right blocks B, C, which are in contact with that block by line segments,
Examine the differences b, c, d, and e with the DC coefficient values of D and E,
If there is even one block whose value exceeds the predetermined value TH, the determination target block A is determined to be the target block to which block distortion improvement processing should be performed.

Figures 5 (a), (b), and (C) show an overview of the operation of the image processing unit, and show changes in DC components for each block and improvement of connections between blocks by image processing.
This is a dimensional model.

Now, as shown in (a) in Fig. 5, are there any adjacent areas? If the difference in DC component level between broncs is 2 units or more, DCrfj. The minute comparison test result shall be set as 1, and if it is less than that, it shall be set as 0.

In the image processing section, when the DC precept comparison test result is 1, low-pass filter processing is performed using a filter whose matrix is exemplified in (b). Pixels shown with diagonal lines in the figure are target pixels, and this filter is applied to each pixel in a block consisting of 8×8 pixels. However, in the area of two pixels around the image, the filter pattern is changed as edge processing.

Note that the filter F shown in FIG. ,1
, which generates the filter output )'(i+j), which is defined by the following equation.

Vti■>=F/64 F ”” (X (i-Lj-21 + X (i-
Lj-H + X (i-Ljhz+ + X (i
aL j"!l ) X 1+( X (i-L+
j-21 +X (i・I+ j-2) + 'X
(i-2,j-l)+X (i+Lj-11
+ X (i*I+j-2) + X
<=? , j4■) + X (i-1+ j-Z)
”X (i and 2+j”l+)×2 +(
11 +X (i.1j1) +X (i-L j
) + X Li◆2+ jl +X (i-1,j・
I) + X (,・1・j◆1) + X (i・j−z
+ ) X 3+(X (i+j-1) + X
(i-1+jl + X (il+j) +X (
i・1+H + X (i+ jl) X 4
In Fig. 5, (C) shows the result of low-pass filter processing, and in (1), the parts where there is a difference of 2 units or more are cut off from the high castle and are smoothly connected to adjacent blocks. However, it is shown that no processing is performed on the image information in areas where the difference is less than that.

〔Effect of the invention〕

As explained above, according to the present invention, in the two-dimensional DCT transform encoding method, the block distortion is improved by estimating the location where block distortion occurs on the receiving side and performing image processing on that part. Therefore, in the block distortion improvement process performed only on the receiving side, it is possible to realize adaptive processing that minimizes the deterioration in image quality that accompanies the block distortion improvement process. In the present invention, there is no increase in the processing load on the transmitting side, and there is no need for uniformity of encoding/decoding algorithms on the transmitting side and the receiving side.

In the case of the method of the present invention, if the block size, expression accuracy of transform coefficients, etc. are clearly specified in the two-dimensional DCT transform encoding method, sufficient image quality improvement effect can be expected, and simple processing is possible. , it becomes possible to contribute to the realization of a high-quality, highly adaptable coding system.

[Brief explanation of the drawing]

Fig. 1 is a diagram showing the principle structure of the present invention, Fig. 2(a),
(b) is a diagram showing the basic structure of the present invention, Figure 3 is a diagram showing an embodiment of the present invention, and Figures 4 (a) and (b) are parts where block distortion is easily visible. A diagram illustrating the judgment of FIG. 5(a). (b) and (C) are diagrams showing an overview of the operation of the image processing section, and Fig. 6 (a). (b) is a diagram showing the basic structure of the DCT transform encoding system, and FIG.
FIG. 2 is a diagram illustrating the flow of basic operation of the T-transform encoding system. 1. 2 is a DCT transform encoding system, ■ is a transmitting side. 2 is the receiving side, 3 is the block distortion generation estimation unit, 4
is an image processing section.

Claims (1)

[Claims] On the transmitting side (1), two-dimensional discrete cosine transform (hereinafter abbreviated as DCT) is performed on image information for each fixed block.
The resulting DCT transform coefficient sequence is encoded and transmitted, and the receiving side (2) decodes it to obtain the
The receiving side (2
), a block distortion occurrence estimator (3) that estimates the occurrence of block distortion in the DCT transform coefficients of each received block and generates an output of the estimation result; A block distortion improvement method characterized by comprising: an image processing unit (4) that performs image processing.