KR20210034225A - Video coding method and apparatus using inloop filter - Google Patents

Abstract

The present invention provides a video coding method and device using an in-loop filter to improve the coding efficiency of a video signal. In applying a deblocking filter of the in-loop filter technologies, the same filter is applied by integrating filters applied to a luminance component and a chrominance component.

Description

Video coding method and apparatus using in-loop filter {VIDEO CODING METHOD AND APPARATUS USING INLOOP FILTER}

The present invention relates to a video signal processing method and apparatus.

Video images are compression-encoded by removing spatio-temporal redundancy and inter-view redundancy, which can be transmitted through a communication line or stored in a format suitable for a storage medium.

The present invention is to improve the coding efficiency of a video signal.

In order to solve the above problems, the present invention provides a video coding method and apparatus using an in-loop filter.

The video signal processing method and apparatus according to the present invention can improve video signal coding efficiency by using an in-loop filter.

FIG. 1 is a diagram showing a block boundary between two different blocks (P block and Q block), and the corresponding block boundary can be classified into a boundary in a vertical direction and a boundary in a horizontal direction.
2 is a diagram illustrating a block boundary between two different blocks (P block and Q block), and the corresponding block boundary can be classified into a vertical boundary and a horizontal boundary.
3 shows a deblocking filter method.

The present invention relates to a method and apparatus for using an in-loop filter in video coding techniques.

More specifically, it relates to a method and apparatus for applying the same filter by integrating a filter applied to a luminance component and a chrominance component in applying a deblocking filter among in-loop filter technologies.

The present invention relates to a method and apparatus for applying a filter applied to a luminance component and a chrominance component in applying a deblocking filter, and a short strong deblocking filter applied to a luminance component and a short strong deblocking filter applied to a chrominance component. We propose a method and apparatus for reducing the cost required for hardware implementation by applying a long deblocking filter as one integrated filter, that is, the same filter.

At this time, it is intended to solve the problem of implementing a logarithmic number of similar filters by allowing the integrated filter proposed in the present invention to be commonly applied to the luminance component and the color difference component.

1 is a diagram illustrating a block boundary between two different blocks (P block and Q block), and the corresponding block boundary can be classified into a vertical boundary and a horizontal boundary.

In FIG. 1, a Q block region refers to a region of a current target block on which encoding and/or decoding is currently performed, and a P block region refers to a block spatially adjacent to a Q block as a reconstructed block.

1 shows the maximum number of pixels in the P block area to which the deblocking filter is applied and the maximum number of pixels in the Q block area, and is a diagram for comparison with an extended deblocking filter and an adaptive deblocking filter to be described later. .

In FIG. 1, the Q block area means the area of the current target block on which the current encoding and/or decoding is performed. In one embodiment of the vertical boundary shown in FIG. 1, the first row of the Q block 100 area ( 130) is an example in which a deblocking filter is applied.

Among the four pixels q0, q1, q2, and q3 belonging to the first row, three pixels q0, q1, and q2 adjacent to the vertical boundary are target pixels on which deblocking filtering is performed. However, the present invention is not limited thereto, and deblocking filtering may be performed on only one, two, or four pixels adjacent to a vertical boundary. The q3 pixels and p0 to p3 are not subjected to deblocking filtering, but may be used for the deblocking filtering of the target pixels q0 to q2, which is hereinafter referred to as a reference pixel.

As shown in FIG. 1, in the Q block area, the length i of the target pixel is longer than the length j of the reference pixel. However, the present invention is not limited thereto, and the length i of the target pixel may be set equal to or shorter than the length j of the reference pixel. Here, the values of i and j may be 1, 2, 3, 4, 5, 6 or more natural numbers. This may be a value pre-committed to the encoding/decoding apparatus, or may be determined based on a predetermined encoding parameter. The coding parameters include: CU size/shape, PU or TU subblock size/shape, whether the Q block region is adjacent to a boundary of a predetermined fragment region, prediction mode (eg, MPM-based intra mode, MIP-based intra Mode, merge mode, affine mode, AMVP mode, Triangle partitioning mode, etc.), component type (luminance, color difference), conversion type (eg, DCT2, DCT8, DST7, conversion skip), filter intensity, filter type, multiple pixels It may include the amount of change or difference between the two. The fragment area may mean at least one of a sub picture, a tile, a slice, a brick, and a CTU. The plurality of pixels may belong to at least one of a Q block area and a P block area. All of the plurality of pixels may be a target pixel or a reference pixel, and at least one of the plurality of pixels may belong to the target pixel.

In the Q block area, the number (n) of filtering target pixels in the first row and the number of filtering target pixels in the second row (m) may be different. n may be set to a value larger than m or may be set to a smaller value. Here, the first row may be any one of a plurality of rows in the Q block area, and the second row may be any one of a plurality of remaining rows except for the first row. For example, the first row may mean a row located at the top and/or bottom of the Q block area.

It goes without saying that the above-described embodiments can be applied in the same/similar manner to the embodiments described later.

In addition, in the example in which the deblocking filter is applied to the first column 131 of the Q block 101 region in the embodiment of the horizontal boundary illustrated in FIG. 1, four pixels (q0) belonging to the first column are similarly applied. Three pixels (q0, q1, q2) adjacent to the horizontal boundary among q1, q2, and q3) are target pixels on which deblocking filtering is performed.

However, in performing the deblocking filter on the corresponding pixels, filtering may be performed by referring to a pixel value of q3, not a target pixel on which the deblocking filtering is performed.

In FIG. 1, in illustrating an example in which the deblocking filter is applied to the Q block region, the first row 130 and the first column 131 are representatively shown, and subsequent rows belonging to the Q block region including the first row The deblocking filter is similarly performed on subsequent columns (second column, third column, etc.) belonging to the Q block region including (second row, third row, etc.) and first column.

In FIG. 1, the P block region refers to a block region spatially adjacent to a vertical boundary or a horizontal boundary of a current target block on which encoding and/or decoding is currently performed, and an embodiment of the vertical boundary shown in FIG. In the middle, an example in which the deblocking filter is applied to the first row 130 of the P block 110 area is shown.

Among the four pixels p0, p1, p2, and p3 belonging to the first row, three pixels p0, p1, and p2 adjacent to the vertical boundary are target pixels on which deblocking filtering is performed.

In addition, in an example in which the deblocking filter is applied to the first column 131 of the P block 111 region in the embodiment of the horizontal boundary illustrated in FIG. 1, four pixels belonging to the first column (p0 , p1, p2, and p3), three pixels p0, p1, and p2 adjacent to the horizontal boundary are target pixels on which deblocking filtering is performed.

However, in performing the deblocking filter on the corresponding pixels, filtering may be performed by referring to the pixel value of p3, not the target pixel on which the deblocking filtering is performed.

In FIG. 1, in showing an example in which the deblocking filter is applied to the P block region, the first row 130 and the first column 131 are representatively shown, and subsequent rows belonging to the P block region including the first row The deblocking filter is similarly performed on subsequent columns (second column, third column, etc.) belonging to the P block region including (second row, third row, etc.) and first column.

FIG. 2 is a diagram showing a block boundary between two different blocks (P block and Q block), and the corresponding block boundary can be classified into a vertical boundary and a horizontal boundary.

In FIG. 2, a Q block region refers to a region of a current target block on which encoding and decoding is currently performed, and a P block region refers to a block spatially adjacent to the Q block as a reconstructed block.

2 is a diagram showing the maximum number of pixels in the P block area and the maximum number of pixels in the Q block area to which the extended deblocking filter is applied.

2 is a diagram showing the maximum number of pixels in the P block area to which the extended deblocking filter is applied and the maximum number of pixels in the Q block area. As shown in FIG. 1, the deblocking filter is at most 4 in the P block area and Q block area. Unlike applying the deblocking filter to 3 pixels by referring to 3 pixels, in the extended deblocking filter, the number of pixels referenced in the P block area and Q block area and the deblocking filter are applied as the block size increases. The number of pixels can be increased.

At this time, the number of target pixels of each P block and Q block adjacent to the block boundary to which the deblocking filter is applied may be 5, as in the embodiments of FIGS. 2A and 2B, or As in the examples of (c) and (d), there may be seven.

2A and 2B are diagrams illustrating an embodiment in which a deblocking filter is applied to 5 pixels by referring to 6 pixels in a P block area and a Q block area according to the size of a current block.

In (a) of FIG. 2, the Q block region refers to the region of the current target block on which encoding and/or decoding is currently performed, and in one embodiment of the vertical boundary shown in (a) of FIG. 2, the Q block An example in which the deblocking filter is applied to the first row 230 of the region (200) is shown.

Among the six pixels (q0, q1, q2, q3, q4, q5) belonging to the first row, five pixels (q0, q1, q2, q3, q4) adjacent to the vertical boundary are target pixels to which deblocking filtering is performed. .

In addition, in an example in which the deblocking filter is applied to the first column 231 of the Q block 201 region among the embodiments of the horizontal boundary shown in FIG. 2B, 6 belonging to the first column is similarly applied. Among the three pixels q0, q1, q2, q3, q4, and q5, five pixels (q0, q1, q2, q3, and q4) adjacent to the horizontal boundary are target pixels on which deblocking filtering is performed.

However, in performing the deblocking filter on the corresponding pixels, filtering may be performed by referring to a pixel value of q5, not a target pixel to which the deblocking filtering is performed.

In (a) of FIG. 2, the P block region refers to a block region spatially adjacent to a vertical boundary or a horizontal boundary of a current target block on which encoding and/or decoding is currently performed, and shown in (a) of FIG. An example of applying a deblocking filter to the first row 230 of the P block 210 region, among embodiments of the vertical boundary, is shown.

Among the six pixels (p0, p1, p2, p3, p4, p5) belonging to the first row, five pixels (p0, p1, p2, p3, p4) adjacent to the vertical boundary are the target pixels on which deblocking filtering is performed. .

In addition, in an example in which the deblocking filter is applied to the first column 231 of the P block 211 region among the embodiments of the horizontal boundary illustrated in FIG. 2B, 6 belonging to the first column is similarly applied. Among the three pixels p0, p1, p2, p3, p4, and p5, five pixels p0, p1, p2, p3, and p4 adjacent to the horizontal boundary are target pixels on which deblocking filtering is performed.

However, in performing the deblocking filter on the corresponding pixels, filtering may be performed by referring to a pixel value of p5, which is not a target pixel on which the deblocking filtering is performed.

2A and 2B illustrate an example in which the deblocking filter is applied to the P block region, the first row 230 and the first column 231 are representatively shown, including the first row. Accordingly, the deblocking filter is similarly performed on subsequent rows (second row, third row, etc.) belonging to the P block region and subsequent columns (second column, third column, etc.) belonging to the P block region including the first column.

2A and 2B illustrate an example in which the deblocking filter is applied to the Q block region, the first row 230 and the first column 231 are representatively shown, including the first row. Thus, the deblocking filter is similarly performed on subsequent rows (second row, third row, etc.) belonging to the Q block region and subsequent columns (second column, third column, etc.) belonging to the Q block region including the first column.

2C and 2D are diagrams illustrating an embodiment in which a deblocking filter is applied to 7 pixels by referring to 8 pixels in a P block area and a Q block area according to the size of a current block.

In (c) of FIG. 2, the Q block region refers to the region of the current target block on which encoding and/or decoding is currently performed, and in one embodiment of the vertical boundary shown in (c) of FIG. 2, the Q block An example in which the deblocking filter is applied to the first row 232 of the area (202) is shown.

Among the eight pixels (q0, q1, q2, q3, q4, q5, q6, q7) belonging to the first row, seven pixels (q0, q1, q2, q3, q4, q5, q6) adjacent to the vertical boundary are It is a target pixel on which blocking filtering is performed.

In addition, in an example in which the deblocking filter is applied to the first column 233 of the Q block 203 region in the embodiment of the horizontal boundary illustrated in FIG. 2D, 8 belonging to the first column is similarly applied. Of the three pixels (q0, q1, q2, q3, q4, q5, q6, q7), seven pixels (q0, q1, q2, q3, q4, q5, q6) adjacent to the horizontal boundary are deblocking filtering. It is the target pixel.

However, in performing the deblocking filter on the corresponding pixels, filtering may be performed by referring to a pixel value of q7 which is not a target pixel on which the deblocking filtering is performed.

In FIG. 2(c), the P block area means a block area spatially adjacent to a vertical boundary or a horizontal boundary of a current target block on which encoding and decoding is currently performed, and the vertical direction shown in FIG. 2(c). In one embodiment of the boundary, an example in which a deblocking filter is applied to the first row 232 of the area of the P block 212 is shown.

Among the eight pixels (p0, p1, p2, p3, p4, p5, p6, p7) belonging to the first row, 7 pixels (p0, p1, p2, p3, p4, p5, p6) adjacent to the vertical boundary are It is a target pixel on which blocking filtering is performed.

In addition, in an example in which the deblocking filter is applied to the first column 233 of the P block 213 region among the embodiments of the horizontal boundary illustrated in FIG. 2D, 8 belonging to the first column is similarly applied. Of the number of pixels (p0, p1, p2, p3, p4, p5, p6, p7), 7 pixels (p0, p1, p2, p3, p4, p5, p6) adjacent to the horizontal boundary are subjected to deblocking filtering. It is the target pixel.

However, in performing the deblocking filter on the corresponding pixels, filtering may be performed by referring to a pixel value of p7 other than a target pixel on which the deblocking filtering is performed.

2C and 2D illustrate an example in which the deblocking filter is applied to the P block region, as representative of the first row 232 and the first column 233, including the first row. Accordingly, the deblocking filter is similarly performed on subsequent rows (second row, third row, etc.) belonging to the P block region and subsequent columns (second column, third column, etc.) belonging to the P block region including the first column.

2C and 2D illustrate an example in which the deblocking filter is applied to the Q block region, as representative of the first row 232 and the first column 233, including the first row. Thus, the deblocking filter is similarly performed on subsequent rows (second row, third row, etc.) belonging to the Q block region and subsequent columns (second column, third column, etc.) belonging to the Q block region including the first column.

In applying the long deblocking filter shown in FIG. 2, filtering targets for p pixels and q pixels around a block boundary may be defined as follows.

[Equation 1]

In the above equation, refMiddle refers to a representative value of a pixel adjacent to a block boundary, and refP and refQ refer to representative values of the outermost pixels of the target pixel of the P block and Q block. The refMiddle, refP, and refQ may be defined as follows according to the number of pixels to which the deblocking filter is applied, that is, the filter length.

[Equation 2]

In addition, in the above formula, f _i and g _j refer to the weight values of the target pixels of the P block and the Q block, and the values of t _c PD _i and T _c QD _j control the amount of change in the resulting pixel values after filtering This is a value for clipping that functions as a function. The values of f _i , g _j , t _c PD _i , and T _c QD _j may be defined as follows according to the number of pixels to which the deblocking filter is applied, that is, the filter length.

[Equation 3]

In FIG. 2, a long deblocking filter (or a long-tap deblocking filter) having a longer filter length compared to the existing deblocking filter is illustrated and described. A long deblocking filter (hereinafter referred to as a first filter) refers to a filter having a number of taps, and a short deblocking filter (hereinafter referred to as a second filter) refers to a filter having b number of taps, and a is greater than b. It's a value. a may be 6, 8, 10, 12, 14 or more, and b may be 4, 6, 8, 10, 12 or more. Any one of the first filter and the second filter may be selectively used, and a predetermined flag may be used for the selection. The flag may be derived based on the above-described encoding parameter, or may be signaled by being encoded in an encoding device. In addition, a flag indicating whether at least one of the first filter or the second filter is available in the encoding/decoding process may be defined, and the flag is a video parameter set (VPS), a sequence parameter set (SPS), and a picture parameter set. It may be signaled at a higher level such as (PPS) and slice.

In this case, the long deblocking filter illustrated and described in FIG. 2 can be applied only to the luma component. Alternatively, the long deblocking filter may be set to be applied only to the color difference component. Alternatively, the long deblocking filter may be set to apply equally to the luminance component and the color difference component. The short deblocking filter can only be applied to the luma component. Alternatively, the short deblocking filter may be set to be applied only to the color difference component. Alternatively, the short deblocking filter may be set to apply equally to the luminance component and the color difference component. This setting may be performed based on at least one of the aforementioned encoding parameters.

For the chroma component as well, as the resolution of the extended image and the size of the coding block increase, it is necessary to apply a long deblocking filter.

However, in order to apply the deblocking filter having a maximum length of 7 or 5 as illustrated and described in FIG. 2 to the chroma component, additional image quality deterioration may occur due to the characteristic of the component.

Therefore, as a method of applying a long deblocking filter to the chroma component, three pixels (p0, p1, p2, q0, q1) are each applied to the P block and Q block, similar to the strong filter among the short deblocking filters that were applied to the luma component in the deblocking filter. , q2) can be filtered.

As an embodiment of applying a long deblocking filter for a chroma component, filtering may be performed on each pixel as shown in the following equation.

[Equation 4]

On the other hand, in an embodiment in which a short strong deblocking filter for a luma component is applied, filtering may be performed on each pixel as shown in the following equation.

[Equation 5]

If the long deblocking filter for the chroma component object and the short strong deblocking filter for the luma component are used differently from each other, the operation purpose and execution method of the corresponding filter are similar, but the number/position of pixels referred for filtering or the reference pixel Due to some differences in weight and threshold values for clipping, it was necessary to implement additional modules.

When the same filter is applied to a long deblocking filter for a chroma component and a short strong deblocking filter for a luma component proposed in the present invention, it is possible to reduce the cost required to implement a codec by preventing the implementation of an additional module. To this end, a flag indicating whether to apply the same filter may be used, or a predetermined table defining filter types applied to the luminance/color difference component may be used. Here, the flag may be signaled by the encoding apparatus or may be derived based on at least one of the aforementioned encoding parameters. The table may define filter types of luminance/color difference components for each index. The index may be signaled by the encoding apparatus or may be derived based on at least one of the aforementioned encoding parameters.

In the present invention, as an embodiment of applying the same filter to a long deblocking filter targeting a chroma component and a short strong deblocking filter targeting a luma component, the following equation may be applied.

[Equation 6]

Alternatively, the following equation may be applied as an embodiment in which the same filter is applied to a long deblocking filter targeting a chroma component and a short strong deblocking filter targeting a luma component in the present invention.

[Equation 7]

Alternatively, using a method of applying the same FIR filter of a similar type different from the above equations for the long deblocking filter targeting the chroma component and the short strong deblocking filter targeting the luma component may also be included in the scope of the present invention. .

Claims (1)

Video coding method using an in-loop filter.

Images