145. The reconstructed block or picture stored in the memory 150 may be provided to the prediction module 110 for performing inter prediction.

FIG. 2 is a block diagram schematically illustrating a prediction module according to an exemplary embodiment of the present invention. Referring to FIG. 2, the prediction module 200 may include an inter prediction module 210 and an intra prediction module 220.

The inter prediction module 210 may perform prediction based on information on at least one of previous or subsequent pictures of a current picture to generate a prediction block. The intra prediction module 220 may perform prediction based on information on a pixel in the current picture to generate a prediction block.

The inter prediction module 210 may select a reference picture for a PU, and a reference block having the same size as the PU by an integer pixel sample unit. Subsequently, the inter prediction module 210 may generate a prediction block that is the most similar to the current PU to have a minimum residual signal and a minimum motion vector size, the prediction unit being generated in a sample unit smaller than an integer, such as a ½ pixel sample unit and a ¼ pixel sample unit. Here, the motion vector may be represented as a unit smaller than an integer pixel.

Information on an index of the reference picture selected by the inter prediction module 210 and the motion vector may be encoded and transferred to the decoder.

FIG. 3 is a block diagram schematically illustrating a video decoding apparatus according to an exemplary embodiment of the present invention. Referring to FIG. 3, the video decoding apparatus 300 may include an entropy decoding module 310, a rearrangement module 315, an dequantization module 320, an inverse transform module 325, a prediction module 330, a filter module 335, and a memory 340.

When a video bit stream is input to the video encoding apparatus, the input video bit stream may be decoded according to a procedure by which the video encoding apparatus has processed video information.

The entropy decoding module 310 may perform entropy decoding on the input bit stream, and an entropy decoding method is similar to the entropy encoding method described above. When entropy decoding is applied, a symbol with a high probability of occurrence may be allocated a low-value index and a corresponding short codeword, while a symbol with a low probability of occurrence may be allocated a high-value index and a corresponding long codeword. Thus, the amount of bits for symbols to be coded may be reduced and video compression performance may be enhanced by entropy coding.

Information for generating a prediction block among information decoded by the entropy decoding module 310 may be provided to the prediction module 330, and residual values on which entropy decoding has been performed by the entropy decoding module may be input to the rearrangement module 315.

The rearrangement module 315 may rearrange the bit stream on which entropy decoding has been performed by the entropy decoding module 310 based on a rearrangement method used in the video encoding apparatus. The rearrangement module 315 may rearrange coefficients represented in a one-dimensional vector form by reconstructing the coefficients into coefficients of a two-dimensional block form. The rearrangement module 315 may receive information related to coefficient scanning performed by the encoding apparatus and perform rearrangement using an inverse scanning method based on a scanning order used by the corresponding encoding apparatus.

The dequantization module 320 may perform dequantization based on quantization parameters and coefficient values of a rearranged block provided by the encoding apparatus.

The inverse transform module 325 may perform inverse DCT and/or inverse DST on DCT and DST performed by the transform module of the encoding apparatus, relative to quantization results performed by the video encoding apparatus. Inverse transform can be performed based on a transmission unit or a partition unit of a picture as determined by the encoding apparatus. In the transform module of the encoding apparatus, DCT and/or DST may be selectively performed based on a plurality of factors, such as a prediction method, the size of the current block and/or a prediction direction. The inverse transform module 325 of the decoding apparatus can perform inverse transform based on transform information resulting from the transform module of the encoding apparatus.

The prediction module 330 may generate a prediction block based on information about the generation of the prediction block provided by the entropy decoding module 310 and information on a previously decoded block and/or picture provided by memory 340. A reconstructed block may be generated using the prediction block generated by the prediction module 330 and the residual block provided by the inverse transform module 325.

The reconstructed block and/or picture may be provided to the filter module 335. The filter module 335 may apply an in-loop filter to the reconstructed block and/or picture. The in-loop filter may include a deblocking filter, an SAO and/or an ALF.

The memory 340 may store the reconstructed picture or block so that the reconstructed picture or block may be used as a reference picture or a reference block, and may also supply the reconstructed picture to an output module.

FIG. 4 is a block diagram schematically illustrating a prediction module of the video decoding apparatus according to an exemplary embodiment of the present invention.

Referring to FIG. 4, the prediction module 400 may include an intra prediction module 410 and an inter prediction module 420.

The intra prediction module 410 may generate a prediction block based on information on a pixel in a current picture when a prediction mode for a corresponding PU is an intra prediction mode.

The inter prediction module 420 may perform inter prediction on a current PU based on information included in at least one of previous or subsequent pictures of a current picture including the current PU using information needed for inter prediction of the current PU provided from the video encoding apparatus, for example, information about a motion vector and a reference picture index, when a prediction mode for a corresponding PU is an inter prediction mode.

Here, when a skip flag or merge flag of a CU received from the encoding apparatus is identified, motion information may be derived according to the identified flag.

Hereinafter, when “video” or “screen” indicate the same meaning as “picture” based on a configuration or expression of the invention, “picture” may be replaced with “video” or “screen.” Further, inter prediction refers to inter frame prediction, and intra prediction refers to intra frame prediction.

As described above with reference to FIGS. 2 and 4, the intra prediction module may perform prediction based on the information on the pixel in the current picture to generate the prediction block with respect to the current block. For example, the intra prediction module may predict pixel values of the current block using pixels in reconstructed blocks disposed to top, left, top left and/or top right sides from the current block. Here, a processing unit on which prediction is performed may be different from a processing unit on which a prediction method and a detail thereof is determined. For example, a prediction method and a prediction mode may be determined on a PU, and prediction may be performed on a TU.

Intra prediction may be carried out based on an intra prediction mode for each PU within the current block (for example, CU). The intra prediction mode may include vertical, horizontal, DC, planar and angular modes according to locations of reference pixels used to predict a pixel value of the current block and/or prediction method. In the vertical mode, prediction may be performed in the vertical direction using pixel values of an adjacent block. In the horizontal mode, prediction may be performed in the horizontal direction using pixel values of an adjacent block. In the DC mode, a prediction block may be generated using an average pixel value of reference pixels. In the angular mode, prediction may be performed according to an angle and/or direction defined in advance for each mode. For intra prediction, a predetermined prediction direction and prediction mode value may be used.

When a prediction mode for the current block is an intra mode, the encoding apparatus and the decoding apparatus may use an intra pulse code modulation (I_PCM) mode for lossless encoding, in addition to the intra prediction mode. When the I_PCM mode is applied to the current block, the encoding apparatus may transmit pixel values in the current block to the decoding apparatus as they are without prediction, transform and/or quantization. Here, the pixel values in the current block may be coded, for example, in raster scan order, and transmitted to the decoding apparatus. The decoding apparatus may not perform prediction, inverse scaling, inverse transform and/or dequantization on the I_PCM mode-applied block. Here, the decoding apparatus may parse the pixel values transmitted from the encoding apparatus and derive a pixel value in a reconstructed block straight from the parsed pixel values.

When the I-PCM mode is used, the encoding apparatus may transmit information on whether the I-PCM mode is applied to the current block (for example, CU) to the decoding apparatus. The information may be indicated, for example, by a PCM flag. Here, the PCM flag may be a flag to indicate whether the I-PCM mode is applied to the current CU, i.e., whether the current CU is coded by the I_PCM. In one exemplary embodiment, the flag may be represented by pcm_flag. The encoding apparatus may entropy-encode the flag and transmit the flag to the decoding apparatus. Here, the decoding apparatus may receive and decode the encoded PCM flag and determine whether the I_PCM mode is applied to the current CU using the decoded PCM flag.

As described above, when the pixel values in the current CU are coded by I-PCM, the decoding apparatus may not perform prediction, inverse scaling, inverse transform and/or dequantization on the current CU. Here, the decoding apparatus may parse the pixel values transmitted from the encoding apparatus and derive a reconstructed pixel value from the parsed pixel values. Here, the reconstructed pixel value may refer to a pixel value reconstructed before the in-loop filter is applied.

Meanwhile, a video signal may generally include three color signals that represent sizes of three primary color components of light. The three color signals may be represented by red (R), green (G) and blue (B). To reduce a frequency band used for video processing, the R, G and B signals may be transformed into luma and chroma signals equivalent to the R, G and B signals. Here, the video signal may include one luma signal and two chroma signals. Here, the luma signal is a component to represent a luminance of a screen, while the chroma signals are a component to represent a color of the screen. The luma signal may be represented by L, while the two chroma signals may be represented by Cb and Cr, respectively. Since a human eye is sensitive to the luma signal but insensitive to the chroma signals, one picture or block may include a fewer number of pixels of a chroma component than that of pixels of a luma component. In a 4:2:0 video format, a number of pixels of a chroma component block corresponding to a luma component block may be ½ of a number of pixels of the luma component block in a horizontal direction and ½ of a number of pixels of the luma component in a vertical direction.

When the current CU is encoded by the I_PCM mode, an example of deriving a reconstructed pixel value of a luma component may be represented as follows.

for
(

i
=
0

;

i
<

currCodingUnitSize
*
currCodingUnitSize

;

i
++

)

R
′

⁢

L
[

xB
+

(

i
⁢

%
⁢

currCodingUnitSize

)

,

yB
+

(

i
/
currCodingUnitSize

)

]

=

pcm_sample
⁢

_luma
[
i
]

Here, currCodingUnitSize may denote a size of a current CU. currCodingUnitSize may be initialized to satisfy 1<<log2CUSize, wherein log2CUSize may be a logarithm of the size of the current CU. Further, R′L may be a reconstructed pixel value of a luma component before the in-loop filter (e.g., deblocking filter) is applied, that is, a luma reconstructed pixel value. (xB, yB) may be coordinates of a most top-left pixel in the current CU. pcm_sample_luma may be a pixel value of a luma component encoded in the I_PCM mode with respect to the current CU and transmitted to the decoding apparatus. Here, for instance, the pixel value of the luma component may be encoded in raster scan order.

When the current CU is encoded by the I_PCM mode, another example of deriving a reconstructed pixel value of a luma component may be represented as follows.

recSamples
L

[

xB
+
i

,

yB
+
j

]

=

pcm_sample
⁢

_luma
[

(

nS
*
j

)

+
 
i

]

≪

(

BitDepth
Y

-

PCMBitDepth
Y

)

,

with
⁢

i

,

j
=

0

..

⁢

nS

-
1

Here, nS may denote a size of a current CU. nS may be initialized to satisfy 1<<log2CbSize, wherein log2CbSize may be a logarithm of the size of the current CU. recSamples_Lmay be a reconstructed pixel value of a luma component before the in-loop filter (e.g., deblocking filter) is applied, that is, a luma reconstructed pixel value. (xB, yB) may be coordinates of a top leftmost pixel in the current CU. pcm_sample_luma may be a pixel value of a luma component encoded in the I_PCM mode with respect to the current CU and transmitted to the decoding apparatus. Here, for instance, the pixel value of the luma component may be encoded in raster scan order. Further, PCMBitDepth_Ymay be a bit number used to indicate each pixel corresponding to pcm_sample_luma, and BitDepth_Ymay be a bit depth of pixels of the luma component in the current CU.

When the current CU is encoded by the I_PCM mode, an example of deriving a reconstructed pixel value of a chroma component may be represented as follows.

for
(

i
=
0

;

i
<

currCodingUnitSizeC
*
currCodingUnitSizeC

;

i
++

)

{

R
′

⁢

Cb
[

(

xB
≫
1

)

+

(

i
⁢

%
⁢

currCodingUnitSizeC

)

,

(

yB
≫
1

)

+

(

i
/
currCodingUnitSizeC

)

]

=

pcm_sample
⁢

_chroma
[
i
]

}

for
(

i
=
0

;

i
<

currCodingUnitSizeC
*
currCodingUnitSizeC

;

i
++

)

{

R
′

⁢

Cr
[

(

xB
≫
1

)

+

(

i
⁢

%
⁢

currCodingUnitSizeC

)

,

(

yB
≫
1

)

+

(

i
/
currCodingUnitSizeC

)

]

=

pcm_sample
⁢

_chroma
[

i
+

currCodingUnitSizeC
*
currCodingUnitSizeC

]

}

Here, currCodingUnitSizeC may denote a size of a chroma component block with respect to a current CU. currCodingUnitSizeC may be initialized to satisfy (1<<log2CUSize)>>1, wherein log2CUSize may be a logarithm of the size of the current CU. Further, R′Cb may be a reconstructed pixel value of a chroma component Cb before the in-loop filter (e.g., deblocking filter) is applied, that is, a chroma Cb reconstructed pixel value. R′Cr may be a reconstructed pixel value of a chroma component Cr before the in-loop filter (e.g., deblocking filter) is applied, that is, a chroma Cr reconstructed pixel value. pcm_sample_chroma may be a pixel value of a chroma component encoded in the I_PCM mode with respect to the current CU and transmitted to the decoding apparatus. Here, for instance, the pixel value of the chroma component may be encoded in raster scan order.

When the current CU is encoded by the I_PCM mode, another example of deriving a reconstructed pixel value of a chroma component may be represented as follows.

recSamples
Cb

[

xB
/
2

+
i

,

yB
/
2

+
j

]

=

pcm_sample
⁢

_chroma
[

(

nS
/
2
*
j

)

+
i

]

≪

(

BitDepth
C

-

PCMBitDepth
C

)

⁢

with
⁢

i

,

j
=

0

..

⁢

nS
/
2

-
1

recSamples
Cr

[

xB
/
2

+
i

,

yB
/
2

+
j

]

=

pcm_sample
⁢

_chroma
[

(

nS
/
2
*

(

j
+
nS

)

)

+
i

]

≪

(

BitDepth
C

-

PCMBitDepth
C

)

⁢

with
⁢

i

,

j
=

0

..

⁢

nS
/
2

-
1

Here, recSamples_Cbmay be a reconstructed pixel value of a chroma component Cb before the in-loop filter (e.g., deblocking filter) is applied, that is, a chroma Cb reconstructed pixel value. recSamples_Crmay be a reconstructed pixel value of a chroma component Cr before the in-loop filter (e.g., deblocking filter) is applied, that is, a chroma Cr reconstructed pixel value. pcm_sample_chroma may be a pixel value of a chroma component encoded in the I_PCM mode with respect to the current CU and transmitted to the decoding apparatus. Here, for instance, the pixel value of the chroma component may be encoded in raster scan order. Further, PCMBitDepth_Cmay be a bit number used to indicate each pixel corresponding to pcm_sample_chroma, and BitDepth_Cmay be a bit depth of pixels of the chroma component in the current CU.

According to the foregoing method of deriving the reconstructed pixel value, the decoding apparatus may parse the pixel values transmitted from the encoding apparatus and copy the parsed pixel values (pcm_sample_luma, pcm_sample_chroma) as they are for use as reconstructed pixel values. When the I-PCM mode is applied to the current CU, the encoding apparatus may code the pixel values in the current CU as they are without prediction, transform and/or quantization and transmit the pixel values to the decoding apparatus. Thus, the I_PCM mode is a lossless encoding mode, and the decoding apparatus may parse the transmitted pixel values and derive the reconstructed pixel values straight from the parsed pixel values.

Meanwhile, to make up for a difference between an original picture and a reconstructed picture by an error that occurs a compression encoding process, such as quantization, the in-loop filter may be applied for the reconstructed picture. As described above, in-loop filtering may be carried out by the filter modules of the encoding apparatus and the decoding apparatus, wherein the filter modules may apply at least one of a deblocking filter, an SAO and an ALF for the reconstructed picture.

FIG. 5 is a flowchart schematically illustrating a deblocking filtering process according to an exemplary embodiment of the present invention.

As described above, the encoding apparatus and the decoding apparatus may reconstruct a video based on a block unit. When the video is reconstructed based on the block unit, block distortion may occur on boundaries between blocks in the reconstructed picture. Thus, the encoding apparatus and the decoding apparatus may use a deblocking filter to remove the block distortion occurring on the boundaries between the blocks in the reconstructed picture.

A deblocking filtering process may be applied to edges of any PU and edges of any CU within the picture. Here, the deblocking filtering process may not be applied to edges on the boundary of the picture. Thus, the encoding apparatus and the decoding apparatus may determine an area to be subjected to the deblocking filter. In detail, the encoding apparatus and the decoding apparatus may determine boundaries between CUs and boundaries between PUs to be subjected to the deblocking filter within a current block on the basis of information on the current block (location, width, height, depth, size and prediction partition mode).

Referring to FIG. 5, the encoding apparatus and the decoding apparatus may determine a boundary strength (bS) with respect to a boundary between blocks and/or an edge of a block based on the determined boundaries between the CUs and the determined boundaries between the PUs (S510). Here, bS may denote a boundary strength between blocks. In the following specification, an “edge” of one block may refer to a “boundary” between blocks for convenience unless specified otherwise.

One picture may include an area where block distortion easily occurs and an area where block distortion hardly occurs. Thus, the encoding apparatus and the decoding apparatus may determine a bS with respect to a boundary between blocks in consideration of a probability that block distortion occurs so as to increase encoding efficiency. For example, in the case where a bS value of a boundary (block edge) between a block P and a block Q is calculated, the encoding apparatus and the decoding apparatus may determine the bS value with respect to the boundary based on whether the deblocking filter is applied to the boundary, a direction of the block edge, locations of the blocks and/or whether the block P and the block Q are encoded in the intra mode.

Here, to be mentioned later, the encoding apparatus and the decoding apparatus may perform filtering in different methods on a boundary of an area with a high probability that block distortion occurs and a boundary of an area with a low probability that block distortion occurs within the same picture, thereby enhancing encoding efficiency.

When the bS is determined, the encoding apparatus and the decoding apparatus may determine whether to perform filtering (S520).

In the case where the bS value of the boundary (block edge) between the block P and the block Q is calculated, the encoding apparatus and the decoding apparatus may determine whether to perform filtering on the basis of a linearity between a sample line within the block P and a sample line in the block Q.

Here, the encoding apparatus and the decoding apparatus may determine whether to perform filtering based on a pixel value on the sample line within the block P, a pixel value on the sample line within the block Q, and a variable β. The encoding apparatus may perform calculation according to a preset method using the pixel value on the sample line in the block P and the pixel value on the sample line in the block Q, thereby deriving a positive variable d. Here, when the variable d is smaller than the variable β, the encoding apparatus may determine to perform filtering. Here, the variable β may be determined based on a quantization parameter (QP) of the current CU. The variable β becomes greater as the QP increases. For instance, when the QP is 0, the variable β may be determined to be 0. Thus, when the QP of the current CU is 0, the variable d is always equivalent to or greater than the variable β, and thus the encoding apparatus and the decoding apparatus may determine not to perform filtering.

Returning to FIG. 5, the encoding apparatus and the decoding apparatus may perform deblocking filtering according to the bS value and the determination of whether to perform filtering (S530).

The encoding apparatus and the decoding apparatus may determine a filter to be used for the boundary between the blocks and/or the edge of the block based on the determined bS value. The filter may be classified into a strong filter and a weak filter. That is, the encoding apparatus and the decoding apparatus may determine a strong filter or weak filter as the filter to be used for the boundary between the blocks and/or the edge of the block. Here, the encoding apparatus and the decoding apparatus perform filtering in different methods on a boundary of an area with a high probability that block distortion occurs and a boundary of an area with a low probability that block distortion occurs within the same picture, thereby enhancing encoding efficiency.

When determining to perform filtering in step S520 of determining whether to perform filtering, the encoding apparatus may perform deblocking filtering on the boundary between the blocks and/or the edge of the block using the determined filter (for example, a strong filter or a weak filter).

Meanwhile, the encoding apparatus and the decoding apparatus may not perform deblocking filtering on a pixel in a block encoded in the I_PCM mode. For example, when the current CU is encoded in the I_PCM mode, the encoding apparatus and the decoding apparatus may omit a deblocking filtering process on pixels in the current CU. As described above, when the I_PCM mode is applied, the encoding apparatus may transmit the pixel values in the current block to the decoding apparatus as they are without prediction, transform and/or quantization. Here, the decoding apparatus may derive reconstructed pixel values, the same as those in original data, from the pixel values transmitted from the encoding apparatus. Thus, to reconstruct a video with optimal quality, an in-loop filtering (e.g., deblocking filtering, SAO and ALF) process may need omitting.

In one exemplary embodiment, when the current CU is encoded by I_PCM, the encoding apparatus and the decoding apparatus may determine not to perform deblocking filtering on the pixels in the CU in step S520 of determining whether to perform filtering. To this end, the encoding apparatus and the decoding apparatus may assume that the QP of the current CU is 0 and determine the variable β. Here, the variable β may be determined as 0, and the encoding apparatus and the decoding apparatus may determine not to perform filtering. That is, in this case, the encoding apparatus and the decoding apparatus may omit the deblocking filtering process. Here, whether the current CU is encoded by I_PCM may be determined, for example, based on a PCM flag.

In another exemplary embodiment, the encoding apparatus and the decoding apparatus may determine a reconstructed pixel value, which is not subjected to deblocking filtering, with respect to a pixel in a block encoded in the I_PCM mode (for example, a CU) as a final pixel value in the deblocking filtering process. Accordingly, the encoding apparatus and the decoding process may omit the deblocking filtering process. Hereinafter, a block encoded in the I_PCM mode is referred to as an I_PCM block.

Here, it may be determined based on a PCM loop filter flag whether the reconstructed pixel value not subjected to deblocking filtering with respect to the I_PCM block is determined as the final pixel value. The PCM loop filter flag may be a flag to instruct whether to apply the in-loop filter to the I_PCM block. In one exemplary embodiment, the flag may be pcm_loop_filter_disable_flag. The encoding apparatus may transmit PCM loop filter flag information about a video sequence to the decoding apparatus, and the decoding apparatus may determine whether to apply the deblocking filter based on the transmitted PCM loop filter flag information.

Meanwhile, as described above, the filter modules of the encoding apparatus and the decoding apparatus may adopt an SAO filter as the in-loop filter in addition to the deblocking filter. SAO filtering may be an in-loop filtering process to compensate a deblocking-filtered image by a pixel unit for an offset difference from an original picture. SAO filtering is to compensate for a coding error, which may be caused by quantization. An SAO may be classified into two types, which are a band offset and an edge offset.

FIG. 6 is a diagram schematically illustrating a band offset.

To apply a band offset, pixels in an SAO-applied unit may be classified according to an intensity of each pixel. A whole intensity range may be divided into a predetermined number of intensity intervals, i.e., bands. Each band may include pixels having an intensity within each intensity interval. Here, an offset may be determined for each band.

A video formed of N-bit pixels may have an intensity range of 0 to 2^N−1. For instance, 8-bit pixels may have an intensity range of 0 to 255. FIG. 6 shows that a whole intensity range is divided into 32 bands having the same intensity interval.

Referring to FIG. 6, each band may have an intensity range of, for example, 8. The 32 bands may be divided into a central first group and a neighboring second group. The first group may include 16 bands, and the second group may also include 16 bands. An offset may be applied for each band, and an offset value for each band may be transmitted to the decoding apparatus.

The decoding apparatus may group pixels and apply the transmitted offset value for each band in the same manner as the encoding apparatus applies a band offset.

A second SAO type is an edge offset mode considering edge information by a pixel. The edge offset mode may be applied in consideration of an edge direction based on a current pixel, and intensities of the current pixels and neighboring pixels.

There are four representative directional edges in a block. The four directional edges may include, for example, a 0-degree edge, a 90-degree edge, a 135-degree edge and a 45-degree edge. Thus, the edge offset for each filtering unit based on an angle or direction of an edge, that is, four types for an SAO application unit, may be used. Hereinafter, the four kinds of edges for the SAO application unit may be referred to as edge types of the edge offset for convenience of description.

FIG. 7 is a block diagram illustrating four representative edge types of the edge offset based on a current pixel (C). Referring to FIG. 7, (a) is a one-dimensional (1D) 0-degree edge, (b) is a 1D 90-degree edge, (c) is a 1D 135-degree edge, and (d) is a 1D 45-degree edge. Four edge offsets may be used based on the four directional edge types shown in FIG. 7. That is, an offset corresponding to one of the four edge types may be applied to each SAO application unit.

When an edge type is determined, a relationship between the current pixel and a neighboring pixel may be considered to apply an edge offset.

FIG. 8 schematically illustrates that a result of comparing intensities of the current pixel and neighboring pixels is divided into four categories.

FIGS. 8(a) to 8(d) show distributions of the current pixel (C) and neighboring pixels in the respective categories. FIG. 8(a) illustrates a category when intensities of two neighboring pixels are greater than an intensity of the current pixel (C). FIG. 8(b) illustrates a category when an intensity of one of two neighboring pixels is smaller than the intensity of the current pixel. FIG. 8(c) illustrates a category when an intensity of one of two neighboring pixels is greater than the intensity of the current pixel. FIG. 8(d) illustrates a category when the intensities of two neighboring pixels are smaller than an intensity of the current pixel.

For example, FIGS. 8(a) and 8(d) show that the intensity of the current pixel is greater or smaller than the intensity of the neighboring pixels. Further, FIGS. 8(b) and 8(c) may occur when the current pixel is located on a boundary of a particular area.

Table 1 schematically illustrates the four categories shown in FIG. 8.

TABLE 1

Category
Condition

1
Intensity of C < Intensity of 2 neighboring pixels

2
Intensity of C < Intensity of 1 neighboring pixel &

Intensity of C = Intensity of 1 neighboring pixel

3
Intensity of C > Intensity of 1 neighboring pixel &

Intensity of C = Intensity of 1 neighboring pixel

4
Intensity of C > Intensity of 2 neighboring pixels

0
None of the above

In Table 1, C represents the current pixel. Also, category 1 corresponds to FIG. 8(a), category 2 corresponds to FIG. 8(b), category 3 corresponds to FIG. 8(c), and category 4 corresponds to FIG. 8(d).

The encoding apparatus may transmit an edge offset value in each category. The decoding apparatus may add each pixel and an edge offset value corresponding to an edge type and a category to reconstruct a pixel. For instance, the decoding apparatus may determine which mode the current pixel belongs to among the four edge types of FIG. 7 and which category the current pixel belongs to among the categories of Table 1 and apply an offset in the corresponding category to the current pixel.

Meanwhile, the SAO process is a process that is performed on the entire slice and/or picture. If the I_PCM mode is applied to the current CU, the encoding apparatus may transmit the pixel values in the current CU to the decoding apparatus as they are without prediction, transform and/or quantization. Here, since the decoding apparatus can derive reconstructed pixel values the same as those of the original data from the pixel values transmitted from the encoding apparatus, it may be more efficient not to perform the SAO process.

Thus, the encoding apparatus and the decoding apparatus may not perform the SAO process on pixels within a block encoded in the I_PCM mode during the SAO application process. That is, the encoding apparatus and the decoding apparatus may omit the SAO process on the pixels in the I_PCM block. Omission of the SAO process based on whether the block is in the I_PCM mode may be applied to at least one of the band offset type and the edge offset type. Here, it may be determined based on, for example, the PCM flag and/or PCM loop filter flag, whether the SAO is applied.

For instance, the encoding apparatus and the decoding apparatus may determine not to perform the SAO process on a pixel in a block encoded in the I_PCM mode. Further, the encoding apparatus and the decoding apparatus may determine a reconstructed pixel value not subjected to the SAO as a final pixel value in the SAO process with respect to the pixel in the block encoded in the I_PCM mode.

Meanwhile, as described above, the filter modules of the encoding apparatus and the decoding apparatus may also employ an ALF as an in-loop filter in addition to the deblocking filter and the SAO. The ALF may perform filtering based on a value obtained by comparing a reconstructed picture via deblocking filtering and/or SAO filtering and an original picture.

FIG. 9 is a flowchart schematically illustrating an adaptive loop filtering process according to an exemplary embodiment of the present invention.

An ALF may compensate for a coding error using a Wiener filter and be applied globally in a slice. The ALF may be applied after the SAO is applied and be applied only in high efficiency (HE).

Referring to FIG. 9, the encoding apparatus and the decoding apparatus may determine whether to apply adaptive loop filtering (S910).

Information about whether to perform filtering may be included in a bit stream and transmitted from the encoding apparatus to the decoding apparatus. For instance, the encoding apparatus may transmit ALF flag information indicating whether adaptive loop filtering is performed on each CU to the decoding apparatus. Here, the ALF flag information may be a flag to instruct whether adaptive loop filtering is performed on the current CU and be represented by, for example, alf_cu_flag. That is, the ALF flag information may be CU-unit on/off information about the ALF. The decoding apparatus may determine whether to apply the ALF for each CU unit using the transmitted ALF flag information.

The information about whether to perform filtering may include CU-unit control flag information. That is, the encoding apparatus may include the CU-unit control flag information in the bit stream to transmit the information to the decoding apparatus. Here, the CU-unit control flag may be a flag to determine whether to perform filtering by each CU, i.e., to instruct whether the ALF is adaptively applied to each CU. For example, the CU-unit control flag may be represented by alf_cu_control_flag.

For example, when alf_cu_control_flag has a value of 1, the encoding apparatus and the decoding apparatus may determine whether to apply the ALF to the current CU based on a value of an ALF flag (for example, alf_cu_flag). That is, when alf_cu_control_flag has a value of 1, the encoding apparatus and the decoding apparatus may turn on/off the ALF process by the CU based on the value of the ALF flag. When alf_cu_control_flag has a value of 0, a CU corresponding to alf_cu_control_flag may not have ALF flag information. That is, there is no CU-unit on/off information about the ALF.

Meanwhile, the encoding apparatus and the decoding apparatus may not perform the ALF process on a pixel in a block encoded in the I_PCM mode. That is, the current CU is encoded in the I_PCM mode, the encoding apparatus and the decoding apparatus may omit the ALF process on the pixels in the current CU. As described above, when the I_PCM mode is applied, the encoding apparatus may transmit the pixel values in the current CU to the decoding apparatus as they are without prediction, transform and/or quantization. Here, the decoding apparatus can derive reconstructed pixel values the same as those of the original data from the pixel values transmitted from the encoding apparatus. Thus, in-loop filtering (for example, deblocking filtering, SAO and ALF) may be omitted to reconstruct a video with optimal quality.

In one embodiment, when the current CU is encoded in the I_PCM mode, the encoding apparatus and the decoding apparatus may determine not to perform the ALF process to the pixels in the CU. Here, it may be determined based on, for example, the PCM flag and PCM loop filter flag, whether the current CU is encoded in the I_PCM mode. The PCM flag and the PCM loop filter flag have been described above, and descriptions thereof are omitted herein.

When the current CU is encoded in the I_PCM mode, the encoding apparatus and the decoding apparatus may determine whether to apply the ALF based on the value of the CU-unit control flag (for example, alf_cu_control_flag, hereinafter, referred to as “alf_cu_control_flag”) according to a different method.

As described, when alf_cu_control_flag has a value of 1, the encoding apparatus and the decoding apparatus may determine whether to apply the ALF to the current CU based on the value of the ALF flag (for example, alf_cu_flag, hereinafter, referred to “alf_cu_flag”). Here, the encoding apparatus may allocate a value of 0 to alf_cu_flag with respect to the CU encoded in the I_PCM mode to transmit alf_cu_flag to the decoding apparatus. In this case, the decoding apparatus may determine whether to apply the ALF based on the value of alf_cu_flag, thereby omitting the ALF process on the CU encoded in the I_PCM mode. Alternatively, when alf_cu_control_flag has a value of 1, the encoding apparatus and the decoding apparatus may ignore the value of alf_cu_flag and the ALF process with respect to the CU encoded in the I_PCM mode.

When alf_cu_control_flag has a value of 0, a CU corresponding to alf_cu_control_flag may not have ALF flag information. That is, there is no CU-unit on/off information about the ALF. Here, the encoding apparatus and the decoding apparatus may infer that the value of alf_cu_flag is 0 with respect to the CU encoded in the I_PCM mode and omit the ALF process.

Returning to FIG. 9, the encoding apparatus and the decoding apparatus may determine a filter shape and a filter coefficient (S920).

When the ALF process is determined to be performed, the encoding apparatus and the decoding apparatus may determine a filter shape and a filter coefficient. Information about the filter shape and/or filter coefficient may be transmitted through a slice header from the encoding apparatus to the decoding apparatus. For example, the filter coefficient may be calculated in the encoding apparatus, where one or more filter coefficients may be provided.

When the filter shape and the filter coefficient are determined, the encoding apparatus and the decoding apparatus may perform adaptive loop filtering on a pixel of a reconstructed picture based on the determined filter shape and filter coefficient (S930). Here, the reconstructed picture may be a video which has been subjected to deblocking filtering and/or SAO process.

FIG. 10 is a flowchart schematically illustrating an operation of the encoding apparatus in a system to which the present invention is applied.

Referring to FIG. 10, the encoding apparatus may reconstruct a block (S1010). For example, the encoding apparatus may transform, quantize a prediction block and dequantize and inverse-transform a residual block generated based on the prediction block and a current block, thereby generating a reconstructed residual block. The encoding apparatus may generate a reconstructed block based on the reconstructed residual block and the prediction block. Alternatively, when the I_PCM mode is applied to the current block, the encoding apparatus may not perform prediction, transform and/or quantization. Here, the encoding apparatus may code pixel values in the current block as they are to transmit the pixel values to the decoding apparatus and generate a reconstructed block directly from the pixel values in the current block.

Subsequently, the encoding apparatus may apply an in-loop filter to the reconstructed block (S1020). The in-loop filter may be used in the filter module of FIG. 1, wherein the filter module may employ a deblocking filter, an SAO or an ALF. Here, the encoding apparatus may not apply at least one of the deblocking filter, the SAO and the ALF to a block encoded in the I_PCM mode and/or pixels in the I_PCM block, which has been described in detail as above.

The encoding apparatus may transmit a bit stream including deblocking filter, SAO and/or ALF-applied video information and video information about the in-loop filter to the decoding apparatus (S1030).

FIG. 11 is a flowchart schematically illustrating an operation of the decoding apparatus in a system to which the present invention is applied.

Referring to FIG. 11, the decoding apparatus may receive a bit stream from the encoding apparatus (S1110). The received bit stream may include not only video information but information needed for reconstruction of the video information.

The decoding apparatus may reconstruct a block based on the received information (S1120). For example, the decoding apparatus may generate a reconstructed block based on a prediction block generated by prediction and a residual block generated via dequantization and inverse transform. Alternatively, when the I_PCM mode is applied to a current block, the decoding apparatus may not perform prediction, transform and/or quantization. Here, the encoding apparatus may code pixel values in the current block as they are to transmit the pixel values to the decoding apparatus, while the decoding apparatus may generate the reconstructed block directly from the transmitted pixel values.