KR101604033B1 - Method and apparatus for decoding an image - Google Patents

Abstract

A method and apparatus for coding and decoding a residual block are disclosed. According to an embodiment of the present invention, a transformed residual block obtained by transforming a residual block into a frequency domain is divided into predetermined frequency band units, and a valid coefficient flag indicating whether an effective transform coefficient exists for each divided frequency band unit .

Description

[0001] The present invention relates to a video decoding method and apparatus,

More particularly, the present invention relates to encoding and decoding of residual blocks.

Background of the Invention [0002] As the development and dissemination of hardware capable of playing back and storing high-resolution or high-definition video content increases the need for video codecs to effectively encode or decode high-definition or high-definition video content. According to the existing video codec, video is encoded according to a limited encoding method based on a macroblock of a predetermined size. Also, existing video codecs encode residual blocks using small-sized conversion units such as 4x4 and 8x8.

The present invention is intended to provide a method and apparatus for more efficiently encoding and decoding effective transform coefficient information in a large-sized transform residual block.

According to another aspect of the present invention, there is provided a residual block coding method including: generating a prediction block of a current block; Generating a residual block which is a difference between the prediction block and the current block; Generating a transformed residual block by transforming the residual block into a frequency domain; Dividing the transform residual block into predetermined frequency band units; And encoding an effective coefficient flag for each frequency band unit indicating whether or not a non-zero effective transform coefficient exists for each of the divided frequency band units.

An apparatus for encoding a residual block according to an exemplary embodiment of the present invention includes: a prediction unit for generating a prediction block of a current block; A subtractor for generating a residual block which is a difference between the prediction block and the current block; A transform unit for transforming the residual block into a frequency domain to generate a transform residual block; And an entropy encoding unit for dividing the transformed residual block by frequency bands and encoding an effective coefficient flag for each band indicating whether there is a non-zero effective transform coefficient for each of the divided frequency bands.

A residual block decoding method according to an embodiment of the present invention extracts an effective coefficient flag indicating whether an effective transform coefficient exists for each predetermined frequency band unit obtained by dividing a transformed residual block of a current block from an encoded bit stream step; Dividing the transform residual block by a predetermined frequency band unit; And determining a frequency band unit in which effective conversion coefficients exist among the frequency band units obtained by dividing the conversion residual block using the effective coefficient flag.

The apparatus for decoding a residual block according to an embodiment of the present invention decodes an effective coefficient indicating whether an effective transform coefficient exists for each frequency band unit obtained by dividing a transform residual block of a current block into a predetermined frequency band unit from an encoded bit stream A parsing unit for extracting a flag; And an entropy decoding unit for dividing the transformed residual block by a predetermined frequency band unit and determining a frequency band unit in which effective transform coefficients exist among the frequency band units using the extracted effective coefficient flag, do.

According to the present invention, it is possible to skip a scanning process in a frequency band in which no effective transform coefficient exists in a transform residual block by generating an effective coefficient flag indicating whether an effective transform coefficient exists in each divided frequency band unit, It is possible to reduce the amount of bits required for coding coefficients.

1 is a block diagram of an image encoding apparatus according to an embodiment of the present invention.
2 is a block diagram of an image decoding apparatus according to an embodiment of the present invention.
FIG. 3 illustrates a hierarchical encoding unit according to an embodiment of the present invention.
4 is a block diagram of an image encoding unit based on an encoding unit according to an embodiment of the present invention.
5 is a block diagram of an image decoding unit based on an encoding unit according to an embodiment of the present invention.
FIG. 6 illustrates a depth-based coding unit and a prediction unit according to an embodiment of the present invention.
FIG. 7 shows a relationship between an encoding unit and a conversion unit according to an embodiment of the present invention.
FIG. 8 illustrates depth-specific encoding information, in accordance with an embodiment of the present invention.
FIG. 9 shows a depth encoding unit according to an embodiment of the present invention.
FIGS. 10A to 10C show a relationship between an encoding unit, a prediction unit, and a frequency conversion unit according to an embodiment of the present invention.
FIG. 11 shows encoding information for each encoding unit according to an embodiment of the present invention.
12A to 12C are views for explaining a process of encoding a transform residual block in a technical field related to the present invention.
13 is a block diagram showing the configuration of a residual block coding apparatus 1300 according to an embodiment of the present invention.
14A to 14J are diagrams illustrating embodiments in which a transform residual block is divided into predetermined frequency band units according to the present invention.
15A and 15B are reference diagrams for explaining the coding process of the effective transform coefficients according to an embodiment of the present invention.
16A and 16B are views for explaining the encoding process of the residual block according to an embodiment of the present invention.
17A and 17B are reference diagrams showing embodiments of encoding information of the conversion residual block generated by the effective coefficient encoding unit 1330. FIG.
18 is a flowchart illustrating a residual block coding method according to an embodiment of the present invention.
19 is a block diagram illustrating an apparatus for decoding residual blocks according to an embodiment of the present invention.
20 is a flowchart illustrating a method of decoding a residual block according to an embodiment of the present invention.

Hereinafter, an image encoding apparatus, an image decoding apparatus, an image encoding method, and an image decoding method according to preferred embodiments of the present invention will be described with reference to the accompanying drawings.

1 is a block diagram of an image encoding apparatus according to an embodiment of the present invention.

1, an image encoding apparatus 100 according to an exemplary embodiment of the present invention includes a maximum encoding unit division unit 110, an encoding depth determination unit 120, and an output unit 130.

The maximum coding unit division unit 110 divides the current picture or the current slice based on the maximum coding unit which is the coding unit of the maximum size. The current picture or current slice is divided into at least one maximum encoding unit. The divided image data may be output to the coding depth determination unit 120 for each of at least one maximum coding unit.

According to an embodiment of the present invention, an encoding unit can be expressed using a maximum encoding unit and a depth. The maximum coding unit indicates the coding unit having the largest size among the coding units of the current picture, and the depth indicates a step in which the coding units are hierarchically divided. The deeper the depth, the more the encoding unit can be divided from the maximum encoding unit to the minimum encoding unit. The depth of the largest encoding unit is the highest depth, and the depth of the minimum encoding unit can be defined as the lowest depth. As the depth increases, the size of the depth-dependent coding unit decreases, so that the sub-coding unit of k depth may include a plurality of sub-coding units having a depth of k + 1 or more.

As described above, according to the maximum size of an encoding unit, the image data of the current picture is divided into a maximum encoding unit, and each maximum encoding unit may include encoding units divided by depth. Since the maximum encoding unit according to an embodiment of the present invention is divided by depth, image data of a spatial domain included in the maximum encoding unit can be hierarchically classified according to depth.

The maximum depth for limiting the total number of times the height and width of the maximum encoding unit can be hierarchically divided and the maximum size of the encoding unit may be preset. The maximum encoding unit and the maximum depth may be set in units of pictures or slices. That is, each picture or slice may have a different maximum coding unit and maximum depth, and the minimum coding unit size included in the maximum image coding unit may be variably set according to the maximum depth. By setting the maximum coding unit and the maximum depth for each picture or slice in this manner, it is possible to improve the compression ratio by coding the image of the flat area using a larger maximum coding unit, and the image with a large complexity can be encoded The compression efficiency of the image can be improved by using the unit.

The encoding depth determination unit 120 encodes at least one divided area in which the area of the maximum encoding unit is divided for each depth, and determines the depth at which the final encoding result is output for each of at least one of the divided areas. That is, the coding depth determination unit 120 selects the depth at which the smallest coding error occurs, and determines the coding depth as the coding depth by coding the image data in units of coding per depth for each maximum coding unit. The coding depth may be determined based on a Rate-Distortion Cost calculation. The determined coding depth and the image data of each coding unit are output to the output unit 130.

The image data in the maximum encoding unit is encoded based on the depth encoding unit according to at least one depth below the maximum depth, and the encoding results based on the respective depth encoding units are compared. As a result of the comparison of the encoding error of the depth-dependent encoding unit, the depth with the smallest encoding error can be selected. At least one coding depth may be determined for each maximum coding unit.

As the depth of the maximum encoding unit increases, the encoding unit is hierarchically divided and reduced, and the number of encoding units increases. In addition, even if encoding units of the same depth included in one maximum encoding unit, the encoding error of each data is measured and it is determined whether or not the encoding unit is divided into lower depths. Therefore, even if the data included in one maximum coding unit has a different coding error according to the position, the coding depth can be determined depending on the position. In other words, the maximum encoding unit may be divided into sub-encoding units of different sizes according to different depths. One or more coding depths may be set for one maximum coding unit and data of the maximum coding unit may be divided according to the coding units of one or more coding depths.

Further, the sub-encoding units of different sizes included in the maximum encoding unit can be predicted or frequency-converted based on the processing units of different sizes. In other words, the image encoding apparatus 100 can perform a plurality of processing steps for image encoding based on various sizes and processing units of various types. In order to encode video data, processing steps such as prediction, frequency conversion, and entropy encoding are performed. Processing units of the same size may be used in all stages, and processing units of different sizes may be used in each step.

For example, in order to predict a coding unit, the image coding apparatus 100 may select a coding unit and a different processing unit. For example, when the size of the encoding unit is 2Nx2N (where N is a positive integer), the processing unit for prediction may be 2Nx2N, 2NxN, Nx2N, NxN, and the like. In other words, predictive coding may be performed based on a processing unit of a type in which at least one of the height or the width of an encoding unit is halved. Hereinafter, a data unit serving as a basis of predictive encoding is referred to as a 'prediction unit'.

The prediction mode may be at least one of an intra mode, an inter mode, and a skip mode, and the specific prediction mode may be performed only for a prediction unit of a specific size or type. For example, the intra mode can be performed only for a 2Nx2N, NxN sized prediction unit, which is a square. In addition, the skip mode can be performed only for a prediction unit of 2Nx2N size. If there are a plurality of prediction units in an encoding unit, a prediction mode having the smallest coding error can be selected by performing prediction for each prediction unit.

Also, the image encoding apparatus 100 can frequency-convert image data based on a processing unit having a different size from the encoding unit. The frequency conversion may be performed based on a data unit having a size smaller than or equal to the encoding unit for frequency conversion of the encoding unit. Hereinafter, a processing unit serving as a basis of frequency conversion is referred to as a 'conversion unit'.

The coding depth determiner 120 measures the coding error of each depth coding unit using a Lagrangian Multiplier based Rate-Distortion Optimization technique and calculates the maximum coding unit having the optimal coding error Can be determined. In other words, the coding depth determiner 120 can determine which type of the maximum coding unit is divided into a plurality of sub-coding units, wherein the plurality of sub-coding units have different sizes depending on the depth.

The output unit 130 outputs, in the form of a bit stream, video data of the maximum encoding unit encoded based on at least one encoding depth determined by the encoding depth determination unit 120 and information on the depth encoding mode.

The information on the depth-dependent coding mode may include coding depth information, partition type information of a prediction unit of a coding unit of coding depth, prediction mode information per prediction unit, size information of a conversion unit, and the like.

The coding depth information can be defined using depth division information indicating whether or not coding is performed at the lower depth coding unit without coding at the current depth. If the current depth of the current encoding unit is the encoding depth, the current encoding unit is encoded in the current depth encoding unit, so that the division information of the current depth can be defined so as not to be further divided into lower depths. On the other hand, if the current depth of the current encoding unit is not the encoding depth, the encoding using the lower depth encoding unit should be tried. Therefore, the division information of the current depth may be defined to be divided into the lower depth encoding units.

If the current depth is not the encoding depth, encoding is performed on the encoding unit divided into lower-depth encoding units. Since there are one or more lower-level coding units in the current-depth coding unit, the coding is repeatedly performed for each lower-level coding unit so that recursive coding can be performed for each coding unit of the same depth.

At least one coding depth is determined in one maximum coding unit and at least one coding mode information is determined for each coding depth so that information on at least one coding mode can be determined for one maximum coding unit. In addition, since the data of the maximum encoding unit is hierarchically divided according to the depth and the depth of encoding may be different for each position, information on the encoding depth and the encoding mode can be set for the data.

Therefore, the output unit 140 according to the embodiment can set the corresponding encoding information for each minimum encoding unit included in the maximum encoding unit. That is, the coding unit of the coding depth includes at least one minimum coding unit that holds the same coding information. By using this, if the neighboring minimum encoding units have the same depth encoding information, it can be the minimum encoding unit included in the same maximum encoding unit.

For example, the encoding information output through the output unit 130 may be classified into encoding information per depth unit and encoding information per prediction unit. The under-coding-by-depth coding unit information may include prediction mode information and partition size information. The encoding information to be transmitted for each prediction unit includes information about the estimation direction of the inter mode, information about the reference picture index of the inter mode, information on the motion vector, information on the chroma component of the intra mode, information on the interpolation mode of the intra mode And the like. Information on the maximum size of a coding unit defined for each picture, slice or GOP, and information on the maximum depth can be inserted into the header of the bitstream.

According to the simplest embodiment of the image coding apparatus 100, the depth-dependent coding unit is a coding unit having a height divided by the height and width of the coding unit of one layer higher depth. That is, if the size of the encoding unit of the current depth (k) is 2Nx2N, the size of the encoding unit of the lower depth (k + 1) is NxN. Therefore, the current encoding unit of 2Nx2N size can include a maximum of 4 sub-depth encoding units of NxN size.

Therefore, the image decoding apparatus 100 according to an exemplary embodiment can determine an optimum shape division type for each maximum encoding unit based on the size and the maximum depth of the maximum encoding unit determined in consideration of the characteristics of the current picture. In addition, since each encoding unit can be encoded by various prediction modes, frequency conversion methods, and the like, an optimal encoding mode can be determined in consideration of image characteristics of encoding units of various image sizes.

If an image having a very high image resolution or a very large data amount is encoded in units of a conventional 16x16 macroblock, the number of macroblocks per picture becomes excessively large. This increases the amount of compression information generated for each macroblock, so that the burden of transmission of compressed information increases and the data compression efficiency tends to decrease. Therefore, the image encoding apparatus according to the embodiment of the present invention can increase the maximum size of the encoding unit in consideration of the image size, and adjust the encoding unit in consideration of the image characteristic, so that the image compression efficiency can be increased .

2 is a block diagram of an image decoding apparatus according to an embodiment of the present invention.

2, an apparatus 200 for decoding an image according to an exemplary embodiment of the present invention includes a receiving unit 210, an image data and coding information extracting unit 220, and an image data decoding unit 230.

The receiving unit 210 receives and parses a bitstream of the encoded video. The image data and encoding information extracting unit 220 extracts image data for each maximum encoding unit from the parsed bit stream and outputs the extracted image data to the image data decoding unit 230. The image data and encoding information extracting unit 220 may extract information on the maximum encoding unit of the current picture or slice from the header of the current picture or slice. Also, the image data and encoding information extracting unit 220 extracts information on the encoding depth and the encoding mode for each maximum encoding unit from the parsed bitstream. The information on the extracted coding depth and coding mode is output to the image data decoding unit 230.

Information on the coding depth and the coding mode per coding unit can be set for one or more coding depth information, and the information on the coding mode for each coding depth includes information on partition type information, prediction mode information, Size information of the conversion unit, and the like. In addition, as the encoding depth information, depth-based segmentation information may be extracted.

The information on the division type of the maximum encoding unit may include information on sub-encoding units of different sizes according to the depth included in the maximum encoding unit, the information on the encoding mode may include information on a prediction unit for each sub- Information on the prediction mode, information on the conversion unit, and the like.

The image data decoding unit 230 decodes the image data of each maximum encoding unit based on the information on the encoding depth and the encoding mode for each maximum encoding unit to reconstruct the current picture. The image data decoding unit 230 can decode the sub-encoding unit included in the maximum encoding unit based on the information on the division type of the maximum encoding unit. The decoding process may include a prediction process including intra prediction and motion compensation, and an inverse frequency conversion process.

The image data decoding unit 230 performs intra prediction or motion prediction in each prediction unit and prediction mode for each coding unit on the basis of the division type information and the prediction mode information of the prediction unit of the coding unit for each coding depth, Compensation can be performed. In addition, the image data decoding unit 230 may perform inverse conversion on each conversion unit for each encoding unit, based on the size information of the conversion unit of the encoding unit by encoding depth, for inverse conversion for each maximum encoding unit.

The image data decoding unit 230 can determine the coding depth of the current maximum encoding unit using the division information by depth. If the partition information indicates that the current depth is to be decoded, the current depth is the depth of the encoding. Therefore, the image data decoding unit 230 can decode the current depth encoding unit for the image data of the current maximum encoding unit using the partition type, the prediction mode, and the conversion unit size information of the prediction unit. That is, it is possible to observe the encoding information set for the minimum encoding unit and to decode the minimum encoding units that hold the encoding information including the same division information, into one data unit.

The image decoding apparatus 200 according to an exemplary embodiment recursively performs encoding for each maximum encoding unit in the encoding process to obtain information on the encoding unit that has generated the minimum encoding error and can use it for decoding the current picture have. That is, it is possible to decode video data in the optimal encoding unit for each maximum encoding unit. Accordingly, even if an image with a high resolution or an excessively large amount of data is used, the information on the optimal encoding mode transmitted from the encoding end is used, and the image data is efficiently encoded according to the encoding unit size and encoding mode, Can be decoded and restored.

FIG. 3 illustrates a hierarchical encoding unit according to an embodiment of the present invention.

Referring to FIG. 3, the hierarchical coding unit according to the present invention may include 32x32, 16x16, 8x8, and 4x4 from a coding unit having a width x height of 64x64. In addition to the square-shaped encoding units, there may be encoding units whose width x height is 64x32, 32x64, 32x16, 16x32, 16x8, 8x16, 8x4, 4x8.

3, the resolution is set to 1920 x 1080, the size of the maximum encoding unit is set to 64, and the maximum depth is set to 2 for the video data 310. For the video data 320, the resolution is set to 1920 x 1080, the maximum size of the encoding unit is set to 64, and the maximum depth is set to 4. [ For the video data 330, the resolution is set to 352 x 288, the maximum size of the encoding unit is set to 16, and the maximum depth is set to 2.

It is desirable that the maximum size of the encoding size is relatively large in order to accurately reflect not only the compression ratio but also the image characteristic when the resolution is high or the data amount is large. Therefore, the maximum size of the encoding size of the video data 310 and 320 having a higher resolution than that of the video data 330 can be selected to be 64.

The maximum depth indicates the total number of layers in the hierarchical encoding unit. Therefore, since the maximum depth of the video data 310 is 2, the encoding unit 315 of the video data 310 is encoded from a maximum encoding unit having a major axis size of 64, Units. On the other hand, since the maximum depth of the video data 330 is 2, the encoding unit 335 of the video data 330 has a depth of two layers from the encoding units having the major axis size of 16, Units.

Since the maximum depth of the video data 320 is 4, the encoding unit 325 of the video data 320 has a depth of four layers from 32 to 16, 8, and 4 Encoding units. Since the image is encoded based on a smaller sub-encoding unit as the depth is deeper, it is suitable to encode an image including a finer scene.

4 is a block diagram of an image encoding unit based on an encoding unit according to an embodiment of the present invention.

The image encoding unit 400 according to an exemplary embodiment of the present invention performs image data encoding operations to be performed in order to determine the encoding depth in the encoding depth determination unit 120 of the image encoding apparatus 100 of FIG.

4, the intra-prediction unit 410 performs intra-prediction on a prediction unit of an intra mode in the current frame 405, and the motion estimation unit 420 and the motion compensation unit 425 perform intra- Inter prediction and motion compensation are performed using the current frame 405 and the reference frame 495 with respect to the unit.

The residual values are generated based on the prediction unit output from the intra prediction unit 410, the motion estimation unit 420 and the motion compensation unit 425, and the generated residual values are input to the frequency conversion unit 430 and the quantization unit 420. [ (440) and output as quantized transform coefficients.

The quantized transform coefficients are restored to a residual value through the inverse quantization unit 460 and the frequency inverse transform unit 470. The restored residual values are passed through the deblocking unit 480 and the loop filtering unit 490, And output to the reference frame 495. [ The quantized transform coefficients are output to the bitstream 455 via the entropy encoding unit 450. 12, the entropy encoding unit 450 according to an exemplary embodiment of the present invention converts a residual block, which is a difference value between a prediction block and an original block, into a frequency domain, Frequency band units and encodes the effective coefficient flag for each frequency band unit indicating whether there is a non-zero effective conversion coefficient for each divided frequency band unit.

In order to perform encoding according to the image encoding method according to an embodiment of the present invention, an intra prediction unit 410, a motion estimation unit 420, a motion compensation unit 425, The inverse quantization unit 460, the inverse quantization unit 470, the deblocking unit 480 and the loop filtering unit 490 of the quantization unit 430, the quantization unit 440, the entropy coding unit 450, the inverse quantization unit 460, , Sub-encoding units according to depth, prediction units, and conversion units. The intra prediction unit 410, the motion estimation unit 420, and the motion compensation unit 425 determine a prediction unit and a prediction mode in an encoding unit in consideration of the maximum size and depth of an encoding unit, and the frequency conversion unit 430 The size of the conversion unit can be considered in consideration of the maximum size and depth of the encoding unit.

5 is a block diagram of an image decoding unit based on an encoding unit according to an embodiment of the present invention.

Referring to FIG. 5, the bitstream 505 is parsed by the parser 510, and the encoded image data to be decoded and the encoding information necessary for decoding are parsed. The encoded image data is output as inverse-quantized data through the entropy decoding unit 520 and the inverse quantization unit 530, and is restored to the residual values through the frequency inverse transform unit 540. [ The residual values are added to the intraprediction result of the intraprediction unit 550 or the motion compensation result of the motion compensation unit 560, and restored for each encoding unit. The reconstructed coding unit is used for predicting the next coding unit or the next picture through the deblocking unit 570 and the loop filtering unit 580.

A parsing unit 510, an entropy decoding unit 520, an inverse quantization unit 530, an inverse frequency transforming unit (hereinafter referred to as an inverse quantization unit) 530, The intra prediction unit 550, the motion compensation unit 560, the deblocking unit 570 and the loop filtering unit 580 are all based on the maximum encoding unit, the depth-dependent sub-encoding unit, the prediction unit, And can process the image decoding processes. In particular, the intra prediction unit 550 and the motion compensation unit 560 determine a prediction unit and a prediction mode in an encoding unit in consideration of the maximum size and depth of an encoding unit, and the frequency inverse transform unit 540 transforms the maximum size And the depth of the conversion unit can be taken into consideration.

FIG. 6 illustrates a depth-based coding unit and a prediction unit according to an embodiment of the present invention.

The image encoding apparatus 100 and the image decoding apparatus 200 according to an embodiment use a hierarchical encoding unit to consider image characteristics. The maximum height, width, and maximum depth of the encoding unit may be adaptively determined according to the characteristics of the image, or may be variously set according to the demand of the user. The size of each coding unit may be determined according to a maximum size of a predetermined coding unit.

The hierarchical structure 600 of the encoding unit according to an embodiment of the present invention shows a case where the maximum height and width of the encoding unit is 64 and the maximum depth is 4. Since the depth is deeper along the vertical axis of the hierarchical structure 600 of the encoding unit according to the embodiment, the height and width of each depth-dependent encoding unit are divided. In addition, along the horizontal axis of the hierarchical structure 600 of the coding unit, a prediction unit which is a partial data unit which is a prediction base of each coding unit of depth is shown.

The maximum coding unit 610 is the largest coding unit among the hierarchical structures 600 of the coding units and has a depth of 0, and the size of the coding units, i.e., the height and the width, is 64x64. A depth-1 encoding unit 620 having a size of 32x32, a depth-2 encoding unit 620 having a size 16x16, a depth-3 encoding unit 640 having a size 8x8, a depth 4x4 having a size 4x4, There is an encoding unit 650. An encoding unit 650 of depth 4 of size 4x4 is the minimum encoding unit.

Referring also to FIG. 6, partial data units are shown along the horizontal axis for each depth, as a prediction unit of an encoding unit. That is, the prediction unit of the maximum encoding unit 610 having a depth 0 size of 64x64 is a partial data unit 610 of size 64x64, partial data units 612 of size 64x32 included in the encoding unit 610 of size 64x64, Partial data units 614 of size 32x64, and partial data units 616 of size 32x32.

The prediction unit of the 32x32 coding unit 620 having the depth 1 is the partial data unit 620 of the size 32x32 included in the coding unit 620 of the size 32x32, the partial data units 622 of the size 32x16, Partial data units 624 of size 16x16, and partial data units 626 of size 16x16.

The prediction unit of the 16x16 coding unit 630 of depth 2 is a partial data unit 630 of size 16x16, partial data units 632 of size 16x8, a size of 8x16 16x16 included in the coding unit 630 of size 16x16, Partial data units 634 of size 8x8, and partial data units 636 of size 8x8.

The prediction unit of the encoding unit 640 having the depth 3 of 8x8 is a partial data unit 640 of the size 8x8, partial data units 642 of the size 8x4, a size 4x8 of the size 8x8 included in the encoding unit 640 of the size 8x8, Partial data units 644 of size 4x4, and partial data units 646 of size 4x4.

Finally, a coding unit 650 of size 4x4 is the minimum coding unit and the coding unit of the lowest depth, and the prediction unit is a data unit 650 of size 4x4.

In order to determine the depth of encoding of the maximum encoding unit 610, the encoding depth determining unit 120 of the image encoding apparatus according to an exemplary embodiment may perform encoding for each depth unit included in the maximum encoding unit 610 Should be performed.

The number of coding units per depth to include data of the same range and size increases as the depth of the coding unit increases. For example, for data containing one coding unit at depth 1, four coding units at depth 2 are required. Therefore, in order to compare the encoding results of the same data by depth, they should be encoded using a single depth 1 encoding unit and four depth 2 encoding units, respectively.

For each depth-of-field coding, encoding is performed for each prediction unit of the depth-dependent coding unit along the horizontal axis of the hierarchical structure 600 of the coding unit, and a representative coding error, which is the smallest coding error at the corresponding depth, is selected . In addition, depths are deepened along the vertical axis of the hierarchical structure 600 of encoding units, and the minimum encoding errors can be retrieved by comparing the representative encoding errors per depth by performing encoding for each depth. The depth at which the minimum coding error occurs among the maximum coding units 610 can be selected as the coding depth and the partition type of the maximum coding unit 610. [

FIG. 7 shows a relationship between an encoding unit and a conversion unit according to an embodiment of the present invention.

The image encoding apparatus 100 and the image decoding apparatus 200 according to an embodiment of the present invention divide and encode or decode an image with a coding unit smaller than or equal to the maximum encoding unit for each maximum encoding unit. The size of the conversion unit for frequency conversion during encoding can be selected based on data units that are not larger than the respective encoding units. For example, when the current encoding unit 710 is 64x64, the frequency conversion can be performed using the 32x32 conversion unit 720. [ In addition, the data of the encoding unit 710 of 64x64 size is encoded by performing the frequency conversion with the conversion units of 32x32, 16x16, 8x8, and 4x4 size of 64x64 or smaller, respectively, and then the conversion unit having the smallest error with the original Can be selected. In addition, the image encoding apparatus 100 and the image decoding apparatus 200 according to an exemplary embodiment of the present invention may group a plurality of prediction units, set one conversion unit, and perform frequency conversion. For example, when the encoding unit has a size of 32x32, the frequency conversion can be performed by using a 64x64 data unit grouped by the adjacent four encoding units of 32x32 size as one conversion unit.

FIG. 8 illustrates depth-specific encoding information, in accordance with an embodiment of the present invention.

The encoding information encoding unit of the image encoding apparatus 100 according to an embodiment of the present invention includes information on the encoding mode, information 800 about the partition type, information on the prediction mode 810 ), And information 820 on the conversion unit size may be encoded and transmitted.

The partition type information 800 indicates a prediction unit for predictive encoding of the current encoding unit and information on the type in which the current encoding unit is divided. For example, the current encoding unit CU_0 of depth 0 and size 2Nx2N includes a prediction unit 802 of size 2Nx2N, a prediction unit 804 of size 2NxN, a prediction unit 806 of size Nx2N, a prediction unit 808 of size NxN ) And can be used as a prediction unit. In this case, information 800 regarding the partition type of the current encoding unit includes a prediction unit 802 of size 2Nx2N, a prediction unit 804 of size 2NxN, a prediction unit 806 of size Nx2N, and a prediction unit 808 of size NxN. Lt; / RTI >

The information 810 on the prediction mode indicates the prediction mode of each prediction unit. The prediction unit indicated by the information 800 relating to the partition type is predicted to be one of the intra mode 812, the inter mode 814 and the skip mode 816 through the prediction mode information 810, for example. Can be set.

In addition, the information 820 on the conversion unit size indicates whether to perform frequency conversion on the basis of which conversion unit the current encoding unit is performed. For example, the conversion unit may be one of a first intra-conversion unit size 822, a second intra-conversion unit size 824, a first inter-conversion unit size 826, have.

The encoding information extracting unit of the video decoding apparatus 200 according to an embodiment of the present invention extracts the information 800 about the partition type, the information 810 about the prediction mode, Information 820 can be extracted and used for decoding.

FIG. 9 shows a depth encoding unit according to an embodiment of the present invention.

Partition information may be used to indicate changes in depth. The division information indicates whether the current-depth encoding unit is divided into lower-depth encoding units.

The prediction unit 910 for the prediction encoding of the coding units of depth 0 and 2N_0x2N_0 has a partition type 912 of 2N_0x2N_0 size, a partition type 914 of 2N_0xN_0 size, a partition type 916 of N_0x2N_0 size, a partition of size N_0xN_0 Type < / RTI >

For each partition type, predictive encoding should be repeatedly performed for each prediction unit of 2N_0x2N_0 size, two 2N_0xN_0 size prediction units, two N_0x2N_0 size prediction units, and four N_0xN_0 size prediction units. For a prediction unit of size 2N_0x2N_0, size N_0x2N_0, size 2N_0xN_0, and size N_0xN_0, motion prediction can be performed in intra mode and inter mode. The skip mode can be performed only for the prediction unit of size 2N_0x2N_0.

If the coding error by the partition type 918 of the size N_0xN_0 is the smallest, the depth 0 is changed to 1 (920), and the coding units 922, 924, 926 and 928 of the partition type of the depth 2 and the size N_0xN_0 are changed The minimum coding error can be repeatedly searched for.

Since encoding is repeatedly performed on the encoding units 922, 924, 926, and 928 of the same depth, encoding of only one of the encoding units of depth 1, for example, will be described. The prediction unit 930 for predicting the coding unit of the depth 1 and the size 2N_1x2N_1 (= N_0xN_0) includes a partition type 932 of size 2N_1x2N_1, a partition type 934 of size 2N_1xN_1, a partition type 936 of size N_1x2N_1, , And a partition type 938 of size N_1xN_1. For each partition type, predictive coding should be repeatedly performed for each prediction unit of a size 2N_1x2N_1, a prediction unit of two sizes 2N_1xN_1, a prediction unit of two sizes N_1x2N_1, and a prediction unit of four sizes N_1xN_1.

If the coding error by the partition type 938 of the size N_1xN_1 is the smallest, the coding units 942, 944, 946 and 948 of the depth 2 and the size N_2xN_2 are changed while changing the depth 1 to the depth 2 (940) The minimum coding error can be repeatedly searched for.

If the maximum depth is d, the depth-based segmentation information can be set until the depth d-1. That is, the prediction unit 950 for predictive coding of the coding unit of the depth d-1 and the size 2N_ (d-1) x2N_ (d-1) A partition type 954 of size 2N_ (d-1) xN_ (d-1), a partition type 956 of size N_ (d-1) x2N_ 1) xN_ (d-1) < / RTI >

(D-1) x2N_ (d-1), two predicted units of two sizes 2N_ (d-1) (d-1) x2N_ (d-1), four sizes N_ (d-1) xN_ (d-1). Since the maximum depth is d, the coding unit 952 of depth d-1 no longer undergoes the division process.

The image encoding apparatus 100 according to an exemplary embodiment of the present invention compares depth-based encoding errors to determine depths for encoding units 912 and selects the depths at which the smallest encoding error occurs. For example, a coding error for a coding unit of depth 0 is predicted for each partition type 912, 914, 916, and 918, and a prediction unit in which the smallest coding error occurs is determined. Similarly, a prediction unit having the smallest coding error can be searched for every depth 0, 1, ..., d-1. At the depth d, the coding error can be determined through predictive coding based on the prediction unit 960, which is a coding unit of size 2N_dx2N_d. In this way, the minimum coding errors of the depths 0, 1, ..., d-1, and d are compared with each other, and the depth with the smallest error is selected to be determined as the coding depth. The coding depth and the prediction unit of the corresponding depth can be encoded and transmitted as information on the coding mode. In addition, since the coding unit must be divided from the depth 0 to the coding depth, only the division information of the coding depth is set to '0', and the division information by depth is set to '1' except for the coding depth.

The encoding information extracting unit 220 of the image decoding apparatus 200 according to an embodiment of the present invention extracts information on the encoding depth and the prediction unit for the encoding unit 912 and uses the information on the encoding depth 912 to decode the encoding unit 912 . The image decoding apparatus 200 according to an exemplary embodiment of the present invention uses the division information according to the depth to determine the depth with the division information of '0' as a coding depth and can use it for decoding using the information about the coding mode for the corresponding depth have.

FIGS. 10A, 10B, and 10C illustrate the relationship between an encoding unit, a prediction unit, and a frequency conversion unit according to an embodiment of the present invention.

The encoding unit 1010 is encoding units for encoding depth determined by the image encoding apparatus 100 according to the embodiment with respect to the maximum encoding unit. The prediction unit 1060 is a prediction unit of each coding depth unit among the coding units 1010 and the conversion unit 1070 is a conversion unit of each coding depth unit.

When the depth of the maximum encoding unit is 0, the depth of the encoding units 1012 and 1054 is 1 and the depth of the encoding units 1014, 1016, 1018, 1028, 1050, The coding units 1020, 1022, 1024, 1026, 1030, 1032 and 1048 have a depth of 3 and the coding units 1040, 1042, 1044 and 1046 have a depth of 4.

A portion (1014, 1016, 1022, 1032, 1048, 1050, 1052, 1054) of the prediction units 1060 is a type in which the coding unit is divided. That is, the prediction units 1014, 1022, 1050 and 1054 are partition types of 2NxN, the prediction units 1016, 1048 and 1052 are partition types of Nx2N, and the prediction units 1032 are partition types of NxN. That is, the prediction unit of the depth-dependent coding units 1010 is smaller than or equal to each coding unit.

The image data of a part 1052 of the conversion units 1070 is subjected to frequency conversion or frequency inverse conversion in units of data smaller in size than the encoding unit. The conversion units 1014, 1016, 1022, 1032, 1048, 1050, 1052, and 1054 are data units of different sizes or types when compared with the prediction units of the prediction units 1060. That is, the image encoding apparatus 100 and the image decoding apparatus 200 according to an exemplary embodiment of the present invention perform prediction and frequency conversion / inverse transform operations on the same encoding unit, respectively, based on separate data units .

FIG. 11 shows encoding information for each encoding unit according to an embodiment of the present invention.

The output unit 130 of the image encoding apparatus 100 according to an exemplary embodiment of the present invention outputs encoding information for each encoding unit and outputs the encoding information to the encoding information extracting unit 200 of the image decoding apparatus 200 according to an embodiment of the present invention 220) can extract the encoding information for each encoding unit.

The encoding information may include division information for the encoding unit, partition type information, prediction mode information, and conversion unit size information. The encoding information shown in FIG. 11 is only an example that can be set in the image encoding apparatus 100 and the image decoding apparatus 200 according to an embodiment of the present invention, and is not limited to the illustrated ones.

The division information may indicate the coding depth of the corresponding encoding unit. That is, since depths that are no longer divided according to the division information are coding depths, partition type information, prediction mode, and conversion unit size information can be defined with respect to the coding depth. If it is necessary to further divide by one division according to the division information, encoding should be performed independently for each of the four higher-depth-depth coding units.

As the partition type information, the partition type of the conversion unit of the coding unit of the coding depth can be represented by 2Nx2N, 2NxN, Nx2N and NxN. The prediction mode may indicate the motion prediction mode as one of an intra mode, an inter mode, and a skip mode. The intra mode and the inter mode can be defined in the partition types 2Nx2N, 2NxN, Nx2N and NxN, and the skip mode can be defined only in the partition type 2Nx2N. The conversion unit size can be set to two kinds of sizes in the intra mode and two kinds of sizes in the inter mode.

The encoding unit-specific encoding information of the belonging encoding depth may be stored for each minimum encoding unit in the encoding unit. Therefore, if encoding information held in each of the adjacent minimum encoding units is checked, it can be confirmed whether or not the encoding information is included in the encoding unit of the same encoding depth. In addition, since the encoding unit of the encoding depth can be identified by using the encoding information held in the minimum encoding unit, the distribution of encoding depths in the maximum encoding unit can be inferred.

In this case, when the current encoding unit is predicted with reference to the neighboring data unit, the encoding information of the minimum encoding unit in the depth encoding unit adjacent to the current encoding unit is directly used, so that the data of the minimum encoding unit can be referred to.

In yet another embodiment, the encoding information of the depth encoding unit may be stored only for the minimum encoding unit represented in the depth encoding unit. In this case, when the current encoding unit is predicted by referring to the surrounding encoding unit, the data adjacent to the current encoding unit in the depth encoding unit may be retrieved using the encoding information of the adjacent depth encoding unit.

The entropy encoding unit 450 of the image encoding apparatus 400 and the entropy decoding unit 520 of the image decoding apparatus 500 of FIG. The encoding and decoding processes will be described in detail. In the following description, a coding unit is a term referring to a block currently coded in an image coding step, and a decoding unit is a term referring to a block to be currently decoded in an image decoding step. The terms encoding unit and decode unit are merely the difference between the encoding step and the decoding step of the image and the encoding unit in the encoding step may be referred to as a decode unit in the decode step. For the sake of uniformity of terms, except for special cases, they shall be uniformly coded in the coding step and the decoding step in the same coding unit. In addition, the method and apparatus for encoding and decoding residual blocks according to an embodiment of the present invention can be applied to encoding and decoding of a residual block in a general image codec. If so, it can be understood through this specification.

As described above, the difference between the prediction block generated in the intraprediction unit 410 or the motion compensation unit 425 of the image coding apparatus 400 according to an embodiment of the present invention and the current block to be currently coded The frequency converter 430 converts the residual block into the frequency domain. For example, the frequency transform unit 430 transforms a residual block into a frequency domain using DCT (Discrete Cosine Transform) to generate a transform residual block. In particular, as described above with reference to FIG. 7, the frequency transforming unit 430 according to an exemplary embodiment of the present invention performs frequency transformation using data units not larger than the current encoding unit as one transform unit, The frequency conversion can be performed by using the data unit grouped by the prediction units as one conversion unit. For example, when the size of the residual block is 64x64, the frequency conversion unit 430 generates a conversion residual block according to the size of the selected one of the 64x64, 32x32, 16x16, 8x8, and 4x4 sizes, A conversion residual block can be generated in units of 128x128 by using a 128x128 data unit grouped by vertically and horizontally adjacent adjacent residual blocks as one conversion unit. The transform residual block is quantized by the quantization unit 440 and then encoded by the entropy encoding unit 450. The quantization unit 440 may be omitted in the case of lossless coding.

12A to 12C are views for explaining a process of encoding a transform residual block in a technical field related to the present invention.

Referring to FIG. 12A, when a frequency conversion is performed on a residual block to generate a transform residual block 1210, the transform coefficients in the transform residual block 1210 are scanned according to a zigzag scan order, The significance map indicating the position of the effective transform coefficient having a non-zero value and the level information of the effective transform coefficient are encoded. For example, when the size of the transform residual block is 4x4, a process of encoding the transform residual block 1220 as shown in FIG. 12B will be described. It is assumed that the transform coefficients at positions indicated by X in Fig. 12B are effective transform coefficients having a non-zero value. In this case, as shown in FIG. 12C, the validity map is one in which the effective conversion coefficient among the conversion coefficients in the residual block 1230 is represented by 1, and the conversion coefficient of 0 is represented by 0 as it is. The validity map is context-based binary arithmetic coding while being scanned according to a predetermined scan order. For example, when a validity map as shown in FIG. 12C is coded according to a raster scanning sequence for scanning from left to right and from top to bottom, a validity map corresponding to an empty string of "111111110101000" Context-based encoding. After the validity map is coded, the level information of the effective coefficient, that is, the sign and the absolute value abs of the effective coefficient are coded.

A method of encoding a transform residual block in the related art is suitable for encoding a relatively small-sized transformed residual block having a size of 8x8 or 4x4. However, as in the image encoding apparatus according to an embodiment of the present invention, , It is not suitable for encoding a conversion residual block having a large size such as 64x64. In the case where all transform coefficients in the transform residual block are scanned and coded according to the related art as described with reference to Figs. 12A to 12C for a transformed residual block having a large size, an empty string corresponding to the validity map And the coding efficiency may be lowered.

Therefore, a residual block coding method and apparatus according to an embodiment of the present invention divides a transform residual block into predetermined frequency band units, and determines whether or not a non-zero effective transform coefficient exists for each divided frequency band unit And the effective coefficient information, that is, the level information of the validity map and the effective coefficient, is encoded only for the frequency band having the value of the effective coefficient flag for each frequency band unit of 1, Thereby enabling encoding of the residual block.

13 is a block diagram showing the configuration of a residual block coding apparatus 1300 according to an embodiment of the present invention. The residual block coding apparatus 1300 of FIG. 13 may correspond to or be included in the entropy coding unit 450 of FIG.

13, the residual block coding apparatus 1300 includes a frequency band division unit 1310, an effective coefficient flag generation unit 1320, and an effective coefficient encoding unit 1330.

The frequency band division unit 1310 divides the conversion residual block into predetermined frequency band units. Referring back to FIG. 12A, the transform residual block 1210 corresponds to a transform coefficient of the low frequency component as the transform coefficient located on the upper left side, and corresponds to the transform coefficient of the high frequency component as it goes to the lower right side. In general, the transform coefficient of the transform residual block 1210 is located on the low-frequency component side, and the transform coefficient of the high-frequency component is 0 in many cases. That is, effective conversion coefficients having non-zero values among the conversion coefficients of the high frequency components are sparse. In particular, as in the image encoding apparatus 400 according to the embodiment of the present invention, frequency conversion is performed by converting units of 16x16, 32x32, 64x64, and larger than the conventional 4x4 and 8x8 transform units, In the case of the transformed residual, the distribution of the effective transform coefficients of the high-frequency components becomes more slender. Accordingly, the frequency dividing unit 1310 divides the transform residual block into predetermined frequency bands, considering the distribution characteristics according to the frequency bands of the transform coefficients existing in the transform residual block.

14A to 14J are diagrams illustrating embodiments in which a transform residual block is divided into predetermined frequency band units according to the present invention.

Referring to FIG. 14A, the frequency band division unit 1310 divides a transformed residual block from a low frequency band to a horizontal frequency H1 and a vertical frequency V1 at predetermined frequency intervals to generate frequency band units 1411 to 1414. In FIG. 14A, the lengths of the transverse and vertical sides of the frequency band units 1411 to 1414 are the same. However, the lengths of the transverse and longitudinal sides may be set to be different from each other. If the remaining frequency band from the horizontal frequency H1 to the maximum horizontal frequency is shorter than the frequency interval corresponding to the length of the side of the frequency band units 1411 to 1414 or the remaining frequency from the vertical frequency V1 to the maximum vertical frequency When the length of the band is smaller than the frequency interval corresponding to the length of the longitudinal side of the frequency band units 1411 to 1414, the frequency band division unit 1310 no longer splits the conversion residual block and corresponds to the high frequency band component And generates one frequency band unit 1415 for performing frequency division multiplexing. In general, the effective conversion coefficients are concentrated in the frequency band units 1411 to 1414 corresponding to the low frequency components, and the distribution of the effective conversion coefficients of the high frequency components is thin. Therefore, even if all of the remaining high frequency band components other than the frequency band units 1411 to 1414 generated by dividing the conversion residual block at predetermined frequency intervals are generated in one frequency band unit 1415, The overhead in encoding the transform coefficients does not increase significantly.

14B, the frequency band division unit 1310 divides the conversion residual block 1420 from the low frequency band to the horizontal frequency H2 and the vertical frequency V2, similarly to the above-described FIG. 14A, And generate the frequency band units 1425, 1426, and 1427 by dividing the transformed residual block of the remaining high frequency band components based on the horizontal frequency H2 and the vertical frequency V2.

14C, the frequency band division unit 1310 divides the conversion residual block 1430 from the low frequency band to the horizontal frequency H3 and the vertical frequency V3 into a frequency band unit 1435 and 1436 of the high frequency components by dividing the transformed residual block of the remaining high frequency band components by the vertical frequency V3.

14D, the frequency band division unit 1310 divides the conversion residual block 1440 from the low frequency band to the horizontal frequency H4 and the vertical frequency V4 in a manner similar to the above-described FIG. 14A, Frequency components 1445 to 1444 of the high frequency components by dividing the transformed residual block of the remaining high frequency band components by the horizontal frequency H4.

As described above, the distribution of the effective conversion coefficients is concentrated in the low frequency band, and becomes thinner toward the high frequency band. Therefore, considering the distribution characteristics of the effective transform coefficients, the frequency band divider 1310 divides the low-frequency band into a plurality of low-frequency bands such that a divided unit size of the low-frequency band is smaller than a divided unit size of the high- (1450). In other words, the frequency band division unit 1310 divides the low frequency band more finely and divides the high frequency band relatively more so as to more accurately encode the effective conversion coefficient distributed intensively in the low frequency band. For example, as shown in FIG. 14E, the frequency band division unit 1310 divides a horizontal frequency H5 and a vertical frequency V5, a horizontal frequency H6 having a value larger than a multiple of the horizontal frequency H5, It is possible to generate the frequency band division units 1451, 1452, 1453, 1454, 1455, 1456, 1457 by dividing the conversion residual block 1450 based on the vertical frequency V6 having a larger value. Assuming that the sizes of the frequency band division units 1451, 1452, 1453, 1454, 1455, 1454, 1454, 1454, 1455, 1456, 1457 are A1451, A1452, A1453, A1454, A1455, A1456, A1457, Is divided to have the maximum size.

Also, as shown in FIG. 14F, the frequency band division unit 1310 may divide the conversion residual block 1460 into frequency band units 1461 having the same size.

Also, as shown in FIG. 14G, the frequency band division unit 1310 divides the conversion residual block 1470 into quadrants, divides the smallest frequency band unit 1471 of the quadrature frequency band units into quadrants Thereby generating frequency band units. The frequency band dividing unit 1310 can divide the lowest frequency band unit 1472 among the frequency band units divided into quadruples into the low frequency band unit 1471. This division process can be repeated until the size of the quadrature frequency band unit becomes a predetermined size.

14H, the frequency band division unit 1310 generates a frequency band unit 1481 of low frequency components from the low frequency to the horizontal frequency H7 and the vertical frequency V7, and outputs the frequency band unit 1481 of the conversion residual block 1480 The remaining high frequency components may be diagonally separated to generate frequency band units 1482 and 1483.

Also, as shown in FIGS. 14I and 14J, the frequency band division unit 1310 may divide the conversion residual blocks 1490 and 1495 by connecting a horizontal frequency and a vertical frequency having predetermined values. In FIG. 14I, the horizontal and vertical frequencies having the same frequency interval are connected to each other in the conversion residual block 1490. In the case of FIG. 14J, the frequency interval is increased toward the high frequency, that is, a1 < a1 and b1, a2 and b2, a3 and b3, and a4 and b4 satisfying the relationship of a2 <a3 <a4 and b1 <b2 <b3 <b4 are connected to divide the conversion residual block 1495.

14A to 14J, the frequency band division unit 1310 may use the distribution characteristic of the effective transform coefficients constituting the transform residual block or the effective transform coefficients existing in each frequency band, It is possible to determine the image characteristic of the transformed residual block by using the number of coefficients and to determine the size of the frequency unit for dividing the transformed residual block for each frequency band using the determined image characteristic of the transformed residual block. Assuming, for example, that the effective conversion coefficient in the conversion residual block exists only in a frequency band lower than the horizontal frequency H8 and the vertical frequency V8, and that there is no effective conversion coefficient in the frequency band larger than the horizontal frequency H8 and the vertical frequency V8 , The frequency band division unit 1310 divides the entire conversion residual block from the low frequency band to the horizontal frequency H8 and the vertical frequency V8 into one frequency band unit or into the same frequency band units, And a frequency band larger than the vertical frequency V8 can be set as a unit of one frequency band.

The present invention is not limited to the embodiments in which the conversion residual block shown in FIGs. 14A to 14J is divided into frequency bands. In the present invention, It will be understood that the present invention is not limited thereto.

On the other hand, in the frequency band dividing unit 1310, the transforming residual block is divided into the same shape on the encoding side and on the decoding side, or a predetermined index is assigned to each of the various division types as shown in FIGS. 14A to 14J The division index for the division information used at the time of coding the conversion residual block on the encoding side can be added to the encoded bit stream. For example, assuming that the integer values of the division indices (div_index) 0 to 9 represent the respective division types shown in Figs. 14A to 14J, and the division type used at the time of coding the current conversion residual block is 14F , The division index information can be added to the coding information of the current transform residual block if div_index = 5 as shown in FIG.

Referring again to FIG. 13, after dividing the transformed residual block by a predetermined frequency band unit by the frequency band dividing unit 1310, the effective coefficient flag generating unit 1320 determines whether there is an effective transform coefficient for each frequency band unit And generates a valid coefficient flag indicating the validity coefficient flag. At this time, it is preferable that the effective coefficient flag generation unit 1320 does not generate the effective coefficient flag separately for the smallest low frequency band unit. For example, when the conversion resume 1410 is divided as shown in FIG. 14A, the effective coefficient flag generating unit 1320 generates the effective coefficient flag generating unit 1320 for the remaining frequency band units 1412 and 1413 except for the smallest low frequency band unit 1411 , 1414, and 1415 are generated. Coeff_exist_1413, Coeff_exist_1414 and Coeff_exist_1415 for each of the frequency band units 1412, 1413, 1414 and 1415, and the effective coefficient flags for the frequency band units 1412, 1413, 1414 and 1415 are denoted as Coeff_exist_1412, The effective coefficient flag generation unit 1320 generates effective coefficient flags for each frequency band unit such as Coeff_exist_1412 = 1, Coeff_exist_1413 = 1, Coeff_exist_1414 = 0, Coeff_exist_1415 = 0. As described above, in the case of the smallest low frequency band unit 1411, it is very likely that there is a possibility that the effective transform coefficient is generally present. Therefore, for the smallest low frequency band unit 1411, It is preferable not to generate a flag. Also, instead of separately generating an effective coefficient flag indicating the presence of an effective transform coefficient in the smallest low frequency band unit 1411, a coded_block_flag indicating whether or not an effective transform coefficient exists in the conventional residual block is used, The presence or absence of the effective conversion coefficient may be indicated in the area 1411. The above-described process of generating the coefficient flag is not limited to the division form shown in Fig. 14A, and can be applied to other division forms shown in Figs. 14B to 14J.

Referring again to FIG. 13, the effective coefficient encoding unit 1330 encodes the effective coefficient indicating that the effective coefficient flag generated by the effective coefficient flag generating unit 1320 has a value of 1, that is, The validity map indicating the position of the effective conversion coefficient, and the level information of the effective conversion coefficient.

15A and 15B are reference diagrams for explaining the coding process of the effective transform coefficients according to an embodiment of the present invention. 15A and 15B illustrate a division mode corresponding to the above-described FIG. 14E in which the conversion residual block is divided into quadrants and the low frequency band is divided into quadrants again to generate frequency band units. However, the present invention is not limited to this, The encoding process of the effective transform coefficients described below can be similarly applied to the frequency bands of other division types shown in Figs.

The effective coefficient encoding unit 1330 may encode the effective transform coefficients by scanning the entire transformed residual block or may perform independent scanning for each frequency band unit to encode the effective transform coefficients on a frequency band basis. 15A, the effective coefficient encoding unit 1330 scans the entire conversion residual block 1510 according to a predetermined scanning order, for example, a raster scanning order, as shown in the figure, The validity map indicating the position of the effective transform coefficients present in the transform unit 1510, and the size and sign information of each effective transform coefficient. In this case, the scanning process in the frequency band unit having the value of the effective coefficient flag of 0, that is, the frequency band unit in which the effective conversion coefficient does not exist, can be skipped.

15B, the effective coefficient coding unit 1330 outputs the validity map and the effective transform coefficient for each frequency band unit in accordance with the division form of the conversion residual block 1520 divided by the frequency band division unit 1310 Level information can be encoded.

16A and 16B are views for explaining the encoding process of the residual block according to an embodiment of the present invention. In Figs. 16A and 16B, it is assumed that the conversion coefficient denoted by x is an effective conversion coefficient, and the portion where there is no indication is a conversion coefficient having a value of 0.

Referring to FIG. 16A, the frequency band division unit 1310 divides the conversion residual block 1610 according to one division type of the division type as shown in FIGS. 14A to 14J. In FIG. 16A, the division type corresponding to FIG. 14E is exemplified. However, the present invention is not limited thereto, and the encoding process of the conversion residual block, which will be described below, may be applied to various division forms. The effective coefficient flag generation unit 1320 sets the effective coefficient flags of the frequency band units 1611, 1612, and 1613 in which the effective conversion coefficients exist, and sets the effective coefficient flags in the frequency band units 1614, 1615, 1616, and 1617 are set to 0, respectively. The effective coefficient encoding unit 1330 encodes the validity map indicating the position of the effective transform coefficient and the level information of each effective transform coefficient while scanning the entire transformed residual block 1610. As described above, the validity map indicates whether the conversion coefficient according to each scan index is an effective conversion coefficient or 0. The level information of the effective transform coefficient includes sign and absolute value information of the effective transform coefficient. For example, the validity map of the frequency band units 1611, 1612, 1613 in which the effective conversion coefficient exists has an empty string value such as "1000100010101110100100100010001" when scanned according to the raster scanning order as shown in Fig. 16A .

16A, when the information of the effective transform coefficient is encoded while scanning the entire transformed residual block 1610, an end-of-block (EOB) indicating whether the effective transform coefficient is the last effective transform coefficient, The flag may be set for the entire conversion residual block 1610, or may be set for each frequency band unit. For example, when the EOB flag is set for the entire conversion residue 1610, only the conversion coefficient corresponding to 1602, which is the last effective conversion coefficient in the scanning order among the conversion coefficients shown in FIG. 16A, Lt; / RTI &gt; If the EOB flag is set for each frequency band unit, a conversion coefficient corresponding to 1601 in the frequency band unit 1611, a conversion coefficient corresponding to 1602 in the frequency band unit 1612, The EOB flags of the transform coefficients 1603 in the field 1613 are set to 1. The EOB flag need not be set for the frequency band units 1614, 1615, 1616, and 1617 in which no effective conversion coefficient exists. When the EOB flag is set for each frequency band unit in which the effective conversion coefficient exists, the scanning of the effective conversion coefficient in the predetermined frequency band unit is completed, and then the scanning of the effective conversion coefficient in the next frequency band unit can be started have. For example, scanning of the transform coefficients existing in the frequency band unit 1602 may be started after the scanning of the last effective transform coefficient 1603 of the frequency band unit 1613 is completed.

16B shows a case in which the effective transform coefficient information is independently encoded for each frequency band unit. The effective coefficient encoder 1330 independently scans each frequency band unit of the transform residual block 1620 and encodes the validity map indicating the position of the effective transform coefficient and the level information of each effective transform coefficient. For example, the validity map of the frequency band unit 1621 has an empty string value such as "1000100010011" when scanned according to the raster scanning order as shown in Fig. 16B. The effective coefficient encoding unit 1330 sets the EOB flag of the effective transform coefficient 1631 corresponding to the last effective transform coefficient among the effective transform coefficients in the frequency band unit 1621 to 1. Similarly, the effective coefficient encoding unit 1330 generates an empty string value such as "101010001" as a validity map of the frequency band unit 1622. Also, the effective coefficient encoding unit 1330 sets the EOB flag of the effective conversion coefficient of the reference numeral 1632 among the effective conversion coefficients in the frequency band unit 1622 to 1. Similarly, the effective coefficient encoding unit 1330 generates an empty string value such as "11001 " as the validity map of the frequency band unit 1623, and sets the EOB flag of the effective conversion coefficient 1633 to 1.

The effective coefficient coding unit 1330 may include an EOB flag indicating that the last effective transform coefficients 1631, 1632, and 1633 of the respective frequency band units are the last effective transform coefficients in the corresponding frequency band unit, (End_Of_WholeBlock) indicating whether the last valid coefficient is the last valid coefficient among the plurality of valid coefficients. 16B, when the frequency band units are independently scanned in the order of 1621, 1622, 1623, 1624, 1625, 1626, and 1627, the effective conversion coefficient 1633 is the last validity of the frequency band unit 1623 Conversion coefficient and the last effective conversion coefficient of the whole conversion residual block 1620. [ Therefore, both the EOB flag and the End_of_wholeBlock flag of the effective conversion coefficient 1633 have a value of one. The last valid conversion coefficients 1631 and 1632 of the frequency band units 1621 and 1622 have a value of 1 for the EOB flag and a value of 0 for the End_of_WholeBlock flag.

When the EOB flag and the End_Of_WholeBlock flag are set for the last effective transform coefficient for each frequency band unit as described above, it is first determined at the time of decoding whether the effective coefficient is present in the frequency band unit by using the effective coefficient flag described above Scanning in frequency band units in which the effective coefficient flag is 0 can be skipped. When the conversion coefficient in the frequency band unit in which the effective coefficient flag is 1, that is, the conversion coefficient in the frequency band unit in which the effective conversion coefficient exists is scanned up to the conversion coefficient having the EOB flag of 1, scanning in the next frequency band unit is started . If the EOB flag is 1 and the effective conversion coefficient of the End_Of_WholeBlock flag is 1, the scanning of the effective conversion coefficient of the entire conversion residual block is completed and the scanning process of the conversion residual block is terminated.

17A and 17B are reference diagrams showing embodiments of encoding information of the conversion residual block generated by the effective coefficient encoding unit 1330. FIG.

Referring to FIG. 17A, the effective coefficient coding unit 1330 can sequentially code the validity map and the effective coefficient flag information generated for each frequency band. When the first frequency band is the smallest frequency band of the conversion residual block, only the information of the validity map 1711 of the first frequency band is encoded for the first frequency band, Encoding of the first frequency band flag indicating whether or not the effective conversion coefficient exists may not be performed separately. 17B, the valid coefficient flags 1721 of each frequency band may be encoded first, and the validity maps 1725 of each frequency band may be encoded later.

18 is a flowchart illustrating a residual block coding method according to an embodiment of the present invention.

Referring to FIG. 18, in step 1810, a prediction block is generated through inter prediction or intra prediction by the intra prediction unit 410 or the motion compensation unit 425.

In step 1820, a residual block, which is a difference value between the predicted block and the current block, is generated by the not shown subtractor.

In operation 1830, the frequency transform unit 430 transforms the residual block into a frequency domain to generate a transform residual block. For example, the residual block can be transformed into the frequency domain through the DCT transform.

In step 1840, the frequency band division unit 1310 divides the conversion residual block into predetermined frequency band units. As described above, the frequency band division unit 1310 can divide the conversion residual block into various forms as illustrated in Figs. 14A to 14J. Specifically, the frequency band division unit 1310 divides the transformed residual block such that the divided unit size of the low frequency band is smaller than the divided unit size of the high frequency band, or the transformed residual block is divided into quarters The transformed residual block may be divided by repeating the process of dividing the lowest frequency band of the residual block again into four equal parts, or the transformed residual block may be divided into the same frequency band units, or a horizontal frequency and a vertical frequency having the same value may be connected And determines a size to be divided for each frequency band of the transformed residual block by using the image characteristic of the transformed residual block determined using the transform coefficients constituting the transformed residual block, The conversion residual block can be divided according to the star division size.

In step 1850, the effective coefficient flag generation unit 1320 generates a valid coefficient flag for each frequency band unit that indicates whether or not a non-zero effective conversion coefficient exists for each divided frequency band unit. It is preferable that the effective coefficient flag is not separately generated for the smallest frequency band unit of the divided frequency band units of the conversion residual block. As described with reference to FIGS. 15A, 15B, 16A, and 16B, the effective coefficient encoding unit 1330 performs a full conversion on the frequency band units in which the effective coefficient flag is not 0, The residual blocks are scanned according to a predetermined scanning order or the level information of the validity map and the effective transform coefficients indicating the positions of the effective transform coefficients are encoded while scanning is performed independently for each frequency band unit.

According to the method and apparatus for encoding a residual block according to an embodiment of the present invention, a transformed residual block having a large size of 16x16 or more is divided and processed in units of frequency bands to obtain a distribution of effective transform coefficients in the transformed residual block The effective transform coefficient information can be encoded more efficiently according to the characteristics. That is, according to the present invention, a large-sized conversion residual block is divided into frequency band units and an effective coefficient flag indicating the presence or absence of an effective conversion coefficient is generated for each divided frequency band unit, It is possible to skip the scanning process in the frequency band in which no frequency band exists and to reduce the amount of bits required for coding the effective transform coefficient.

19 is a block diagram illustrating an apparatus for decoding residual blocks according to an embodiment of the present invention. The residual block decoding apparatus 1900 according to an embodiment of the present invention may correspond to or be included in the entropy decoding unit 520 of FIG.

19, the residual block decoding apparatus 1900 includes a frequency band division unit 1910, an effective frequency band determination unit 1920, and an effective coefficient decoding unit 1930.

The frequency band division unit 1910 divides the conversion residual block into predetermined frequency band units. Specifically, the frequency band division unit 1910 divides the conversion residual block such that the divided unit size of the low frequency band is smaller than the divided unit size of the high frequency band, as illustrated in Figs. 14A to 14J, The transformed residual block is divided into four equal parts and the transformed residual block is repeated by repeating the process of again dividing the lowest frequency band among the quadrupled transformed residual blocks into quadrants, , The horizontal and vertical frequencies having the same value are connected to each other to divide the conversion residual block or the conversion residual block is determined by using the conversion characteristic of the conversion residual block determined using the conversion coefficients constituting the conversion residual block The size to be divided according to the frequency band is determined, The conversion residual block can be divided. In order to divide the conversion residual block into any of the various division types, a predetermined division index is set in advance according to the type determined by the encoding apparatus and the decoding apparatus, for each of the various division types, and at the time of encoding, The frequency band partitioning unit 1910 may determine from the division index information included in the bitstream the current conversion residual block is divided into a certain type.

The valid frequency band determination unit 1920 extracts a valid coefficient flag indicating whether an effective transform coefficient exists for each frequency band unit obtained by dividing the transformed residual block from the bit stream. Using the effective coefficient flag, the effective frequency band determination unit 1920 can determine a frequency band unit in which effective conversion coefficients exist among frequency band units obtained by dividing the conversion residual block. For example, if it is assumed that the coded conversion residual block is as shown in FIG. 16B, only the frequency band units denoted by reference numerals 1621, 1622, and 1623 have the effective coefficient flag of 1 and reference numerals 1624, 1625, 1626, Since the effective coefficient flag of the frequency band unit denoted by 1627 has a value of 0, the effective frequency band determiner 1920 can determine the frequency band unit in which the effective transform coefficient exists from the extracted effective coefficient flag for each frequency band .

The effective coefficient decoding unit 1930 decodes the effective transform coefficients in the frequency band unit determined by the effective frequency band determining unit 1920 to include the effective transform coefficients. Specifically, the effective coefficient decoding unit 1930 extracts the validity map indicating the position of the effective transform coefficients from the bit stream and the level information of the effective transform coefficient. 15A and 15B, the effective coefficient decoding unit 1930 scans the entire conversion residual block or scans each frequency band unit according to a predetermined scanning order independent for each divided frequency band unit The position of the effective transform coefficient in the transform residual block is determined using the extracted validity map information, and the value of the effective transform coefficient is restored by using the level information.

20 is a flowchart illustrating a method of decoding a residual block according to an embodiment of the present invention.

In step 2010, the effective frequency band determination unit 1920 extracts an effective coefficient flag indicating whether an effective transform coefficient exists for each predetermined frequency band unit obtained by dividing the transformed residual block of the current block from the encoded bit stream.

In step 2020, the frequency band division unit 1910 divides the conversion residual block into predetermined frequency band units. The frequency band division unit 1910 divides the conversion residual block such that the divided unit size of the low frequency band is smaller than the divided unit size of the high frequency band as illustrated in Figs. 14A to 14J, It is possible to divide the transformed residual block by repeating the process of dividing the dual block into quarters and dividing the lowest frequency among the quadrupled transformed residual blocks into quadruples again or by dividing the transformed residual block into the same frequency band units, And the transformed residual block is divided into a plurality of subframes by dividing the transformed residual block by connecting the horizontal frequency and the vertical frequency having the value of the transformed residual block or by using the transformed residual block image characteristic determined using the transform coefficients constituting the transformed residual block, And determines a size to be divided according to the divided size of the determined frequency band, It divides the frozen block. This type of division can be determined by using the division index information added to the bitstream or in a predetermined form as in the encoding side as described above. The order of execution of step 2010 and step 2020 may be changed.

In step 2030, the effective frequency band determination unit 1910 determines a frequency band unit in which effective conversion coefficients exist among the divided frequency band units using the extracted effective coefficient flag. The effective coefficient decoding unit 1930 restores the effective transform coefficient using the validity map and the level information of the effective transform coefficient for the frequency band unit determined to have the effective coefficient.

The image coding and decoding method according to the present invention can also be implemented as computer-readable codes on a computer-readable recording medium. A computer-readable recording medium includes all kinds of recording apparatuses in which data that can be read by a computer system is stored. Examples of the computer-readable recording medium include ROM, RAM, CD-ROM, magnetic tape, floppy disk, optical data storage, and the like. The computer readable recording medium may also be distributed over a networked computer system so that computer readable code can be stored and executed in a distributed manner.

The present invention has been described with reference to the preferred embodiments. It will be understood by those skilled in the art that various changes in form and details may be made therein without departing from the spirit and scope of the invention as defined by the appended claims. Therefore, the disclosed embodiments should be considered in an illustrative rather than a restrictive sense. The scope of the present invention is defined by the appended claims rather than by the foregoing description, and all differences within the scope of equivalents thereof should be construed as being included in the present invention.

Claims (4)

The effective coefficient flag of the specific sub-conversion residual block indicating whether or not at least one effective conversion coefficient exists in the specific sub-conversion residual block among the sub-conversion residual blocks included in the conversion residual block is obtained from the bit stream step;
When the effective coefficient flag of the specific sub-transform residual block indicates that the effective transform coefficient exists in the specific sub-transform residual block, based on the position information and the level information of the effective transform coefficient obtained from the bit stream Obtaining transform coefficients of the specific sub-transform residual block;
If the effective coefficient flag of the specific sub-conversion residual block indicates that the effective conversion coefficient does not exist in the specific sub-conversion residual block, setting the conversion coefficients belonging to the specific sub-conversion residual block to 0; And
Transforming the transformed residual block including the specific sub-transform residual block based on the transform coefficients,
The transform coefficients of the specific sub-transform residual block are part of the transform coefficients included in the transform residual block,
The effective coefficient flag
Wherein the specific sub-transform residual block is obtained when the specific sub-transform residual block does not correspond to the smallest frequency band among the sub-transform residual blocks. The method according to claim 1,
Wherein the position information is a significance map indicating a position of the effective transform coefficients. 3. The method of claim 2,
And the position of the effective transform coefficient in the sub-transform residual block is indicated according to a predetermined scan order independent of each sub-transform residual block. The effective coefficient flag of the specific sub-conversion residual block indicating whether or not at least one effective conversion coefficient exists in the specific sub-conversion residual block among the sub-conversion residual blocks included in the conversion residual block is obtained from the bit stream An entropy decoding unit;
When the effective coefficient flag of the specific sub-transform residual block indicates that the effective transform coefficient exists in the specific sub-transform residual block, based on the position information and the level information of the effective transform coefficient obtained from the bit stream When the effective coefficient flag of the specific sub-transform residual block indicates that the effective transform coefficient does not exist in the specific sub-transform residual block, An effective coefficient decoding unit for setting the transform coefficients belonging to the residual block to 0; And
And an inverse transform unit for inversely transforming the transform residual block including the specific transform transform residual block based on the transform coefficients,
The transform coefficients of the specific sub-transform residual block are part of the transform coefficients included in the transform residual block,
The effective coefficient flag
And when the specific sub-transform residual block does not correspond to the smallest frequency band among the sub-transform residual blocks.

Images