KR20150093969A - Method and apparatus for early mode decision of video - Google Patents

Abstract

The method of determining encoding of a coding unit according to the present invention includes the following steps of: coding by applying a first prediction mode as to the coding unit; calculating a rate-distortion cost and a bit rate when the first prediction mode is applied; determining whether the rate-distortion cost and a preset cost are less than a threshold rate, or the bit rate is less than a preset bit rate threshold rate; and determining the prediction mode as to the coding unit as the first prediction mode if the rate-distortion cost and a preset cost are less than a threshold rate, or the bit rate is less than a preset bit rate threshold rate. Accordingly, provided are a method for encoding of an image and an encoding apparatus for decreasing a calculation amount when determining the prediction mode.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a method for fast encoding an image,

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to an image encoding method and an apparatus using the same, and more particularly, to a method and apparatus for encoding an image to determine an encoding mode at a high speed.

With the recent expansion of HD (High Definition) resolution (1280x1024 or 1920x1080) broadcasting service not only in Korea but also in the world, many users are now getting used to high definition and high definition video. It is spurring development. In addition, as interest in UHD (Ultra High Definition) with resolution more than four times that of HDTV with HDTV increased, video standardization organizations became aware of the necessity of compression technology for higher resolution, high image quality.

In addition, there is a need for a new standard that can achieve the same picture quality through the compression efficiency higher than H.264 / AVC used in HDTV, cell phone, and Blu-ray player,

Currently, Moving Picture Experts Group (MPEG) and Video Coding Experts Group (VCEG) jointly standardize High Efficiency Video Coding (HEVC), the next generation video codec. The video including UHD video is twice as high as H.264 / AVC And aims at encoding with compression efficiency. In addition to HD and UHD images, 3D broadcasting and mobile communication networks can provide high-quality images at lower frequencies.

Since the first JCT-VC (Joint Collaboration Team Video Coding) conference was held in April 2010, HEVC has made a codec named HM (HEVC Test Model) to measure the performance of standards through contributions of various organizations.

In HEVC technology, many techniques have been added to achieve high coding performance, which has greatly increased the complexity of the encoder. Therefore, in order to use HEVC technology in the market, it is necessary to improve the coding speed by reducing the complexity of the encoder.

An object of the present invention is to provide an image coding method and apparatus capable of quickly determining whether coding is to be performed on a coding block.

An object of the present invention is to provide an image encoding method and apparatus capable of reducing the amount of computation when determining a prediction mode.

It is another object of the present invention to provide an image coding method and apparatus capable of reducing unnecessary inter-picture coding prediction mode calculation.

A coding method of a coding unit according to the present invention includes coding a coding unit by applying a first prediction mode to the coding unit; Calculating a rate-distortion cost and a bit amount when the first prediction mode is applied; Determining whether the rate-distortion cost is less than a predetermined cost threshold or the bit amount is less than a predetermined bit amount threshold; Determining a prediction mode for the coding unit to be the first prediction mode when the rate-distortion cost is less than a predetermined cost threshold or the bit amount is less than a preset bit amount threshold.

According to an embodiment of the present invention, there is provided an image encoding method and apparatus capable of quickly determining whether to perform coding on a coding block.

According to an embodiment of the present invention, there is provided an image encoding method and an encoding apparatus capable of reducing a computation amount when a prediction mode is determined.

According to an embodiment of the present invention, there is provided an image encoding method and apparatus capable of reducing unnecessary inter-picture encoding prediction mode calculation.

1 is a block diagram illustrating a configuration of an image encoding apparatus according to an embodiment of the present invention.
2 is a block diagram illustrating a configuration of an image decoding apparatus according to an embodiment of the present invention.
3 is a conceptual diagram schematically showing an embodiment in which one unit is divided into a plurality of lower units.
4 is a diagram for explaining quad-tree coding.
5 is a diagram showing a form of a prediction unit that a coding unit may include according to the present invention.
6 is a control flowchart for explaining a method for determining an encoding mode for each CU.
7 is a control flowchart for explaining a method for determining an encoding mode for individual CUs according to the present invention.
8A and 8B are tables showing experimental results of the encoding mode decision method according to the present invention.
9A and 9B are tables showing experimental results of the conventional encoding mode determination method.

Hereinafter, embodiments of the present invention will be described in detail with reference to the drawings. In describing the embodiments of the present invention, if the detailed description of related known structures or functions is deemed to obscure the subject matter of the present specification, the description may be omitted.

When an element is referred to herein as being "connected" or "connected" to another element, it may mean directly connected or connected to the other element, Element may be present. In addition, the content of " including " a specific configuration in this specification does not exclude a configuration other than the configuration, and means that additional configurations can be included in the scope of the present invention or the scope of the present invention.

The terms first, second, etc. may be used to describe various configurations, but the configurations are not limited by the term. The terms are used for the purpose of distinguishing one configuration from another. For example, without departing from the scope of the present invention, the first configuration may be referred to as the second configuration, and similarly, the second configuration may be named as the first configuration.

In addition, the constituent elements shown in the embodiments of the present invention are shown independently to represent different characteristic functions, which do not mean that each constituent element is composed of separate hardware or a single software constituent unit. That is, each constituent unit is included in each constituent unit for convenience of explanation, and at least two constituent units of each constituent unit may form one constituent unit or one constituent unit may be divided into a plurality of constituent units to perform a function. The integrated embodiments and the separate embodiments of each component are also included in the scope of the present invention unless they depart from the essence of the present invention.

In addition, some of the components are not essential components to perform essential functions in the present invention, but may be optional components only to improve performance. The present invention can be implemented only with components essential for realizing the essence of the present invention, except for the components used for the performance improvement, and can be implemented by only including the essential components except the optional components used for performance improvement Are also included in the scope of the present invention.

First, in order to facilitate the description and to facilitate understanding of the invention, terms used in this specification will be briefly described.

A unit is a unit of image encoding and decoding. In other words, a coding unit or a decoding unit in image coding / decoding refers to a divided unit when one image is divided into subdivided units and is then encoded or decoded. A block, a macroblock, a coding unit or a prediction unit, a transform unit, a coding block, a prediction block or a transform block, etc. . One unit may be divided into smaller-sized lower units.

A transform unit is a basic unit or a unit unit for performing coding / decoding of residual blocks such as transform, inverse transform, quantization, inverse quantization, and transform coefficient coding / decoding, and one transform unit is divided So that the size can be divided into a plurality of small conversion units. In addition, it can be used in the same sense as a conversion block, and a form including a syntax element related to a conversion block for luminance and chrominance signals may be referred to as a conversion unit.

A quantization matrix means a matrix used in a quantization or inverse quantization process to improve the subjective or objective image quality of an image. The quantization matrix is also called a scaling list.

The quantization matrix used for quantization / dequantization may be transmitted as a bitstream, or a default matrix already possessed by the image encoding apparatus and / or the image decoding apparatus may be used. The information of the quantization matrix to be transmitted may be collectively transmitted according to the size of the quantization matrix through a sequence parameter set (SPS) or a picture parameter set (PPS) or the size of a transform block to which a quantization matrix is applied. For example, 4x4 quantization matrices for a 4x4 transform block may be sent, 8x8 matrices for an 8x8 transform block may be sent, 16x16 matrices for a 16x16 transform block may be sent, and 32x32 matrices for a 32x32 transform block may be sent.

The quantization matrix applied to the current block may be obtained by (1) copying a quantization matrix of the same size, or (2) by prediction from a previous matrix coefficient in the quantization matrix. A matrix of the same size may be a previously quantized, decoded, or used quantization matrix, a reference quantization matrix, or a basic quantization matrix. Or a combination including at least two of a quantization matrix, a reference quantization matrix, and a basic quantization matrix which have been previously encoded or decoded or used.

A parameter set corresponds to header information in a structure in a bitstream, and has a meaning collectively referred to as a sequence parameter set, a picture parameter set, an adaptation parameter set, and the like.

The quantization parameter may be a value used in quantization and inverse quantization, and the quantization parameter may be a value mapped to a quantization step size.

The basic matrix may denote a predetermined quantization matrix predefined in the image encoding apparatus and / or the image decoding apparatus, and the basic quantization matrix to be described later in this specification may be used in the same meaning as the basic matrix. The non-default matrix means a quantization matrix that is not previously defined in the image encoding apparatus and / or the image decoding apparatus, but is transmitted from the image encoding apparatus to the image decoding apparatus, that is, transmitted / received by the user And a non-basic quantization matrix to be described later in this specification can be used in the same sense as a non-basic matrix.

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

1, the image encoding apparatus 100 includes a motion prediction unit 111, a motion compensation unit 112, an intra prediction unit 120, a switch 115, a subtractor 125, a transform unit 130, A quantization unit 140, an entropy encoding unit 150, an inverse quantization unit 160, an inverse transformation unit 170, an adder 175, a filter unit 180, and a reference image buffer 190.

The image encoding apparatus 100 may encode an input image in an intra mode or an inter mode and output a bit stream. Intra prediction is intra prediction, and inter prediction is inter prediction. In the intra mode, the switch 115 is switched to the intra mode, and in the inter mode, the switch 115 can be switched to the inter mode. The image encoding apparatus 100 may generate a prediction block for an input block of the input image, and may then code the difference between the input block and the prediction block. At this time, the input image may mean an original picture.

In the intra mode, the intra prediction unit 120 may generate a prediction block by performing spatial prediction using the pixel value of the already coded block around the current block.

In the inter mode, the motion predicting unit 111 may obtain a motion vector by searching an area of the reference image stored in the reference image buffer 190, which is best matched with the input block, in the motion estimation process. The motion compensation unit 112 may generate a prediction block by performing motion compensation using a motion vector. Here, the motion vector is a two-dimensional vector used for inter prediction, and can represent an offset between a current block and a block in the reference image.

The subtractor 125 can generate the residual block by the difference between the input block and the generated prediction block. The transforming unit 130 may perform a transform on the residual block to output a transform coefficient. The quantization unit 140 may quantize the input transform coefficient using at least one of a quantization parameter and a quantization matrix to output a quantized coefficient. At this time, the quantization matrix may be input to the image encoding apparatus, and the input quantization matrix may be determined to be used in the image encoding apparatus.

The entropy encoding unit 150 may output a bit stream by performing entropy encoding based on the values calculated by the quantization unit 140 or the encoding parameter values calculated in the encoding process. When entropy encoding is applied, a small number of bits are allocated to a symbol having a high probability of occurrence, and a large number of bits are allocated to a symbol having a low probability of occurrence, thereby expressing symbols, The size of the column can be reduced. Therefore, the compression performance of the image encoding can be enhanced through the entropy encoding. The entropy encoding unit 150 may use an encoding method such as Exponential-Golomb Code, Context-Adaptive Variable Length Coding (CAVLC), and Context-Adaptive Binary Arithmetic Coding (CABAC) for entropy encoding.

Since the image encoding apparatus 100 according to the embodiment of FIG. 1 performs inter-prediction encoding, that is, inter-view prediction encoding, the currently encoded image needs to be decoded and stored for use as a reference image. Accordingly, the quantized coefficients are inversely quantized in the inverse quantization unit 160 and inversely transformed in the inverse transformation unit 170. The inverse quantized and inverse transformed coefficients become reconstructed residual blocks and added to the prediction block through an adder 175 to generate reconstructed blocks.

The restoration block passes through the filter unit 180 and the filter unit 180 applies at least one of a deblocking filter, a sample adaptive offset (SAO), and an adaptive loop filter (ALF) can do. The filter unit 180 may be referred to as an in-loop filter. The deblocking filter can remove block distortion occurring at the boundary between the blocks. The SAO may add a proper offset value to the pixel value to compensate for coding errors. ALF can perform filtering based on the comparison between the reconstructed image and the original image. The restoration block having passed through the filter unit 180 may be stored in the reference image buffer 190.

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

2, the image decoding apparatus 200 includes an entropy decoding unit 210, an inverse quantization unit 220, an inverse transform unit 230, an intra prediction unit 240, a motion compensation unit 250, an adder 255, a filter unit 260, and a reference image buffer 270.

The video decoding apparatus 200 receives the bit stream output from the video encoding apparatus, decodes the video stream into the intra mode or the inter mode, and outputs the reconstructed video, that is, the reconstructed video. In the intra mode, the switch is switched to the intra mode, and in the inter mode, the switch can be switched to the inter mode. The image decoding apparatus 200 may obtain a reconstructed residual block from the input bitstream, generate a prediction block, and add the reconstructed residual block and the prediction block to generate a reconstructed block, i.e., a reconstructed block .

The entropy decoding unit 210 may entropy-decode the input bitstream according to a probability distribution to generate symbols including a symbol of a quantized coefficient type. The entropy decoding method is similar to the entropy encoding method described above.

When the entropy decoding method is applied, a small number of bits are assigned to a symbol having a high probability of occurrence, and a large number of bits are assigned to a symbol having a low probability of occurrence, so that the size of a bit string for each symbol is Can be reduced.

The quantized coefficients are inversely quantized in the inverse quantization unit 220 using the quantization parameters and inversely transformed in the inverse transformation unit 230. The reconstructed residual blocks can be generated as a result of inverse quantization / inverse transformation of the quantized coefficients.

The quantization matrix used for inverse quantization is also called a scaling list. The inverse quantization unit 220 may generate a dequantized coefficient by applying a quantization matrix to the quantized coefficients.

In this case, the inverse quantization unit 220 may perform inverse quantization in response to the quantization applied in the image encoding apparatus. For example, the inverse quantization unit 220 can perform inverse quantization by applying the quantization matrix applied in the image coding apparatus to the quantized coefficients inversely.

The quantization matrix used in the inverse quantization in the video decoding apparatus 200 may be received from the bitstream, or a basic matrix already held by the video encoding apparatus and / or the video decoding apparatus may be used. The information of the quantization matrix to be transmitted may be collectively received by the quantization matrix size through the sequence parameter set or the picture parameter set or by the conversion block size to which the quantization matrix is applied. For example, 4x4 quantization matrices for a 4x4 transform block may be received, 8x8 matrices for an 8x8 transform block may be received, 16x16 matrices for a 16x16 transform block may be received, and 32x32 matrices for a 32x32 transform block may be received.

In the intra mode, the intraprediction unit 240 may generate a prediction block by performing spatial prediction using the pixel value of the already decoded block around the current block. In the inter mode, the motion compensation unit 250 can generate a prediction block by performing motion compensation using a motion vector and a reference image stored in the reference image buffer 270. [

The reconstructed residual block and the prediction block are added through the adder 255, and the added block can be passed through the filter unit 260. The filter unit 260 may apply at least one of a deblocking filter, SAO, and ALF to a restoration block or a restored picture. The filter unit 260 may output a reconstructed image, that is, a reconstructed image. The restored image is stored in the reference image buffer 270 and can be used for inter prediction.

3 is a conceptual diagram schematically showing an embodiment in which one unit is divided into a plurality of lower units. In the figure, for convenience of explanation, the processing unit of the image is described by taking a coding unit (CU) as an example.

The coding unit CU is a processing unit for processing the above-described video signal, for example intra / inter prediction, transform, quantization and / or entropy coding In this paper, The size of the coding unit used in coding one image may not be constant.

The coding unit may have a rectangular shape, and one coding unit may be divided into several coding units again. For example, one coding unit having a size of 2N x 2N may again be divided into coding units having four NxN sizes. The division of such a coding unit can be made recursively, and not all the coding units need be divided into the same form. However, there may be restrictions on the maximum or minimum size of the coding unit for convenience in coding and processing. If the maximum size of the coding unit is determined, it is called the maximum coding unit size, and if the minimum size is determined, it is called the minimum coding unit size.

In the HEVC encoder, coding is performed based on a CTU (Coding Tree Unit). CTUs can have sizes of 64x64, 32x32, and 16x16, and generally use 64x64. Recursive tree coding is performed based on the CTU. In the CTU, prediction and transcoding are performed based on a CU (Coding Unit) corresponding to each block divided into a Quadtree type. The CU can have values of 64x64, 32x32, 16x16, and 8x8, and has a structure that can minimize the rate-distortion cost while having a quadtree structure in the CTU.

As shown in FIG. 3, a CU that requires encoding in order to find a structure having a quadtree structure and a minimum rate-distortion cost for 64x64 CTUs can be finely divided.

In the YUV 4: 2: 0 format, encoding is performed for one 64x64 CU, four 32x32 CU, 16 16x16 CU, and 64 8x8 CUs to find the optimal structure for the Y component.

For one coding unit, information indicating whether or not the coding unit is divided may be specified. For example, if the flag indicating the division is 1, the block corresponding to the corresponding node is divided into four blocks again. If the flag is 0, the processing process for the corresponding coding unit can be performed without further division.

4 is a diagram for explaining quad-tree coding.

The structure of the coding unit described above can be represented using a recursive tree structure. That is, a coding unit that is divided into other coding units, with one image or a maximum size coding unit root, has as many child nodes as the number of the divided coding units. Thus, the coding unit which is not further divided becomes a leaf node. Assuming that only one square division is possible for one coding unit, one coding unit may be divided into a maximum of four different coding units, so that the tree representing the coding unit structure will be in the form of a quad tree.

In the encoder, the maximum and minimum coding unit sizes may be determined according to the characteristics (for example, resolution) of the video image or the efficiency of coding, and information about the maximum and minimum coding unit sizes may be included in the bitstream.

For example, when the maximum coding unit size and the maximum depth of the tree are defined, if the square division is performed, the height and width of the coding unit are half the height and width of the coding unit of the parent node, You can get the minimum coding unit size. Conversely, the minimum coding unit size and the maximum depth of the tree can be predefined and utilized, and the maximum coding unit size can be derived and used if necessary. Since the size of a unit changes in a square of 2 in a square partition, the size of the actual coding unit is expressed as a log value of 2 or less, which can increase the transmission efficiency.

The decoder may obtain information indicating whether the current coding unit is partitioned. This information can only be obtained (sent) under certain conditions to increase efficiency. For example, if the current coding unit is an SCU, it is no longer divided into smaller coding units, so in this case, it is not necessary to acquire information indicating that it is divided.

If the information indicates that the coding unit has been divided, the size of the divided coding unit is half of the current coding unit, and is divided into four square coding units based on the current processing position. The above processing can be repeated for each divided coding unit.

A unit or block can be hierarchically partitioned with depth information based on a tree structure. Each divided subunit may have depth information. The depth information may include information on the size of the lower unit, as the unit indicates the number and / or degree of division.

Referring to 410 of FIG. 4, the uppermost node may be referred to as a root node and may have the smallest depth value. At this time, the uppermost node may have a depth of level 0, and may represent the first unpartitioned unit.

A lower node having a depth of level 1 may represent a unit in which the first unit is divided once and a lower node having a depth of level 2 may represent the unit in which the first unit is divided twice. For example, the unit a corresponding to the node a in FIG. 4 is a once-divided unit in the initial unit and may have a depth of level 1.

A level 3 leaf node may represent a unit in which the first unit is divided three times. For example, the unit d corresponding to the node d in 420 of FIG. 4 is a unit divided three times in the initial unit and can have a depth of level 3. Therefore, the leaf node of level 3, which is the lowest node, can have the deepest depth.

5 is a diagram showing a form of a prediction unit that a coding unit may include according to the present invention.

The image prediction for coding can be made on a coding unit that is no longer divided (i.e., the leaf node of the coding unit tree). The current coding unit is divided into one or more prediction units (PU) or prediction blocks or partitions, and this behavior itself is also referred to as a partition. The prediction unit is a basic unit for performing prediction, and the types of prediction blocks available are different in the intra-picture coding unit and the inter picture coding unit.

(a) shows a prediction unit for a coding unit that performs intra prediction. As shown, the coding unit 2Nx2N for performing intra prediction can divide a 2Nx2N prediction unit having the same size as the coding unit and a 2Nx2N coding unit into a prediction divided into NxN.

(b) shows a prediction unit for a coding unit that performs inter picture prediction. As shown in the figure, the coding unit 2Nx2N for performing inter picture prediction includes a prediction unit divided into 2Nx2N and 2NxN using the same size as the coding unit, a prediction unit divided into Nx2N, a prediction unit divided into NxN, A prediction unit divided into nRx2N, a prediction unit divided into 2NxnU, and a prediction unit divided into 2NxnD.

Currently, JCT-VC is working on standards for High Efficiency Video Coding (HEVC) technology as the next generation video standard. HEVC aims at compression efficiency of more than 50% compared to H.264 / AVC which provides the best compression ratio. Many coding tools have been proposed for this purpose. Conventional video codecs are coded in macroblocks (16x16), while HEVC is optimized for CU (Coding Unit), PU (Prediction Unit) and TU (Transform Unit) Of the block size.

On the other hand, the image encoding method according to the HEVC requires a large amount of computation to calculate the inter picture coding mode. This is because the basic size of the block coding process is divided into the child CUs of the same size as the tree structure, and the calculation is performed on the PU and TU each time the image coding method according to the HEVC is H.264 / AVC. It is because.

The CU is divided into 32x32, 16x16, and 8x8 blocks from a maximum size 64x64 (LCU: Largest Coding Unit), and each CU calculates the optimum block size of the optimal PU and TU. Therefore, SKIP, Inter_2Nx2N, Inter_2NxN, Inter_N2N, Inter_2NxnU, Inter_2NxnD, Inter_nRx2N, Inter_nLx2N, and Intra_PCM modes are supported in one CU size.

If the current CU size is 8x8, Rate-Distortion Optimization (RDO) is performed for Inter_NxN and Intra_NxN modes to determine the most optimal block mode in one CU size.

There is a need for a method capable of efficiently terminating the inter-picture coding mode efficiently because the complexity of inter picture coding itself is considerably larger than that of intra picture coding.

Accordingly, the present invention provides a method of reducing the complexity of an encoder by early determining a block division process in a current CU (Coding Unit) during video coding, and an apparatus using the method.

When coding these CTUs, it is necessary to calculate the RDO for each size and find a coding unit structure that can code the best, that is, the best quality for the smallest bit amount. For example, the coding unit structure may be determined to be 64x64, the coding unit may be determined to be 32x32, or a more complex structure may be split. In this specification, a coding unit estimated with an optimal block size means having a coding unit structure that is divided into coding units of various sizes for the best coding among arbitrary CTUs.

6 is a control flowchart for explaining a method for determining an encoding mode for each CU.

The M (i) shown at the bottom right of FIG. 6 represents a prediction mode and a division type applicable to the CU. The first mode M (1) means 2Nx2N to which the skip mode and the merge mode are applied, and the second mode (M (2)) means inter prediction, which is not a skip mode and merge mode, Indicates the partition type.

The second mode M (2) means 2Nx2N to which the motion vector prediction is applied, the third mode M (3) means Nx2N to which inter prediction is applied, the fourth mode M (4) Means 2NxN to which inter prediction is applied.

The fifth through eighth modes represent an asymmetric division type. M (6) means 2NxnU to which inter prediction is applied, and seventh mode (M (7)) means 2N x nD to which inter prediction is applied. Denotes the nLx2N to which the inter prediction is applied, and the eighth mode (M (8)) denotes the nRx2N to which the inter prediction is applied.

The ninth and tenth modes indicate the intra prediction mode. The ninth mode (M (9)) means 2Nx2N to which intra prediction is applied, and the tenth mode (M (10)) means NxN to which intra prediction is applied.

First, 1 is substituted into i (S610), and the CU is coded using the first mode (M (i)) (S620).

That is, in order to obtain the prediction mode for the CU, the CU is sequentially applied from the first mode to the CU one by one.

It is determined whether i is 1 (S630). If i is 1, J (i) is assigned to Jbest (S640).

J (i) denotes the RD cost incurred in the encoding mode.

When J (i) is substituted for Jbest, i + 1 is substituted for i (S660).

If i is not 1, it is determined whether J (i) is smaller than Jbest (S650).

If J (i) is smaller than Jbest, Jbest is set to J (i) (S640) and i + 1 is substituted for i (S660).

Even if J (i) is not smaller than Jbest, i + 1 is substituted for i (S660).

It is determined whether or not the newly set i is greater than 10 (S670), and if i is greater than 10, the encoding mode determination step ends. If the total coding mode is 10, it is determined in step S670 whether i is larger than 10, and the value to be compared according to the total number of coding modes can be changed.

If i is less than 10, it is determined whether M (i) for the newly set i should be considered (S680).

If M (i) for i is to be considered, step S620 is entered and the CU is coded using mode M (i) (S620).

On the other hand, if M (i) for i does not need to be considered, the process goes to step S660 in which i + 1 is substituted for i.

6, one CTU can be divided into a quad tree structure like 420 of FIG. 420 of FIG. 4 is an example of a quadtree structure having a minimum rate-distortion cost in 64x64 CTUs determined after all the encoding processes. In this specification, the rate-distortion cost and the RD cost can be used in the same sense.

As described, since a maximum of 10 coding modes are checked in each CU to find a coding mode with a minimum rate-distortion (RD) cost, a quad tree structure with a minimum rate-distortion cost is found in the CTU The process has a lot of complexity.

On the other hand, various methods have been proposed to quickly find the mode and shape having the minimum rate-distortion cost among the encoding modes.

Here, the mode is a concept including both intra prediction and inter prediction. There are skip mode, merge mode, MVP (motion vector prediction) mode, and the like, as types of prediction methods according to inter prediction. The shape means a type when the CU is divided into the prediction blocks, which means any one of FIG. 5B.

In one of the various methods, when inter prediction is performed during the process of determining the optimal encoding mode of the CU, if the coded block flag (cbf) values of Y, U, and V are all 0, That is, a method of omitting the operation for the remaining shape.

According to another example, if the DMV (Differential Motion Vector) and cbf are both zero after calculating the inter prediction mode in order to obtain the optimal coding mode of the CU, the calculation for the other modes and modes is omitted.

Since only the cbf is checked in the first method, the amount of encoded bits generated for the signaling of the prediction information is not considered. Therefore, loss of coding efficiency may occur.

In the second method, the amount of bits for coding a motion vector is further considered in the prediction information. However, there is a restriction condition for performing the process for obtaining the inter prediction mode first.

7 is a control flowchart for explaining a method for determining an encoding mode for individual CUs according to the present invention.

M (i) shown in the lower right of FIG. 7 represents a prediction mode and a division type applicable to the CU. The first mode M (1) means 2Nx2N to which the skip mode and the merge mode are applied, and the second mode M (2) to the eighth mode M (8) Inter prediction, that is, a division type to which motion vector prediction is applied. The ninth (M (9)) and tenth (M (10)) modes indicate the intra prediction mode. The details are the same as in Fig.

First, 1 is substituted into i (S710), and the CU is coded using the first mode (M (i)) (S720). In this process, the RD cost and the bit amount are calculated for the prediction mode of i = 1.

That is, in order to obtain the prediction mode for the CU, the CU is sequentially applied from the first mode to the CU one by one.

i is 1 (S730). If i is 1, J (i) is substituted for Jbest and B (i) is substituted for Bbest (S740).

If i is not 1, it is determined whether J (i) is smaller than Jbest (S750).

J (i) denotes an RD cost generated in each coding mode in the process of determining an encoding mode, and B (i) denotes a bit amount.

When the RD cost and the bit amount are calculated for the prediction mode of i at i in S720, whether the JTH corresponding to the threshold value of the RD cost is greater than Jbest or the BTH corresponding to the bit amount threshold is greater than Bbest (S760).

According to the present invention, in a process of determining an encoding mode, when the RD cost and the bit amount generated in each encoding mode are compared with JTH and BTH, which are threshold values, if either one of them is smaller than the threshold value, So that the encoding mode is determined early.

That is, when the coding unit is coded using the specific prediction mode, if the RD cost and the bit amount are compared with the respective threshold values JTH and BTH, and if either one of them is smaller than the threshold value, The specific prediction mode can be determined as the prediction mode of the coding unit.

This method can be applied to a picture coded in the inter prediction mode.

JTH can be determined by equation (1) as a threshold value for RD cost.

Equation (1)

In Equation (1), i denotes an encoding mode, 2N denotes a transverse length (number of pixels) of a CU, and TL denotes a temporal layer of a picture to be encoded. WTL is a weighting factor according to TL, and Dpred is a distortion value in a co-located CTU of a picture located closest to the playback order among coded pictures. When a plurality of pictures located nearest to each other in the reproduction order are used, recently coded pictures are used. lambda mode is a Lagrangian constant, and B is a value for considering the generated bit amount at a bit cost.

BTH represents a threshold value for the generated bit amount.

When coding in a random access configuration according to an embodiment of the present invention, the WTL required to determine the JTH value is 0.7 in TL 2, 0.9 in TL 3, 1.0 in TL 4, and 10 in B Can be set.

Also, the threshold value BTH for the bit amount can be set as shown in equation (2).

Equation (2)

If Jbest is smaller than JTH or Bbest is smaller than BTH, the encoding mode is determined early by omitting the operation for the remaining encoding mode. That is, the step of determining the encoding mode is ended.

On the other hand, if Jbest is larger than JTH and Bbest is larger than BTH, i + 1 is substituted for i (S770).

It is determined whether or not the newly set i is greater than 10 (S780), and if i is greater than 10, the encoding mode determination step ends.

If i is less than 10, it is determined whether M (i) for the newly set i should be considered (S790).

If M (i) for i is to be considered, step S720 is entered and the CU is coded using mode M (i) (S720).

On the other hand, if M (i) for i does not need to be considered, the process goes to step S770 in which i + 1 is substituted for i.

8A and 8B are tables showing experimental results of the encoding mode decision method according to the present invention. That is, FIGS. 8A and 8B are tables in which the result values for various images are summarized in the case of applying FIG.

In FIGS. 8A and 8B, Encoding Time Reduction (ΔETR) is a value obtained by the following equation (3), and represents a decrease in the time required for coding. The BD-rate (Bjontegaard Delta-rate) is a value indicating the degree of loss in coding efficiency.

(Equation 3)

? ETR (%) = (T1-T2) / T2 * 100%

T1: Time spent coding using the designed fast coding mode

T2: Time spent coding without using the designed fast coding mode

As a result, it can be confirmed that the coding efficiency of the A group (class A) is 0.43% while improving the coding rate (AVG pf class A) of about 44%. In the B group (class B), the coding efficiency (overall AVG) is 0.43% while improving the coding rate (AVG pf class B) of about 50%.

The overall AVG coding speed of the A to F groups improved by about 44.8% and the coding efficiency loss was found to be 0.42%.

9A and 9B are tables showing experimental results of the conventional encoding mode determination method. In the fast encoding mode determination method compared with the present invention, JCTVC-F045 determines that the coded block flag (cbf) values of Y, U, and V are all 0 when inter prediction is performed in the process of determining the optimal encoding mode of the CU. If the DMV (Differential Motion Vector) and cbf are both 0 after performing the inter prediction mode, then JCTVC-G543 does not perform calculation for different shapes and modes Means to omit.

According to the existing first fast encoding mode determination method, the coding efficiency is improved by about 39%, and the coding efficiency loss is 0.94%. According to the second fast coding mode determination method, the coding efficiency is improved by about 33%, and the coding efficiency loss is 0.25%.

8A and 8B are compared with FIGS. 9A and 9B, it can be seen that the coding rate is much improved and the encoding efficiency is reduced when the fast coding mode determining method according to the present invention is compared.

In the above-described exemplary system, the methods are described on the basis of a flowchart as a series of steps or blocks, but the present invention is not limited to the order of the steps, and some steps may occur in different orders or simultaneously . In addition, the above-described embodiments include examples of various aspects. For example, combinations of the embodiments are also to be understood as one embodiment of the present invention.

100: image encoding device 111: motion prediction unit
112: motion compensation unit 120: intra prediction unit
115: Switch 125:
130: conversion unit 140: quantization unit
150: an entropy encoding unit 160: an inverse quantization unit
170: inverting section 180: filter section

Claims (1)

A coding method of a coding unit,
Coding the coding unit by applying a first prediction mode to the coding unit;
Calculating a rate-distortion cost and a bit amount when the first prediction mode is applied;
Determining whether the rate-distortion cost is less than a predetermined cost threshold or the bit amount is less than a predetermined bit amount threshold;
And determining the prediction mode for the coding unit to be the first prediction mode when the rate-distortion cost is smaller than a predetermined cost threshold value or the bit amount is smaller than a preset bit amount threshold value. Wherein the encoding unit decodes the encoded data.

Images