WO2014081261A1

WO2014081261A1 - Method and device for encoding/decoding video using motion information merging

Info

Publication number: WO2014081261A1
Application number: PCT/KR2013/010745
Authority: WO
Inventors: 문주희; 이재영; 한종기
Original assignee: 인텔렉추얼 디스커버리 주식회사
Priority date: 2012-11-23
Filing date: 2013-11-25
Publication date: 2014-05-30
Also published as: KR20150105346A

Abstract

A method for decoding video using motion information, according to one embodiment of the present invention, comprises the steps of: decoding partition information from a received code; determining a partition form of a largest coding unit (LCU) according to the partition information; parsing an area flag from the code according to the partition form of the LCU; and determining whether to merge motion information on coding units included in the LCU on the basis of the area flag, and decoding the coding units.

Description

Method and apparatus for video encoding / decoding using motion information merging

The present invention relates to a video codec, and more particularly, to a video encoding providing method and apparatus that inherit information from neighboring blocks of a coding block during encoding and decoding.

In general video encoding and decoding, when a current coding block is an inter coded block, MVP is calculated based on a motion vector (MV) and a reference image index of neighboring blocks, or a merge skip mode is considered. When the current coding block is an intra coded block, an MPM for intra coding the current block is calculated based on intra prediction modes of neighboring blocks. However, in this general method, there is a problem that the kind of information inherited from neighboring blocks is limited.

An embodiment of the present invention provides a method of inheriting more various information from neighboring blocks than in the video codec.

However, the technical problem to be achieved by the embodiment of the present invention is not limited to the technical problems as described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, the video encoding providing method according to the first aspect of the present invention increases the type and number of information that can be inherited from neighboring blocks, and a plurality of previously encoded and decoded Information is extracted from the pictures and used for encoding.

In addition, a video decoding method using motion information according to an embodiment of the present invention for achieving the above technical problem, the video decoding method, comprising: decoding the segmentation information from the received code; Determining a partition type of a large coding unit (LCU) according to the partition information; Parsing an area flag from the code according to the partition type of the LCU; And determining whether to merge motion information for coding units included in the LCU based on the region flag, and decoding the coding units.

According to the present invention, coding efficiency can be improved by inheriting more various information from neighboring blocks than now.

1 is a block diagram illustrating an example of a configuration of a video encoding apparatus.

2 is a block diagram illustrating an example of a structure of an inter prediction encoding apparatus.

3 is a block diagram illustrating an example of a configuration of an inter prediction decoding apparatus.

FIG. 4 is a diagram for explaining an example of a method of considering a merge / merge skip.

5 is a view for explaining a method of determining a merge / merge skip according to an embodiment of the present invention.

6 to 8 are diagrams illustrating embodiments of peripheral regions that are merged / merge skipped.

9 shows a general coding block structure and a method of determining the same.

10 shows an example of merge candidates.

11 illustrates a header structure of a coding unit (CU) and a prediction unit (PU).

12 illustrates an image segmentation difference according to resolution.

FIG. 13 is a diagram for explaining an LCU segmentation method for encoding integrated Merge information.

FIG. 14 is a diagram for explaining difference of region division information of neighboring LCUs. FIG.

15 shows a method of encoding split information of a current LCU.

16 shows an example of a CU encoding order in a divided region.

17 is a diagram illustrating coding of CUs and PUs in a split LCU according to an embodiment of the present invention.

18A illustrates Merge Flag transmission in a normal region.

18B illustrates Merge Flag transmission in the Merge region.

19 illustrates an LCU unit decoding flow chart according to an embodiment of the present invention.

DETAILED DESCRIPTION Hereinafter, exemplary embodiments of the present disclosure will be described in detail with reference to the accompanying drawings so that those skilled in the art may easily implement the present disclosure. As those skilled in the art would realize, the described embodiments may be modified in various different ways, all without departing from the spirit or scope of the present invention. In the drawings, parts irrelevant to the description are omitted for simplicity of explanation, and like reference numerals designate like parts throughout the specification.

Throughout this specification, when a portion is "connected" to another portion, this includes not only "directly connected" but also "electrically connected" with another element in between. do.

Throughout this specification, when a member is located “on” another member, this includes not only when one member is in contact with another member but also when another member exists between the two members.

Throughout this specification, when a part is said to "include" a certain component, it means that it can further include other components, without excluding the other components unless specifically stated otherwise. As used throughout this specification, the terms "about", "substantially" and the like are used at, or in the sense of, numerical values when a manufacturing and material tolerance inherent in the stated meanings is indicated, Accurate or absolute figures are used to assist in the prevention of unfair use by unscrupulous infringers. As used throughout this specification, the term "step to" or "step of" does not mean "step for."

Throughout this specification, the term “combination of these” included in the expression of the makushi form means one or more mixtures or combinations selected from the group consisting of constituents described in the expression of the makushi form, wherein the constituents It means to include one or more selected from the group consisting of.

As an example of a method of encoding a real image and its depth information map, standardization is jointly performed by the Moving Picture Experts Group (MPEG) and the Video Coding Experts Group (VCEG), which have the highest coding efficiency among the video coding standards developed to date. Encoding may be performed by using high efficiency video coding (HEVC).

1 is a block diagram illustrating an example of a configuration of a video encoding apparatus and illustrates a coding structure diagram of HEVC.

As shown in FIG. 1, HEVC includes various new algorithms such as coding units and structures, inter prediction, intra prediction, interpolation, filtering, and transform methods.

2 is a block diagram illustrating an example of a structure of an inter prediction encoding apparatus. The inter prediction encoding apparatus includes a motion information determiner 110, a motion information encoding mode determiner 120, a motion information encoder 130, and a prediction. The block generator 140, the residual block generator 150, the residual block encoder 160, and the multiplexer 170 may be configured.

Referring to FIG. 2, the motion information determiner 110 determines motion information of the current block. The motion information includes a reference picture index and a motion vector. The reference picture index represents any one of the previously coded and reconstructed pictures. When the current block is unidirectional inter prediction coded, it indicates any one of the reference pictures belonging to list 0 (L0).

On the other hand, when the current block is bidirectional predictively coded, the current block may include a reference picture index indicating one of the reference pictures of list 0 (L0) and a reference picture index indicating one of the reference pictures of list 1 (L1). . In addition, when the current block is bidirectional predictively coded, the current block may include an index indicating one or two pictures of reference pictures of the composite list LC generated by combining the list 0 and the list 1.

The motion vector indicates the position of the prediction block in the picture indicated by each reference picture index. The motion vector may be pixel unit (integer unit) or sub pixel unit. For example, it may have a resolution of 1/2, 1/4, 1/8 or 1/16 pixels. If the motion vector is not an integer unit, the prediction block is generated from pixels of an integer unit.

The motion information encoding mode determiner 120 determines whether to encode the motion information of the current block in the skip mode, the merge mode, or the AMVP mode.

The skip mode is applied when a skip candidate having the same motion information as the motion information of the current block exists and the residual signal is zero. Also, the skip mode is applied when the current block is the same size as the coding unit. The current block can be viewed as a prediction unit.

The merge mode is applied when there is a merge candidate having the same motion information as that of the current block. The merge mode is applied when a residual signal exists when the current block has a different size or the same size as the coding unit. The merge candidate and the skip candidate may be the same.

AMVP mode is applied when skip mode and merge mode are not applied. An AMVP candidate having a motion vector most similar to the motion vector of the current block is selected as an AMVP predictor.

The motion information encoder 130 encodes the motion information according to a method determined by the motion information encoding mode determiner 120. If the motion information encoding mode is the skip mode or the merge mode, the merge motion vector encoding process is performed. When the motion information encoding mode is AMVP, the AMVP encoding process is performed.

The prediction block generator 140 generates a prediction block by using motion information of the current block. If the motion vector is an integer unit, the block corresponding to the position indicated by the motion vector in the picture indicated by the reference picture index is copied to generate a prediction block of the current block.

However, when the motion vector is not an integer unit, pixels of the prediction block are generated from integer unit pixels in the picture indicated by the reference picture index. In this case, in the case of the luminance pixel, the prediction pixel may be generated using an interpolation filter of 8 taps. In the case of chrominance pixels, a prediction pixel may be generated using a 4-tap interpolation filter.

The residual block generator 150 generates a residual block by using the current block and the prediction block of the current block. When the size of the current block is 2Nx2N, the residual block is generated by using the current block and a prediction block having a size of 2Nx2N corresponding to the current block.

However, when the size of the current block used for prediction is 2NxN or Nx2N, the prediction block for each of the 2NxN blocks constituting the 2Nx2N is obtained, and then the final prediction block having the size of 2Nx2N is obtained using the two 2NxN prediction blocks. Can be generated. In addition, a residual block of 2Nx2N may be generated by using the prediction block having a size of 2Nx2N. In order to solve the discontinuity of the boundary portion of two prediction blocks having a size of 2N × N, overlapping pixels of the boundary portion may be smoothed.

The residual block encoder 160 divides the generated residual block into one or more transform units. Each transform unit is then transform coded, quantized, and entropy coded. In this case, the size of the transform unit may be determined by the quadtree method according to the size of the residual block.

The residual block encoder 160 transforms the residual block generated by the inter prediction method using an integer-based transform matrix. The transformation matrix is an integer based DCT matrix. The residual block encoder 160 uses a quantization matrix to quantize coefficients of the residual block transformed by the transform matrix. The quantization matrix is determined by the quantization parameter. The quantization parameter is determined for each coding unit of a predetermined size or more. The predetermined size may be 8x8 or 16x16.

Therefore, when the current coding unit is smaller than the predetermined size, only the quantization parameter of the first coding unit in the coding order among the plurality of coding units within the predetermined size is encoded, and since the quantization parameter of the remaining coding units is the same as the parameter, no need.

The coefficients of the transform block are quantized using the quantization matrix determined according to the determined quantization parameter and the prediction mode.

The quantization parameter determined for each coding unit of the predetermined size or more is predictively coded using the quantization parameter of the coding unit adjacent to the current coding unit. The left coding unit of the current coding unit and the upper coding unit may be searched to generate a quantization parameter predictor of the current coding unit using one or two valid quantization parameters.

For example, the first valid quantization parameter found in the above order may be determined as a quantization parameter predictor. In addition, the first quantization parameter that is valid may be determined as a quantization parameter predictor by searching the left coding unit and the coding unit immediately before the coding order.

The coefficients of the quantized transform block are scanned and converted into one-dimensional quantization coefficients. The scanning method may be set differently according to the entropy encoding mode. For example, when coded with CABAC, the inter prediction coded quantization coefficients may be scanned in one predetermined manner (zigzag or diagonal raster scan). On the other hand, when encoded by CAVLC, scanning may be performed in a manner different from that described above. For example, the scanning method may be determined according to a zigzag in case of inter and an intra prediction mode in case of intra.

In addition, the coefficient scanning scheme may be determined differently according to the size of the transform unit. The scan pattern may vary depending on the directional intra prediction mode. The scan order of the quantization coefficients scans in the reverse direction.

The multiplexer 170 multiplexes the motion information encoded by the motion information encoder 130 and the residual signals encoded by the residual block encoder. The motion information may vary according to an encoding mode. That is, in case of skip or merge, only the index indicating the predictor is included. However, in the case of AMVP, it includes a reference picture index, a differential motion vector, and an AMVP index of the current block.

3 is a block diagram illustrating an example of a configuration of an inter prediction decoding apparatus. The inter prediction decoding apparatus 200 includes a demultiplexer 210, a motion information encoding mode determiner 220, and a merge mode motion information decoder 230. ), The AMVP mode motion information decoder 240, the prediction block generator 250, the residual block decoder 260, and the reconstructed block generator 270.

Referring to FIG. 3, the demultiplexer 210 demultiplexes currently encoded motion information and encoded residual signals from a received bitstream. The demultiplexer 210 transmits the demultiplexed motion information to the motion information encoding mode determiner 220 and transmits the demultiplexed residual signal to the residual block decoder 260.

The motion information encoding mode determiner 220 determines the motion information encoding mode of the current block. When skip_flag of the received bitstream has a value of 1, the motion information encoding mode determiner 220 determines that the motion information encoding mode of the current block is encoded by the skip encoding mode. The motion information encoding mode determiner 220 has a skip_flag of the received bitstream having a value of 0, and the motion information encoding mode of the current block in which the motion information received from the demultiplexer 210 has only a merge index. It is determined that it is encoded by.

If the skip_flag of the received bitstream has a value of 0 and the motion information received from the demultiplexer 210 has a reference picture index, a differential motion vector, and an AMVP index, the motion information encoding mode determiner 220 includes: It is determined that the motion information encoding mode of the current block is encoded in the AMVP mode.

The merge mode motion information decoder 230 is activated when the motion information encoding mode determiner 220 determines that the motion information encoding mode of the current block is a skip or merge mode.

The AMVP mode motion information decoder 240 is activated when the motion information encoding mode determiner 220 determines that the motion information encoding mode of the current block is the AMVP mode.

The prediction block generator 250 generates the prediction block of the current block by using the motion information reconstructed by the merge mode motion information decoder 230 or the AMVP mode motion information decoder 240. If the motion vector is an integer unit, the block corresponding to the position indicated by the motion vector in the picture indicated by the reference picture index is copied to generate a prediction block of the current block.

However, when the motion vector is not an integer unit, the pixels of the prediction block are generated from the integer unit pixels in the picture indicated by the reference picture index. In this case, in the case of the luminance pixel, the prediction pixel may be generated using an interpolation filter of 8 taps. In the case of chrominance pixels, a prediction pixel may be generated using a 4-tap interpolation filter.

The residual block decoder 260 entropy decodes the residual signal. Inverse scanning of the entropy decoded coefficients generates a two-dimensional quantized coefficient block. The reverse scanning method may vary depending on the entropy decoding method.

That is, the inverse scanning scheme of the inter prediction residual signal when the CABAC decoding is performed and the CAVLC decoding may be different. For example, when the CABAC-based decoding is performed, the diagonal raster inverse scan method may be used, and when the CAVLC-based decoding is performed, the zigzag inverse scanning method may be used. In addition, the inverse scanning scheme may be determined differently according to the size of the prediction block.

The residual block decoder 260 dequantizes the generated coefficient block using an inverse quantization matrix. Restore quantization parameters to derive the quantization matrix. The quantization step size is reconstructed for each coding unit of a predetermined size or more.

The predetermined size may be 8x8 or 16x16. Therefore, when the current coding unit is smaller than the predetermined size, only the quantization parameter of the first coding unit in the coding order of the plurality of coding units within the predetermined size is restored, and since the quantization parameter of the remaining coding units is the same as the parameter, the encoding is performed. There is no need to do it.

The quantization parameter of the coding unit adjacent to the current coding unit is used to recover the quantization parameter determined for each coding unit of the predetermined size or more. The first valid quantization parameter may be determined as a quantization parameter predictor of the current coding unit by searching in the order of the left coding unit and the upper coding unit of the current coding unit. Further, the first valid quantization parameter may be determined as a quantization parameter predictor by searching the left coding unit and the coding unit immediately before the coding order.

The quantization parameter of the current prediction unit is recovered using the determined quantization parameter predictor and the differential quantization parameter.

The residual block decoder 260 inversely transforms the inverse quantized coefficient block to restore the residual block.

The reconstruction block generation unit 270 generates a reconstruction block by adding the prediction block generated by the prediction block generation unit 250 and the residual block generated by the residual block decoding unit 260.

When the current coding block is an inter coded block, MVP is calculated based on a motion vector (MV) and a reference image index of blocks already coded around it, or a merge mode and a merge skip mode are considered.

Referring to FIG. 4, motion vectors of a left block A1, a lower left block A2, an upper right block B0, an upper block B1, and an upper left block B2 of the current block are defined as Spatial candidate motion vectors. In addition, the motion vector at the lower right (H) and the middle (C3) positions of the Col-located block in the reference frame is considered as a temporal candidate motion vector. Considering these candidate MVs, it is considered whether the current coding block can be encoded in Merge or Merge Skip mode.

The concept of merge or merge skip considered in the methods described with reference to FIG. 4 is limited to be applied to an ultra high resolution image, and an enlarged concept of merge and merge skip needs to be used.

According to an embodiment of the present invention, unlike merge or merge skip, which is limited in consideration of similarities with neighboring blocks adjacent to the current coding block, in the present invention, neighboring blocks located in a large area around the current coding block Considering the similarity of, we propose extended merge and merge skip concept.

In other words, when considering Merge / Merge Skip, since the range of neighboring blocks referencing the information is limited to the blocks in the upper row and the blocks on the left, based on the current block, encoding in the process of encoding a super high resolution image Efficiency is limited. In order to solve this problem, the present invention determines Merge / Merge Skip in consideration of various information of a wider range of neighboring blocks.

Referring to FIG. 5, when B (i, j) is a current coding block, the current block is encoded using coding information of blocks that have already been encoded and reconstructed around the neighbor.

Considering Merge / Merge Skip, since the range of neighboring blocks referencing information is limited to the blocks on the top row and the blocks on the left, based on the current block (i.e., B (i, j-1), Only coding information of B (i-1, j-1), B (i-1, j), and B (i-1, j + 1) is used, and encoding efficiency is limited in the process of encoding an ultra high resolution image. .

To solve this, one embodiment of the present invention uses a wider range of peripheral blocks. In this case, the information of the neighboring blocks used for merging is added not only motion information (MV, MVD, reference picture number, etc.) but also quantization parameter, PU partition information, merge information, and the like. Decide In other words, the concept of Merge extends much more diversely, and Merge's mode is also supported in various ways.

That is, the number of neighboring blocks that can be referred to may be increased to refer to B (i-k, j-m) located farther. Here, k and m may have a variable value according to the reference block region.

6 to 8 illustrate embodiments of neighboring regions that are merged / merge skipped, and blocks located in a wider area having similar motion information may be expressed only by a simple merge / merge skip flag, thereby increasing compression efficiency. .

That is, according to embodiments of the present invention, the encoding efficiency is increased in the process of encoding a super high resolution image by using information of the blocks in the immediately upper row and the blocks on the left immediately based on the current encoding block for limited information reference. More neighboring blocks are used for informational reference.

For example, referring to FIG. 6, the current coding block to be encoded may be merged / merge skipped with blocks of a wider area as well as neighboring blocks directly attached thereto.

For example, referring to FIG. 7, a current coding block to be encoded may be merged / merge skipped with blocks at arbitrary positions.

For example, referring to FIG. 8, a current coding block to be encoded may be merged / merge skipped with further neighboring blocks except directly attached neighboring blocks.

초고해상도 영상을 위한 확장된 영역의 Merge 플래그 표현 방법Merge Flag Representation of Extended Region for Ultra-High Resolution Images

Hereinafter, a method of extending the concept of merge for an ultra high resolution image will be described.

As described above, by using coding units of various sizes, the coding may be performed by considering the spatial resolution and the block characteristics of an image effectively.

9 shows a general coding block structure and a method of determining the same.

As shown in FIG. 9, in general, when the resolution of the image is small or the pixel values change locally, it may be efficient to perform intra and inter prediction using coding units having a small size. The use of a small coding unit increases the amount of header bits required for encoding, but has the advantage of reducing the amount of bits required for encoding quantization errors and transform coefficients due to relatively accurate prediction.

On the contrary, in a region having a large spatial resolution or a small change in pixel values, using a large coding unit may increase coding efficiency. In this case, even when a large coding unit is used, the prediction error does not tend to increase as much as when a small coding unit is predicted. Therefore, when encoding these blocks, it is preferable to use a large coding unit to save the header bit amount. It can be efficient.

As described above, in addition to the inter mode and the skip mode, a merge mode may be used for encoding an image.

10 shows an example of merge candidates.

As shown in FIG. 10, the merge mode may be used to encode one of motion information of prediction units present around a current prediction unit. Accordingly, index information indicating which motion prediction information of the neighboring prediction unit is used in the current prediction unit may be further encoded. Merge candidates of A to E located around the current prediction unit represent motion information of spatially neighboring prediction units, and T may mean a merge candidate of a position corresponding to the current block in another frame in time.

In addition, the SKIP mode of HEVC means Merge Skip. In SKIP, although the motion vector of the neighboring prediction unit is the same as in the Merge mode, there is a difference in that the residual signal of the current coding block is not transmitted. That is, Skip may be a mode that expresses information of a coding unit using only a SKIP flag and a merge index.

11 illustrates a header structure of a coding unit (CU) and a prediction unit (PU). In CU, SKIP flag can be used to distinguish whether CU is SKIP or not. If the SKIP flag is 1 in the CU, only the Merge Index is encoded / decoded in the header of the PU without further encoding / decoding other Herder information of the CU.

If the current CU is not SKIP and is an Inter CU, the PU can put a merge flag to distinguish whether or not the current PU is in merge mode. If the current PU is in Merge mode, the Merge Index is decoded / decoded. If not, the motion information is decoded / decoded.

In this case, if the ultra high resolution image and the low resolution image are divided into blocks of the same size, the correlation between the blocks of the ultra high resolution image may be higher than that of the low resolution image.

12 illustrates an image segmentation difference according to resolution.

As shown in FIG. 12, even if the same area is divided into blocks of the same size, there is a problem in that the ultra-high resolution image is divided using more blocks than the low resolution image.

In other words, in a super high resolution image, PUs constituting neighboring CUs are encoded in a similar mode. That is, in a super high resolution image, when one PU is selected as the merge mode, the probability of neighboring PUs also being determined as the merge mode is higher than that of the low resolution image. In such a situation, using an existing encoding method that repeatedly transmits a Merge Flag that informs Merge information for each PU may reduce coding efficiency.

Therefore, in the present invention, for a large coding unit unit (LCU), a LCU is divided into a plurality of areas according to the image characteristics inside the LCU, and then merges a plurality of PUs located in each area. Instead of transmitting information repeatedly, we propose an encoding / decoding method that can replace Merge flags repeatedly transmitted for each PU header by transmitting only information that corresponding PUs are included in a merge region.

Hereinafter, an encoding method according to an embodiment of the present invention will be described.

(A) How to divide the internal area of LCU

Each LCU does not repeatedly transmit merge information of neighboring PUs included in the LCU, but instead encodes information indicating that the PUs are included in the merge region. That is, after grouping a plurality of PUs that are spatially neighbored and encoded in a merge mode into a set, only information indicating that the PUs are located in a merge region is encoded.

FIG. 13 illustrates a method of dividing CUs included in an LCU into several sets. In FIG. 13, when classifying the CUs inside the LCU, a straight line is used to divide the straight line, which always passes through the center point of the LCU. This straight line has eight slopes passing through the center point of the LCU, and the specific slope type is shown in FIG. 13.

As shown in FIG. 13, LCUs are divided by selecting up to two types of straight lines among a total of eight types of slopes, and thus LCUs can be divided into up to four areas. If the LCU does not need to be partitioned, No Partition, which is 0 in FIG. 13, is selected without partitioning. The leftmost figure in the lower row of FIG. 13 illustrates an example in which the LCU is divided into four regions using the first slope dividing line and the fourth slope dividing line. The middle figure of the lower row of FIG. 13 is an example of dividing the LCU into two regions using only the fourth slope dividing line. The right figure of the lower row of FIG. 13 is an example of considering the LCU as one area without dividing.

When one LCU is divided into several regions as shown in FIG. 13, each region may be classified into one of a normal region and a merge region. All CUs in the region classified as the normal region are encoded using the general encoding method as it is. CUs in an area classified as a merge area may be Inter CUs or Intra CUs, but in the case of Inter CUs, the PUs inside the Inter CUs are encoded only in the merge mode. Since the Inter CUs in the merge region include a plurality of PUs encoded in merge mode, encoding efficiency is improved by transmitting information indicating that these PUs belong to the merge area, instead of repeatedly transmitting merge flags of the PUs.

As described above, in the case of ultra-high resolution images, since neighboring blocks have a high probability of being data having similar properties with each other, PUs encoded in merge mode may occur continuously at neighboring positions. In this case, by not transmitting a merge flag Encoding efficiency can be improved.

(B) Method of encoding internal region segmentation information in LCU

After dividing the LCU as shown in FIG. 13, information about the divisional shape should be encoded. This section describes a method of encoding information about this division shape.

As can be seen from FIG. 13 or FIG. 14, the split information of each LCU is represented by two split straight lines. These split straight lines are called the first split straight line and the second split straight line. Since these two split straight lines are in every LCU, split straight line information of a specific LCU can be predictively coded using split straight lines of the previous LCU.

The index of the first split straight line of the current LCU may be one of 0 to 8, the index of FIG. 13. According to an embodiment of the present invention, the encoder does not encode the index itself as it is, but uses the partition information of the previous LCU to make prediction partition information for the first partition information of the current LCU, and then the actual partition information and the prediction partition. The difference information between the information can be encoded.

Specifically, the encoder uses the index of the first split straight line of the previous LCU as a prediction index for encoding the first split straight line of the current LCU. If the index of the first split straight line of the current LCU is equal to the index of the first split straight line of the previous LCU, MPP_flag (1) = 1 is indicated. Otherwise, if the index of the first split straight line of the current LCU is different from the index of the first split straight line of the previous LCU, MPP_flag (1) = 0 and encode information on the first split straight line of the current LCU. . At this time, if the index of the first split straight line of the previous LCU is greater than the index of the first split straight line of the current LCU, the first split straight line index of the current LCU is represented by a 3-bit binary number. If the index of the first split straight line is smaller than the index of the first split straight line of the current LCU, the first split straight line index-1 of the current LCU is represented by a 3-bit binary number. This process is also used for the second segmented straight-line coding of the LCU.

15 shows a method of encoding split information of a current LCU.

The method described above may be represented as shown in FIG. 15. As illustrated in FIG. 15, at least 1 bit and at most 4 bits are required to encode an index of one division straight line. Therefore, in order to encode both the first and second partition information indicating the partition information of one LCU, a minimum of 2 bits and a maximum of 8 bits may be required.

(C) Method of Merge Information Coding

16 shows an example of a CU encoding order in a divided region.

Each area of the LCU divided into up to four can be distinguished whether the area is a merge area or a normal area by using area_wide_merge_flag, which is 1 bit.

As shown in FIG. 16, when the corresponding area is the merge area, area_wide_merge_flag is set to 1, and when the corresponding area is normal, area_wide_merge_flag is set to 0. The merge region means that all the PUs of the CU determined as the inter prediction mode are encoded in the merge mode. In this case, an intra CU may also exist in the merge region. All CUs existing in the normal region can be encoded in one mode among existing modes such as SKIP / Inter / Intra and PU can also be encoded in all available modes. Area_wide_merge_flag that distinguishes whether the corresponding area is a merge area is encoded by the number of areas where the LCU is divided.

The CU scan order within the LCU can use the scan order promised by the encoder and the decoder. 16 illustrates an example of an encoding sequence of a CU that can be used when applying the illustrated method according to an embodiment of the present invention, and is an exemplary method of determining an encoding sequence based on a raster scan. The present invention is not limited to the CU scan order inside the LCU.

As shown in FIG. 17, all the PUs in the CU determined as the inter prediction mode in the area determined as the merge area are encoded in the merge mode.

18 illustrates a difference between Merge Flag transmissions in a normal region and a merge region.

In particular, FIG. 18A illustrates a method of transmitting a merge flag. In addition, FIG. 18B describes a method of transmitting a flag in a normal region. In this area, a merge flag is transmitted for each PU.

18B illustrates a method of transmitting a merge flag in a merge region. As can be seen from this figure, the merge mode is encoded in the merge mode only by the fact that it belongs to the merge region even if the merge flag is not transmitted separately. In FIG. 18B, the Merge Index does not indicate whether a merge mode is used, but may be a syntax representing which Motion Vector Predictor is merged when the merge mode is encoded.

When the Merge flag is expressed in the above manner, when a PU encoded in merge mode occurs continuously, the information can be transmitted by transmitting only the information that the CU and the PU are included in the merge area, without transmitting the merge flag each time. Encoding efficiency is improved.

(D) LCU Encoding Method

As mentioned earlier, one LCU can be split from one to up to four. When dividing an LCU into two, there are eight branches that can be divided. In addition, if one LCU is divided into four, the number of branches that can be divided is 8 X 7 = 72 branches. How to determine the best partition type among the various partition type forms that are possible is not limited in the context of the present invention. In general, among division types that can be divided, a division type that generates the lowest rate distortion cost may be selected. In addition, it is possible to determine the optimal partitioning form in various ways, and the present invention does not limit the method of determining the optimal partitioning form of the LCU.

After determining the type of partitioning the LCU and the number of partitions, the partition type is encoded using the method of encoding the LCU internal region partition information described above. This process allows the current LCU to be divided into several areas and how it is divided.

In the next step, area_wide_merge_flag is transmitted as many as the number of partitions. In this case, if the corresponding area is the merge area, area_wide_merge_flag is transmitted to 1, and if the corresponding area is the normal area, area_wide_merge_flag is transmitted to 0.

After transmitting the area_wide_merge_flags like this, the CUs in the LCU are encoded using the method described in <Figure 10>.

The CU located at the boundary between the partitioned area and the partitioned area sets a partitioned area including more parts of the current CU as an area to which the current CU belongs. If a CU contains the same area in two areas and the characteristics of the two areas are different, the CU is set to the normal area.

(2) Decoding method

As illustrated in FIG. 19, in the decoding process of the LCU unit, information about the first and second division straight lines should be decoded in order to decode the information about the current LCU division type.

To do this, first check the MPP_flag for the first split straight line, and if the value is 1, it is interpreted that the first split straight line of the current LCU is the same as the first split straight line of the immediately decoded LCU. Otherwise, if the MPP_flag is 0, it is interpreted that the first split line of the current LCU is different from the first split line of the previous LCU.

If interpreted this way, an additional 3 bits are parsed to obtain a temporary value for the index of the first split straight line of the current LCU. If this temporary value is smaller than the first partitioning index of the previous LCU, this temporary value is used as the first partitioning straight line information of the current LCU. However, if this temporary value is equal to or greater than the first partition information index of the previous LCU, the value obtained by adding + 1 to this temporary value is decoded as the first partition information of the current LCU. This process is repeated when decoding the second split straight line of the current LCU. Through this process, it is possible to know in which partition type the current LCU is divided.

Since the number of partitions can be known through the above process, parsing and decoding area_wide_merge_flag as many as this number. If the area_wide_merge_flag decoded according to each area is 1, the modes of all the PUs constituting the CU determined as the inter-screen prediction mode in the area are merge, and if 0, the CUs determined as the inter-screen prediction mode of the area. The mode can be decoded into all interpretable prediction modes. This information is used to decode CUs inside the current LCU.

The method according to the present invention described above may be stored in a computer-readable recording medium that is produced as a program for execution on a computer, and examples of the computer-readable recording medium include ROM, RAM, CD-ROM, magnetic tape , Floppy disks, optical data storage devices, and the like, and also include those implemented in the form of carrier waves (eg, transmission over the Internet).

The computer readable recording medium can be distributed over network coupled computer systems so that the computer readable code is stored and executed in a distributed fashion. In addition, functional programs, codes, and code segments for implementing the method can be easily inferred by programmers in the art to which the present invention belongs.

In addition, although the preferred embodiment of the present invention has been shown and described above, the present invention is not limited to the specific embodiments described above, but the technical field to which the invention belongs without departing from the spirit of the invention claimed in the claims. Of course, various modifications can be made by those skilled in the art, and these modifications should not be individually understood from the technical spirit or prospect of the present invention.

Claims

In the video decoding method,

Decoding the segmentation information from the received code;

Determining a partition type of a large coding unit (LCU) according to the partition information;

Parsing an area flag from the code according to the partition type of the LCU; And

Determining whether to merge motion information for coding units included in the LCU based on the region flag, and decoding the coding units.
The method of claim 1,

The split type of the LCU is determined according to at least one split straight line.
The method of claim 1,

The splitting information includes index information of a split straight line of a previous coding unit.
The method of claim 3,

The decoding step

Decoding the split straight line of the LCU by using the index information as a prediction index for the split straight line of the LCU;

Video decoding method.
The method of claim 1,

The LCU is divided into a plurality of areas according to a dividing line having a predetermined slope.
The method of claim 1,

If the LCU is divided into a plurality of regions, each region is either a normal region or a merge region.
The method of claim 6,

The decoding step

Identifying a merge region in the LCU based on the region flag, and merging motion information of the merge region;

Video decoding method.
In the video decoding apparatus,

A decoder which decodes partition information from the received code;

A determination unit to determine a partition type of a large coding unit (LCU) according to the partition information; And

A parsing unit for parsing an area flag from the code according to the division type of the LCU;

And the decoder determines whether motion information for coding units included in the LCU is merged based on the region flag, and decodes the coding units.
The method of claim 8,

The splitting type of the LCU is determined according to at least one split straight line.
The method of claim 8,

And the splitting information includes index information of a split straight line of a previous coding unit.
The method of claim 10,

And the decoder to decode the segmented straight line of the LCU using the index information as a prediction index for the segmented straight line of the LCU.
The method of claim 8,

The LCU is divided into a plurality of areas according to a dividing line having a predetermined slope.
The method of claim 8,

When the LCU is divided into a plurality of regions, each region is one of a normal region and a merge region.
The method of claim 13,

And the decoding unit identifies a merge region in the LCU based on the region flag and merges motion information of the merge region.