Method and apparatus for combined intra prediction modes

By using the intra prediction mode combination of adjacent blocks in the video encoder, the problems of low efficiency and high syntax cost of predicting curved objects in the prior art are solved, and more efficient video encoding and decoding are achieved.

CN112005552BActive Publication Date: 2024-12-03INTERDIGITAL VC HOLDINGS INC
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN201980027026.6
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Priority Date
2018-04-27
Filing Date
2019-04-03
Publication Date
2024-12-03
Estimated Expiration
2039-04-03

AI Technical Summary

Technical Problem

Existing video encoders are inefficient when predicting the boundaries of curved objects, require segmented predictions and transmit additional parameters, resulting in high syntax costs and unable to effectively utilize the advantages of smooth intra-predictors.

Method used

Use intra prediction mode combinations based on adjacent video blocks in an encoder or decoder to add to the most likely mode list to improve prediction efficiency and reduce syntax costs.

Benefits of technology

By combining prediction modes of adjacent blocks, the efficiency of intra prediction is improved and the syntax cost is reduced, especially for small blocks or low-texture areas, with little additional syntax elements required.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN112005552B_ABST
    Figure CN112005552B_ABST
Patent Text Reader

Abstract

An encoder and decoder for a digital video signal uses a combined intra-frame prediction mode for certain images. In at least one embodiment, the combined intra-frame prediction mode can be generated by the intra-frame prediction modes of adjacent blocks. The combined intra-frame prediction mode can be added to the most probable mode list. Various embodiments describe techniques for combining intra-frame prediction modes, including a combination of reference samples and a combination of directions to form a prediction. In another embodiment, if the two modes combined to form the combined mode are directional modes with an angle difference of less than 90 degrees, the combined mode is added to the most probable mode list. Another embodiment uses a linear combination of intra-frame prediction modes, and another embodiment performs a linear combination that depends on the distance of the prediction from the blocks to the left and above.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present principles relate to video compression as well as video encoding and decoding. Background Art

[0002] The field of the following embodiments is video coding, focusing on a robust approach to intra-frame encoding and decoding with more diverse prediction modes than current encoders and decoders. For example, to predict the border of a curved object, existing encoders would have to select small blocks along the edge of the object to predict it in a piecewise manner.

[0003] JVET (Joint Video Exploration Team) is exploring an improved video coding standard called H.266, Versatile Video Coding (VVC), or H.266 / VCC. In JVET or H.266 / VCC, filters have been proposed to smooth intra predictors. For example, position-dependent prediction combinations use smoothed and unfiltered reference samples to combine predictions. This enables filtering of predictions that originally corresponded to directional padding of possibly smoothed reference samples. However, this approach has limitations and may require the transmission of additional parameters, reducing its advantages.

[0004] The described embodiments propose to use such intra prediction using a combined mode in an encoder or a decoder. Summary of the invention

[0005] These and other drawbacks and deficiencies of the prior art are addressed by the present principles, which are directed to methods and apparatus for combined intra-prediction modes.

[0006] According to an aspect of the present principles, a method is provided. The method comprises the steps of predicting a current video block based on a combination of intra-prediction modes from video blocks adjacent to the current video block; and encoding the current video block using the prediction, wherein the combination of intra-prediction modes is added to a most probable mode list used during encoding.

[0007] According to another aspect of the present principles, a second method is provided. The method comprises the steps of: predicting a current video block based on a combination of intra-prediction modes from video blocks adjacent to the current video block; and decoding the current video block using the prediction, wherein the combination of intra-prediction modes is added to a most probable mode list used during decoding.

[0008] According to another aspect of the present principles, there is provided an apparatus. The apparatus comprises a memory and a processor. The processor may be configured to encode or decode a block of a video image by predicting a current video block based on a combination of intra-prediction modes from video blocks adjacent to the current video block; and, encode / decode the current video block using the prediction, wherein the combination of intra-prediction modes is added to a most probable mode list used during encoding / decoding.

[0009] These and other aspects, features and advantages of the present principles will become apparent from the following detailed description of exemplary embodiments, which is to be read in conjunction with the accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A standard generic video compression scheme is shown to which the described embodiments may be applied.

[0011] Figure 2 Example intra modes from neighboring video blocks are shown.

[0012] Figure 3 One embodiment of adding a mode to a Most Probable Mode (MPM) list is shown.

[0013] Figure 4 The construction of the most likely pattern list in JEM (Joint Exploration Model) is shown.

[0014] Figure 5 An example of segmenting an intra-coded frame using a candidate codec for H.266 is shown.

[0015] Figure 6 An example of a current block whose neighboring blocks are encoded using a directional prediction mode is shown.

[0016] Figure 7 The intra mode predictor is shown using (a) the mode of the left block, (b) the mode of the above block, (c) the direction corresponding to the average of left and above, (d) the proposed position-dependent combination.

[0017] Figure 8 An example of the interpolation direction on the current block to be predicted is shown.

[0018] Fig. 9 An example of angular mode interpolation based on a sub-block grid is shown.

[0019] Fig.10 One embodiment of a method using the described aspects is shown.

[0020] Fig.11Another embodiment of a method using the described aspects is shown.

[0021] Fig.12 One embodiment of an apparatus utilizing the described aspects is shown. DETAILED DESCRIPTION

[0022] The present invention is incorporated into the framework of video compression, in particular at the intra prediction stage. Existing hybrid codecs such as H.264 / MPEG4-AVC and H.265 / HEVC use intra prediction, which aims to encode the current block by first building a prediction using previously reconstructed neighboring samples in the same image, and then encoding the difference between the prediction and the source (also called the residual). In the exemplary reference Joint Exploration Model (JEM) being studied by the Joint Video Exploration Team (JVET), the encoder can choose from 67 intra prediction modes.

[0023] - Planar mode creates a smooth linear combination of the reference samples over the current block.

[0024] -DC mode assigns the average value of the reference samples to all samples in the current block.

[0025] - The remaining 65 directional modes propagate the reference samples to the current block along a specific direction.

[0026] The syntax element encoding the mode index may then represent too high a cost. The Most Probable Mode (MPM) approach enables the cost of this information to be reduced by providing a more lightweight syntax for statistically common modes and reusing the mode selected for the reconstructed neighboring blocks. If an MPM is selected, the index is transmitted in a short list of MPMs. Several neighboring blocks are then considered and an optimal mode is selected.

[0027] In one embodiment described herein, when at least two neighboring blocks are available, it is proposed to add an intra mode to the MPM list, which is a combination of the prediction modes of these neighboring blocks. For example, when two neighboring blocks are encoded with different directional modes, it is proposed to create a hybrid mode that is a linear mapping of the direction on the surface of the current block.

[0028] The problem addressed by these embodiments is to improve the efficiency of intra prediction of a block whose causal neighboring blocks bring supplementary information by adding an extra mode that is a combination of those neighboring prediction modes.

[0029] The present invention aims to implement new filling of reference samples onto the current block to establish intra-frame prediction. The existing directional mode is to propagate reference samples in a fixed direction.

[0030] Then, for example, to predict the boundary of a curved object, an existing encoder would have to select small patches along the edge of the object to predict it in a piecewise manner.

[0031] In JVET, filters have been proposed to smooth the intra predictors. For example, position-dependent prediction combinations use smoothed and unfiltered reference samples to combine predictions. This enables filtering of predictions that originally correspond to directional padding of possibly smoothed reference samples. However, the algorithm does not combine two modes that can be considered on the same surface.

[0032] In another approach, parameter modifications to the rectilinear mode are proposed to better align the prediction with the content of the current block. However, additional parameters must be transmitted to signal the prediction process to the decoder. For small or low-texture blocks, the residuals can be compacted and compressed well by transform and entropy coding. However, the syntax to encode the prediction mode is non-negligible and these additional parameters become cost-prohibitive. In the proposed solution, almost no additional syntax is needed.

[0033] To better predict various spatial directions of objects and textures, at least one of the described embodiments proposes adding additional modes in the context of the Most Probable Mode (MPM) framework. The new modes may be signaled as additional modes to the list of most probable modes used during encoding.

[0034] When two reconstructed neighboring blocks are available and use different prediction modes, one embodiment uses a model that combines the two prediction direction modes on the current block surface and takes into account the different ways of encoding the neighboring blocks.

[0035] As a first example, but not limited to this case, consider blocks to the left and top of the current block that are encoded using different directional modes. Then, depending on the neighboring directions and the positions of the neighboring blocks considered, a model is selected to make the apparent direction of the current block vary.

[0036] Adding an additional mode to the list of 67 modes that already exist would represent a prohibitive syntax cost. Advantageously, this mode corresponds to an additional mode that can be added only if there are neighboring modes. It can be added as an additional MPM without extending the complete list of intra modes and their corresponding syntax.

[0037] The description of at least one proposed solution is based on an exemplary current version of reference software designed by the Joint Video Exploration Team (JVET) to design the next MPEG / ITU standard for video coding, currently known as H.266. However, the present invention is not limited to this potential version of H.266 and can be applied to any future video standard that includes block-by-block intra prediction.

[0038] In existing codecs, intra prediction is constructed using extremely basic operations. The efficiency of the process lies in the multiple modes that the encoder can choose. However, encoding the syntax elements of this decision can become costly. For example, in the JVET exploratory software, 67 intra modes are considered to predict each block of the luminance component.

[0039] - Planar mode creates a smooth linear combination of the reference samples over the current block.

[0040] -DC mode assigns the average value of the reference samples to all samples in the current block.

[0041] - The remaining 65 directional modes propagate the reference samples to the current block along a specific direction.

[0042] Therefore, the Most Probable Mode (MPM) has been introduced to predict the intra prediction mode from already decoded information.

[0043] In JEM, the encoding of the intra mode index is performed by an improved version of the most probable mode of HEVC. The current method is based on another algorithm. For the current block, 5 neighboring positions are considered, i.e. Figure 2 The left side (L), top (A), lower left (BL), upper right (AR), and upper left (AL) are shown in FIG.

[0044] The modes to be added are checked in the following order:

[0045] -L, A, Plane, DC, BL, AR, AL.

[0046] - For each entry in the current list: if the pattern is directional, add pattern -1, then add pattern +1 (i.e., add the neighboring block's pattern first. If there is space in the list, add the pattern with indices to the left and right of the neighboring block's pattern, since neighboring blocks can share similar structures with slightly different angles).

[0047] -Added remaining default modes: VER, HOR, 2, DIA.

[0048] If there are modes to be added (dirMode > 0 and dirMode < 65) and they are not already included in the list, they are added to the current last position (pushed back), as in Figure 3 As shown in . First, 110 checks the existence of the mode to be added. If the neighboring block does not exist or is not intra-coded, the list remains unchanged. Otherwise, module 120 checks whether the mode is already included in the current list. If not, module 130 includes it at the last position in the list, otherwise, the list remains unchanged.

[0049] Figure 3The process described in corresponds to Figure 4 In box 150, Figure 4 Describes the overall process of constructing the MPM list. Figure 4 As described in , potential modes are added until the MPM list is full (160).

[0050] If the best mode belongs to this list, the syntax for encoding is very small. The encoder only needs to send a flag to inform the decoder that the mode belongs to the MPM list. Then, the MPM index encoded with the truncated unary code is very lightweight. Especially for small blocks, and in general for blocks with few residuals, i.e., when the syntax cost is not negligible compared to the transform coefficients, this encourages the encoder to choose those modes because their cost in terms of syntax is low.

[0051] In existing codecs, intra prediction aims to construct a first version of the current block, called the predictor, which minimizes the energy of the residual signal to be transmitted in the bitstream. The residual is usually transformed in the frequency domain and then quantized and encoded. Therefore, the goal is not to provide a visually pleasing predictor, but a construction using available samples that optimizes the cost of transmitting data. For example, in the case where the block is located at an object boundary or on a curved texture, the encoder has the following options:

[0052] - Use best match direction mode.

[0053] - Use DC or plane.

[0054] - If possible, split the block so that smaller sub-blocks are better predicted, at the expense of more syntax elements.

[0055] For example, a curved boundary of an object will cause the encoder to select small patches along the edge so that straight-line directional patterns can approximate the shape. Textures and edges are mostly encoded with small patches, except for some straight-line edges. For example, in Figure 5 In , the edges of the character are encoded with small patches because they cannot be predicted efficiently using existing patterns. It can be noticed that on the jacket, several larger patches appear because the straight line pattern is well predicted using the straight line direction pattern. However, this rarely happens in natural videos.

[0056] From the above, it can be asserted that curved edges and textures are not easily predicted using existing intra modes. Adding a "curve" fill mode will expand the number of intra modes, thereby expanding the syntax, and expanding the combinations on the encoder and decoder side, resulting in higher complexity. On the other hand, MPM provides a way to combine adjacent modes with a reduced syntax. According to an embodiment, the MPM list can contain a combined mode, or two modes contained in the MPM can be combined.

[0057] Then, it is proposed to add an intra mode corresponding to a combination of modes of adjacent blocks when the modes of adjacent blocks are different.

[0058] Take a block whose left and top neighboring blocks are predicted using different directional modes, such as Figure 6 as shown in .

[0059] Figure 7 Depicted with Figure 6 An exemplary propagation related to the intra-mode direction is shown in FIG. Figure 7 (a) and Figure 7 (b) corresponds to the intra predictor when using the modes selected for the left and above blocks, respectively. Figure 7 (c) shows what happens if we use a directional pattern that corresponds to the average of the patterns of the left and upper blocks. Figure 7 (d) shows an exemplary result of the proposed combination. In this example, the combination (model) is calculated as:

[0060]

[0061] Among them, θ(x,y) represents the angle of the direction of each pixel position, x and y represent the horizontal and vertical coordinates respectively, θ L and θ A represent the left and up directional modes respectively. Figure 7 In , when the directional mode fills points between reference samples and the value of the closest reference sample is selected, no smoothing filter is used and no interpolation is performed. The sample value corresponds to the resulting angle applied to the pixel.

[0062] For the current codec, the encoder can choose one of them with low syntax cost. Then, several embodiments described below include different ways to signal the combination. For example, a syntax element can describe how to combine two existing modes with low syntax cost.

[0063] In some embodiments, different models are described in which linear combinations of actual sample values ​​are considered.

[0064] In a first embodiment for encoding an image block, an intra prediction process is used that propagates neighboring samples, and:

[0065] - Check if at least two of the previously coded / decoded blocks can be combined to create at least one new propagation of reference samples

[0066] - Derive the intra predictor for the current block from the reference samples and the new mode

[0067] - Encodes syntax elements that describe which neighboring blocks' intra modes are used and how.

[0068] In a second embodiment, a decoder for decoding an image block uses an intra prediction process that propagates neighboring samples, and:

[0069] - Decodes syntax elements that describe which neighboring blocks' intra modes are used and how.

[0070] - Derive the intra predictor for the current block from the reference samples and the described mode.

[0071] During intra prediction, the embodiment checks in the adjacent left mode and above mode. If the modes are different, at least one mode is added to the list corresponding to the combination of the two modes.

[0072] In at least a third embodiment for encoding or decoding an image block, embodiment 1 or 2, respectively, is used, wherein each new mode is a combination of modes included in the MPM list.The required syntax element then describes the MPM list index of the used mode.

[0073] In at least a fourth embodiment, and where no additional syntax is required, new combined patterns are added to the MPM list. During the MPM list construction process, adjacent patterns are checked and patterns are added to the list when a combination is possible.

[0074] In at least a fifth embodiment, the fourth embodiment can be implemented, wherein only the blocks to the left and above (see Figure 2 ) The new mode can be included anywhere in the MPM list, such as after the left and above modes or after plane and DC, the only restriction is that the encoder and decoder share the same construction process.

[0075] In at least the sixth embodiment, the five preceding embodiments may further include a process of adding a new mode depending on the neighboring block intra modes considered, for example, when the mode is directional and the angle difference is less than 90°.

[0076] In at least the seventh embodiment, the preceding six embodiments may further include at least one new mode, which is a linear combination of sample values ​​when predicted using different modes.

[0077] For each p(i,j) in the current block, (i,j) is the coordinate within the block, (i=0,j=0) is at the top left sample point,

[0078] p(i,j)=α.p L (i,j)+(1-α)p A (i,j),

[0079] Among them, pL (i,j) and p A (i,j) corresponds to the prediction of position (i,j) using the left and above modes, respectively.

[0080] In at least an eighth embodiment, the fifth embodiment may be used, wherein the combination is linear and depends on the horizontal or vertical distance of the corresponding block to the left or above.

[0081]

[0082] In at least the ninth embodiment, the fifth embodiment is used, where the combination is in terms of direction rather than pixel value. In this embodiment, the directions of the patterns are combined. For each pixel of the current block, the combination direction is determined, rather than the combination of reference samples. For example, the angular direction can be linearly interpolated from the upper right to the lower left position. Then, for each position in the current block, there is a single interpolation direction.

[0083]

[0084] In at least a tenth embodiment, the eighth embodiment may be implemented wherein the change of direction is applied at a sub-block level, such as Fig. 9 As shown in . Each subblock is classically predicted using a given directional pattern. The center of each subblock can be considered to calculate its angular pattern. In HEVC and JVET exploration software, 4×4 subblocks are considered for inter prediction. Under the described embodiments, subblocks of any size can also be considered for intra prediction.

[0085] exist Fig.10 One embodiment of the described aspects is illustrated in , which shows a method 1000 for encoding a video block using a combination of intra-prediction modes. The method begins at start block 1001, and control proceeds to block 1010 for predicting a current video block based on a combination of intra-prediction modes from video blocks neighboring the current video block. Control proceeds from block 1010 to block 1020 for encoding the current video block using the prediction, wherein the combination of intra-prediction modes is added to a most probable mode list used during encoding.

[0086] exist Fig.11 Another embodiment of the described aspects is illustrated in , which shows a method 1100 for decoding a video block using a combined intra prediction mode. The method begins at start block 1101, and control proceeds to block 1110 for predicting a current video block based on a combination of intra prediction modes from video blocks neighboring the current video block. Control proceeds from block 1110 to block 1120 for decoding the current video block using the prediction, wherein the combination of intra prediction modes is added to a most probable mode list used during decoding.

[0087] exist Fig.12 One embodiment of the described aspects is illustrated in FIG. 1 , which shows an apparatus 1200 for encoding or decoding a video block using a combined intra-frame prediction mode. The apparatus includes a processor 1210 and a memory 1220. The processor 1210 is configured to encode to perform Fig.10 step, that is, encoding is performed using a combined intra prediction mode.

[0088] When the processor 1210 is configured to decode, it performs Fig.11 step, i.e., decoding is performed using the combined intra prediction mode.

[0089] The functions of the various elements shown in the drawings may be provided using dedicated hardware and hardware capable of executing software in association with appropriate software. When provided by a processor, the functions may be provided by a single dedicated processor, by a single shared processor, or by multiple individual processors, some of which may be shared. In addition, the explicit use of the term "processor" or "controller" should not be interpreted as referring exclusively to hardware capable of running software, and may implicitly include, but is not limited to, digital signal processor ("DSP") hardware, read-only memory ("ROM") for storing software, random access memory ("RAM"), and non-volatile memory.

[0090] Other hardware, conventional and / or custom, may also be included. Similarly, any switches shown in the figure are conceptual only. Their functions may be implemented by the operation of program logic, by dedicated logic, by the interaction of program control and dedicated logic, or even manually, and the specific technique may be selected by the implementer, as more specifically understood from the context.

[0091] This description illustrates the present principles. Thus, it will be appreciated that those skilled in the art will be able to devise various arrangements that, although not explicitly described or shown herein, embody the principles of the invention and are included within its spirit and scope.

[0092] All examples and conditional language recited herein are intended for teaching purposes to help readers understand the principles of the invention and the concepts contributed by the inventors to advance the art, and are to be construed as not being limited to such specifically recited examples and conditions.

[0093] In addition, all statements describing the principles, aspects and embodiments of the present principles and specific examples thereof are intended to cover their structural and functional equivalents. In addition, such equivalents are intended to include currently known equivalents and equivalents developed in the future, i.e., any element that performs the same function, regardless of the structure.

[0094] Thus, for example, those skilled in the art will appreciate that the block diagrams presented herein represent conceptual views of illustrative circuits embodying the present principles. Similarly, it will be appreciated that any flow charts, flow diagrams, state transition diagrams, pseudo-code, etc. represent various processes that can be substantially represented in a computer-readable medium and executed by a computer or processor, whether or not such a computer or processor is explicitly shown.

[0095] In the claims herein, any element represented as a means for performing a particular function is intended to encompass any manner of performing that function, including, for example, a) a combination of circuit elements that perform that function, or b) any form of software, thus including firmware, microcode, etc., in combination with appropriate circuitry for running the software to perform that function. The present principles defined by such claims reside in the fact that the functions provided by the various recited components are combined and brought together in the manner required by the claims. Therefore, any components capable of providing those functions are considered equivalent to those shown herein.

[0096] References in the specification to "one embodiment" or "an embodiment" of the present principles and other variations thereof mean that a particular feature, structure, characteristic, etc. described in conjunction with the embodiment is included in at least one embodiment of the present principles. Therefore, the phrases "in one embodiment" or "in an embodiment" and any other variations appearing in various places throughout the specification are not necessarily all referring to the same embodiment.

Claims

1. A method, include: predicting a current video block based on a combination of intra prediction modes from video blocks neighboring the current video block; as well as encoding the current video block using the prediction, wherein the combination of intra prediction modes is added to a list of most probable modes used during encoding, wherein the combination of intra prediction modes is added to the list of most probable modes when a combination of the intra prediction modes from video blocks neighboring the current video block is possible as neighboring modes of intra prediction modes of video blocks neighboring the current video block are checked and a new combination mode exists, and syntax elements indicate which neighboring blocks' intra prediction modes are used by video blocks neighboring the current video block and how the intra prediction modes are used, wherein the modes are added to the list of most probable modes for a subset of possible combinations; The subset includes patterns from blocks to the left and above the current video block.

2. The method according to claim 1, in, If modes in the combination of intra-prediction modes of video blocks neighboring the current block are different, the combination of intra-prediction modes is added to the most probable mode list.

3. The method according to claim 1, in, The syntax element specifies use of a most probable mode list, and wherein an index of a most probable mode corresponds to a combination of intra prediction modes from video blocks neighboring the current video block.

4. The method according to claim 1, in, Combinations of intra prediction modes are additional modes in the most probable mode list. 5 . The method of claim 1 , if a combined mode is based on a directional intra mode with an angle difference less than 90 degrees, adding the combined mode to the most probable mode list.

6. The method according to claim 1, in, The combination is a linear combination of the sample values ​​when predicted using different modes.

7. The method according to claim 1, in, The combinations are formed based on the directions.

8. A device for encoding a block of image data, include: Memory, and The processor is configured as: predicting a current video block based on a combination of intra prediction modes from video blocks neighboring the current video block; as well as encoding the current video block using the prediction, wherein the combination of intra prediction modes is added to a list of most probable modes used during encoding, wherein the combination of intra prediction modes is added to the list of most probable modes when a combination of the intra prediction modes from video blocks neighboring the current video block is possible as neighboring modes of intra prediction modes of video blocks neighboring the current video block are checked and a new combination mode exists, and syntax elements indicate which neighboring blocks' intra prediction modes are used by video blocks neighboring the current video block and how the intra prediction modes are used, wherein the modes are added to the list of most probable modes for a subset of possible combinations; The subset includes patterns from blocks to the left and above the current video block.

9. The device according to claim 8, in, If modes in the combination of intra-prediction modes of video blocks neighboring the current block are different, the combination of intra-prediction modes is added to the most probable mode list.

10. The device according to claim 8, in, The syntax element specifies use of a most probable mode list, and wherein an index of a most probable mode corresponds to a combination of intra prediction modes from video blocks neighboring the current video block.

11. The device according to claim 8, in, Combinations of intra prediction modes are additional modes in the most probable mode list. 12 . The apparatus of claim 8 , further comprising adding a combined mode to a most probable mode list if the combined mode is based on a directional intra mode with an angle difference less than 90 degrees.

13. The device according to claim 8, in, The combination is a linear combination of the sample values ​​when predicted using different modes.

14. The device according to claim 8, in, The combinations are formed based on the directions.

15. A method, include: predicting a current video block based on a combination of intra prediction modes from video blocks neighboring the current video block; as well as decoding the current video block using the prediction, wherein the combination of intra prediction modes is added to a list of most probable modes used during decoding, wherein the combination of intra prediction modes is added to the list of most probable modes when a combination of the intra prediction modes from video blocks neighboring the current video block is possible as neighboring modes of intra prediction modes of video blocks neighboring the current video block are checked and a new combination mode exists, and syntax elements indicate which neighboring blocks' intra prediction modes are used by video blocks neighboring the current video block and how the intra prediction modes are used, wherein the modes are added to the list of most probable modes for a subset of possible combinations; The subset includes patterns from blocks to the left and above the current video block.

16. The method according to claim 15, in, If modes in the combination of intra-prediction modes of video blocks neighboring the current block are different, the combination of intra-prediction modes is added to the most probable mode list.

17. The method according to claim 15, in, The syntax element specifies use of a most probable mode list, and wherein an index of a most probable mode corresponds to a combination of intra prediction modes from video blocks neighboring the current video block.

18. The method according to claim 15, in, Combinations of intra prediction modes are additional modes in the most probable mode list.

19. The method of claim 15, adding a combined mode to a most probable mode list if the combined mode is based on a directional intra mode with an angle difference less than 90 degrees.

20. The method according to claim 15, in, The combination is a linear combination of the sample values ​​when predicted using different modes.

21. The method according to claim 15, in, The combinations are formed based on the directions.

22. A device for decoding a block of image data, include: Memory, and The processor is configured as: predicting a current video block based on a combination of intra prediction modes from video blocks neighboring the current video block; as well as decoding the current video block using the prediction, wherein the combination of intra prediction modes is added to a list of most probable modes used during decoding, wherein the combination of intra prediction modes is added to the list of most probable modes when a combination of the intra prediction modes from video blocks neighboring the current video block is possible as neighboring modes of intra prediction modes of video blocks neighboring the current video block are checked and a new combination mode exists, and syntax elements indicate which neighboring blocks' intra prediction modes are used by video blocks neighboring the current video block and how the intra prediction modes are used, wherein the modes are added to the list of most probable modes for a subset of possible combinations; The subset includes patterns from blocks to the left and above the current video block.

23. The device according to claim 22, in, If modes in the combination of intra-prediction modes of video blocks neighboring the current block are different, the combination of intra-prediction modes is added to the most probable mode list.

24. The device according to claim 22, in, The syntax element specifies use of a most probable mode list, and wherein an index of a most probable mode corresponds to a combination of intra prediction modes from video blocks neighboring the current video block.

25. The device according to claim 22, in, Combinations of intra prediction modes are additional modes in the most probable mode list.

26. The apparatus of claim 22, further comprising adding a combined mode to a most probable mode list if the combined mode is based on a directional intra mode with an angular difference less than 90 degrees.

27. The device according to claim 22, in, The combination is a linear combination of the sample values ​​when predicted using different modes.

28. The device according to claim 22, in, The combinations are formed based on the directions.

29. A non-transitory computer readable medium comprising computer program instructions, in, When the computer program instructions are executed by a processor, the processor performs the method according to any one of claims 1 to 7.

30. A computer program product comprising instructions which, when executed by a computer, cause the computer to perform the method of any one of claims 15 to 21.