Patent Grant
11856153
Implementations of the present disclosure relate to an intra prediction technology in the video encoding field, and more particularly, to an intra prediction method and device and a computer storage medium.
In a luma prediction process for the next generation video coding standard H.266 or versatile video coding (VVC), in order to reduce the number of bits for entropy coding, a most probable modes (MPM) list will be constructed to store prediction modes of neighbouring blocks. Based on the principle that spatially neighbouring blocks have a high similarity, there is a high probability that the prediction mode selected by the current block will be the same as a certain mode existing in the MPM list. Therefore, the prediction mode of the current block can be encoded with fewer bits. However, due to the existence of wide angular modes of non-square blocks, an actual angle direction represented by an angular mode index may be different from the original meaning, which may cause that the angular modes with the same number in the MPM list represent different prediction directions for the neighbouring blocks and the current block, and there are many types of situations, so the accurate expression and use of the prediction mode of the current block are affected. Moreover, in a chroma prediction process, a direct mode (DM) will borrow a prediction mode of a luma block which is located at a center position of the current chroma block. Regardless of whether the prediction mode of the luma block is a wide angular mode or not, the DM will borrow the original angular mode index, which may cause a deviation between the actually used angular mode of the chroma block and an angular mode of its corresponding luma block. That is to say, in the wide angular mode, the angular mode index may correspond to different actual angular modes, such that angle conversion is more complicated in the luma prediction process, and there may be a deviation in the angular mode of the luma block borrowed in the chroma prediction, resulting in the problem of inaccurate prediction.
Implementations of the present disclosure provide an intra prediction method and device, and a computer-readable storage medium, so as to effectively improve the accuracy of intra prediction while improving the efficiency of coding and decoding.
Technical schemes of the implementations of the present disclosure are implemented as follows.
An implementation of the present disclosure provides an intra prediction method including:
In the above scheme, when the sampling is performed at 65 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 66′, the actual angular modes corresponding to the relative angle numbers are 65 consecutive actual angular modes in a range from −14 to 80, and selection of the 65 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one in sequence.
In the above scheme, when the sampling is performed at 33 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 34′, the actual angular modes corresponding to the relative angle numbers are 33 consecutive actual angular modes in a range from −7 to 41, and selection of the 33 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers corresponding to the actual angular modes one to one in sequence.
In the above implementation, when the sampling is performed at 129 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 130′, the actual angular modes corresponding to the relative angle numbers are 129 consecutive actual angular modes in a range from −28 to 158, and selection of the 129 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one in sequence.
An implementation of the present disclosure further provides an intra prediction method including:
In the above scheme, determining the actual angular modes corresponding to the reference blocks represented by the relative angle numbers according to the width-height relationships, the prediction direction ranges corresponding to the preset width-height relationships and the preset angle sampling points includes:
In the above scheme, performing the intra prediction for the current block based on the angular prediction modes includes:
In the above scheme, the intra prediction includes at least one of luma intra prediction and chroma intra prediction.
An implementation of the present disclosure provides an intra prediction device including:
An implementation of the present disclosure provides a computer-readable storage medium having intra prediction instructions stored therein, wherein the intra prediction instructions, when executed by a processor, implement the intra prediction methods described above.
In the implementations of the present disclosure, by using the above technical implementation scheme, the intra prediction device can use a unified approach for the actual angular modes to process the reference blocks with different width-height relationships in the intra prediction process, to cause the actual angles to correspond to the actual angular modes one to one, such that when a certain angle is represented, the angular modes of blocks of each shape can be specifically determined according to length to width ratios whether in a luma prediction process or in a chroma prediction process, to simplify angle conversion in wide angular modes, unify an angle value meaning represented by each mode, and eliminate a deviation, thereby improving the accuracy of the intra prediction effectively while improving the efficiency of coding and decoding.
Technical schemes in implementations of the present disclosure will be clearly and completely described below in conjunction with the drawings in the implementations of the present disclosure. It may be understood that the specific implementations described herein are only intended to explain the relevant application, not limit the present disclosure. In addition, it should also be noted that for convenience of description, only the parts related to the relevant application are shown in the drawings.
Designations used in the present disclosure will be explained first below.
VVC/H.266: next generation video coding standard.
VTM: test model of VVC, which is a reference software test platform of VVC.
MPM: most probable mode.
DM: direct mode, which is a chroma prediction mode.
VER: vertical direction angular prediction mode, which corresponds to an intra prediction mode index 50 in VTM3.0.
HOR: horizontal direction angular prediction mode, which corresponds to an intra prediction mode index 18 in VTM3.0.
DIA: diagonal direction angular prediction mode, which corresponds to an intra prediction mode index 34 in VTM3.0.
VDIA: inverse diagonal direction angular prediction mode, which corresponds to an intra prediction mode index 66 in VTM3.0.
In the implementation of the present disclosure, a function of prediction coding is to construct a predictive value of the current block by using the existing reconstructed picture in space or time during video encoding, and only transmit a difference value between an original value and the predictive value, to achieve the purpose of reducing the amount of transmitted data. In luma prediction, the original value and predictive value herein may be an original value and predictive value of luma. In chroma prediction, the original value and predictive value herein may be an original value and predictive value of chroma.
A function of intra prediction is to construct a predictive value of the current block using sample units in the above row and sample units in the left column which are neighbouring to the current block. Each sample unit of the current block is predicted using the restored neighbouring samples around the current block (i.e., the sample units in the above neighbouring row and the sample units in the left neighbouring column of the current block).
For example, the current block is a luma block. When a luma predictive value of the current block is constructed using neighboring samples, luma prediction for the current block is carried out using a plurality of prediction directions in turn to obtain a luma predictive value matrix corresponding to each of the prediction directions. A difference matrix corresponding to each of the prediction directions is determined based on each luma predictive value matrix and a luma original value matrix of the current block. Evaluation parameter values corresponding to the prediction directions are determined based on each difference matrix, and the evaluation parameter values are used to indicate the prediction effect of the corresponding prediction directions on the current block. A target prediction direction is determined from the plurality of prediction directions based on each of the evaluation parameter values. For example, on the premise of ensuring the video restoration quality, a prediction direction in which the minimum quantity of bits for picture coding can be obtained is determined as the target prediction direction. Then the target prediction direction is written into a bitstream.
Illustratively, 67 intra prediction directions, i.e., prediction modes, are supported by VVC, where the intra prediction directions with index numbers 2-66 are shown in
It should be noted that in order to adapt to the increasing demand for video resolution and express directions of video contents more precisely and accurately, 33 luma intra prediction angular modes defined in H.265/HEVC is expanded to 65 luma intra prediction angular modes in H.266/VVC. The newly added angular modes are indicated by dashed arrows in
In the implementation of the present disclosure, an intra prediction direction with an index number 66 it taken as an example to propose a method of constructing a luma predictive value of each of the sample units of the current block. Data in the above neighbouring row of the current block are sample units that have been predicted. Each of the sample units of the current block is filled according to the sample unit with the top-right diagonal (i.e., the prediction direction with the index number 66).
In addition, there are two ways to construct prediction blocks in a relative flat manner, the DC mode and the PLANAR mode, respectively. In the DC mode, the whole current block is filled with an average value of feature values (e.g., chroma values or luma values) in the previous row or the left column, while in the PLANAR mode, the current block is filled in a gradual manner.
Luma modes are predicted according to 0-66 directions in
On the basis of the known basic concepts described above, a video encoding system is provided.
An implementation of the present disclosure provides a video decoding system.
The implementation of the present disclosure mainly acts on the intra prediction unit 213 of the video encoding system 21 and the intra prediction unit 223 of the video decoding system 22. That is to say, if a better prediction effect can be obtained in the video encoding system 21 through the intra prediction method in accordance with the implementation of the present disclosure, then accordingly, the video decoding recovery quality can be improved at the decoding end as well.
Based on this, the technical schemes of the present disclosure will be further set forth in detail in conjunction with the drawings and implementations.
It should be noted that an intra prediction device in accordance with an implementation of the present disclosure may be either an encoder or a decoder, and the implementation of the present disclosure is not limited thereto.
An implementation of the present disclosure provides an intra prediction method, which may include:
The intra prediction method provided by the implementation of the present disclosure relates to a process of writing prediction numbers of prediction modes of the current block into a bitstream for encoding and decoding.
Wide angular modes will be described below by taking 65 angle intra prediction modes. In the wide angular modes, as shown in
There are 85 angular direction modes and DC and planar modes in VTM2.0.1, wherein 20 angle directions are beyond a range between −135 degrees and 45 degrees, that is, wide angles. The angle directions within a range from −135 degrees (mode 2) to 45 degrees (mode 66) in the clockwise direction is designed for square blocks and includes diagonal directions of all square blocks (modes 2, 34 and 66). However, for non-square blocks, their diagonal directions are not always covered. In addition, angle directions of the square blocks start from the bottom-left diagonal direction to the top-right diagonal direction, while angle directions of the non-square blocks do not.
As shown in
A unified wide angular mode proposed by the L0279 proposal is accepted in the latest H.266/VVC reference software VTM3.0. The proposal puts forward three improvements:
wherein W is a width of a block (coding block) and H is a length of a block (coding block).
It should be noted that the unified method proposed in the L0279 proposal has modified the number of traditional modes that need to be replaced with wide angular modes, such that the angle range after the expansion of the wide angles is just between the bottom-left diagonal direction and the top-right diagonal direction (for example, between 2 and 66), as shown in Table 1. At the same time, the method has also modified appropriately directions of the expanded wide angular modes and the traditional angular modes that need to be replaced, such that the diagonal directions of the current coding block in the case of various width to height ratios are included.
Here, when the width to height ratio (or height to width ratio, the same goes for the following) is 2, 6 modes need to be replaced; when the width to height ratio is 4, 10 modes need to be replaced; when the width to height ratio is 8, 12 modes need to be replaced; and when the width to height ratio is 16, 14 modes need to be replaced.
That is to say, as shown in
In the implementation of the present disclosure, the actual angular modes represented by the relative angle numbers are used when the intra prediction device configures or identifies the angular modes; wherein the relative angle numbers successively represents the corresponding actual angular modes after sampling is performed at the preset angle sampling points starting from the starting angles within the prediction direction ranges corresponding to the preset width-height relationships; and the starting angles are determined according to the width-height relationships of the processed blocks and the prediction direction ranges corresponding to the preset width-height relationships.
Specifically, the intra prediction device can determine the starting angles of the angular modes of the reference blocks according to the width-height relationships of the processed blocks and the prediction direction ranges corresponding to the preset width-height relationships; determine angle offset ranges of the reference blocks based on the preset angle sampling points; and determine the actual angular modes corresponding to the processed blocks represented by the relative angle numbers according to the starting angles and the angle offset ranges.
It should be noted that in the implementation of the present disclosure, the width-height relationships and the prediction direction ranges determine the starting angles. The prediction direction ranges indicate what the angular modes of the preset angle sampling points selected from all angular modes are. Prediction direction range is known in the prior art. Referring to Table 1, for example, when the width to height ratio is 2, the prediction direction range is the 65 angular modes within a range from 8-72.
In some implementations of the present disclosure, when the sampling is performed at 65 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 66′, the actual angular modes corresponding to the relative angle numbers are 65 consecutive actual angular modes in a range from −14 to 80, and selection of the 65 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one.
In the implementation of the present disclosure, 2′-66′ is used to represent the actual angular modes corresponding to the relative angles represented by numbers of 2-66. That is to say, the actual angular modes represented by consecutive numbers in the range from 2′-66′ are 65 actual angular modes corresponding to 65 actual angles in an angle range from [a starting value angle+a lower limit value of an angle offset range, a starting value angle+an upper limit value of the angle offset range].
In the implementation of the present disclosure, the angle offset range is [0−the number of the preset angle sampling points−1]. For example, when t the number of the preset angle sampling points is 65, the angle offset range is [0-64].
Illustratively, in the implementation of the present disclosure, taking the number of the preset angle sampling points, i.e., angle directions, being 65 as an example, the preset width-height relationships, meanings of starting angular modes represented by 2′, the relative angle numbers, and meanings representing the actual angular modes are shown in Table 2.
It can be understood that in such a representation method, absolute numbers of all angular modes are represented in a range [−14, 80] with reference to the representation method for the actual angular modes in the L0279 proposal. No matter what width to height ratio the current coding block has, only 65 consecutive angular modes, i.e., 65 angle numbers, are contained. The selection range (that is, the prediction direction range) of the 65 angle numbers will be different due to the width to height ratios. In the representation method of the implementation of the present disclosure, the modes from the bottom-left diagonal direction to the top-right diagonal direction are fixed to be 2′-66′ (relative angle numbers) regardless of whether there is wide angular mode expansion. However, due to different width to height ratios, the meanings of the actual angular modes represented by the starting angle numbers of the relative angle numbers will be different. Although ranges of the relative angle numbers fall into different intervals due to differences in the prediction direction ranges, they all belong to the range [−14,80]. For example, when W/H=2, a relative angle number in the bottom-left diagonal direction is 2′, which actually represents the first value of a starting angle+an angle offset range, that is, an actual angular mode of 8+0 (the angle numbers in
The representation method of the implementation of the present disclosure can avoid a deviation between an angle direction of an MPM list storage mode and an angle direction of an actual neighbouring block due to the fact that both of the angle directions are represented by the same angular mode index, in a luma prediction process. Meanwhile, it avoids a deviation between a luma direction borrowed at the center position of the current chroma block and an actual luma direction in a chroma prediction process.
It should be noted that in the representation method using the relative angle numbers provided in the implementation of the present disclosure, the relative angle numbers are put into a bitstream for transmission in a form of syntax elements during coding, and a decoder appoints the meanings of the relative angle numbers with different width to height ratios during decoding, so the actual angular modes corresponding to the relative angle numbers can be analyzed through the received relative angle numbers.
In some implementations of the present disclosure, when the sampling is performed at 33 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 34′, the actual angular modes corresponding to the relative angle numbers are 33 consecutive actual angular modes in a range from −7 to 41, and selection of the 33 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one.
Illustratively, in the implementation of the present disclosure, taking the number of the preset angle sampling points, i.e., angle directions, being 33 as an example, the preset width-height relationships, meanings of starting angular modes represented by 2′, the relative angle numbers, and meanings representing the actual angular modes are shown in Table 3.
In some implementations of the present disclosure, when the sampling is performed at 129 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 130′, the actual angular modes corresponding to the relative angle numbers are 129 consecutive actual angular modes in a range from −28 to 158, and selection of the 129 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one.
Illustratively, in the implementation of the present disclosure, taking the number of the preset angle sampling points, i.e., angle directions, being 129 as an example, the preset width-height relationships, meanings of starting angular modes represented by 2′, the relative angle numbers, and meanings representing the actual angular modes are shown in Table 4.
In the implementation of the present disclosure, the number of the preset angle sampling points is not limited.
An implementation of the present disclosure provides an intra prediction method, which, as shown in
In S101, width-height relationships of reference blocks of a current block, prediction direction ranges corresponding to preset width-height relationships and preset angle sampling points are acquired.
In S102, actual angular modes corresponding to the reference blocks represented by relative angle numbers are determined according to the width-height relationships, the prediction direction ranges corresponding to the preset width-height relationships and the preset angle sampling points, to cause actual angles to correspond to the actual angular modes one to one.
In S103, angular prediction modes corresponding to the reference blocks are obtained based on the actual angular modes corresponding to the reference blocks.
In S104, intra prediction for the current block is performed based on the angular prediction modes.
In the implementation of the present disclosure, the reference blocks are data blocks within a preset range where the current block is located and for which the intra prediction is completed, and the reference blocks may be at least one.
It should be noted that the intra prediction device performs the intra prediction, which includes at least one of luma intra prediction and chroma intra prediction. The luma intra prediction uses neighbouring blocks as the reference blocks, and the chroma intra prediction can borrow luma blocks at the center positions of the previous chroma blocks as the reference block. Chroma prediction modes can include DM, LM, LM_T, LM_L and the like.
In some implementations of the present disclosure, the prediction modes include at least one of luma intra prediction directions and chroma intra prediction directions.
It can be understood that when the prediction directions are the luma intra prediction directions, the prediction directions of the reference blocks are luma directions, thus, when the intra prediction is performed for the current block in S104, the intra prediction is performed for the luma of the current block actually. Similarly, when the prediction directions are chroma intra prediction directions, the prediction directions of the reference blocks are chroma directions, thus, when the intra prediction is performed for the current block in S104, the intra prediction is performed for the chroma of the current block actually.
In S101, no matter whether the intra prediction device performs the luma intra prediction or the chroma intra prediction, the intra prediction device can obtain the width-height relationships of the reference blocks of the current block, the prediction direction ranges corresponding to the preset width-height relationships and the preset angle sampling points in the process of obtaining the prediction modes of the current block.
In the implementation of the present disclosure, the width-height relationships of the reference blocks may be width to height ratios or height to width ratios, and the implementation of the present disclosure is not limited thereto.
In the implementation of the present disclosure, the prediction direction ranges corresponding to the preset width-height relationships are the actual angle ranges including the number of consecutive preset angle sampling points between the bottom-left diagonal direction and the top-right diagonal direction, which are obtained for different width-height relationships of a processed block.
In the implementation of the present disclosure, the number of the preset angle sampling points is the number of sampling points between the bottom-left diagonal direction and the top-right diagonal direction.
It should be noted that the intra prediction device can obtain the width-height relationships of the reference blocks of the current block and the prediction direction ranges corresponding to the preset width-height relationships, determine the preset angle sampling points in each of the prediction direction ranges, and select the relative angle numbers between the prediction directions and starting angles.
Illustratively, 65 angle intra prediction modes are taken as an example. The prediction directions of the 65 angle intra prediction modes are defined to be between −135 degrees (mode 2) to 45 degrees (mode 66) in the clockwise direction, where [2′, 66] here are the relative angle numbers.
In the implementation of the present disclosure, in the case that the number of the preset angle sampling points is certain, lengths of the prediction direction ranges of all angular modes are the same. No matter what width-height relationship the current block has, only a number, which is the number of the consecutive preset angle sampling points, of angles (i.e., the relative angle numbers) are contained, except that a number, which is the number of the preset angle sampling points, of actual angular modes selected from the length of the prediction direction ranges will be different due to the width-height relationships.
In S102, after the intra prediction device obtains the width-height relationships of the reference blocks of the current block, the prediction direction ranges corresponding to the preset width-height relationships and the preset angle sampling points, the intra prediction device can determine the actual angular modes corresponding to the reference blocks represented by relative angle numbers according to the width-height relationships, the prediction direction ranges corresponding to the preset width-height relationships and the preset angle sampling points, to cause the actual angles to correspond to the actual angular modes one to one. S102 is implemented in S1021-1023 specifically below.
In S1021, the starting angles of the angular modes of the reference blocks are determined according to the width-height relationships and the prediction direction ranges corresponding to the preset width-height relationships.
In S1022, angle offset ranges of the reference blocks are determined based on the preset angle sampling points.
In S1023, the actual angular modes corresponding to the reference blocks represented by the relative angle numbers are determined according to the starting angles and the angle offset ranges.
After obtaining the width-height relationships and the prediction direction ranges corresponding to the preset width-height relationships, the intra prediction device can determine the starting angles of the angular modes of the reference blocks according to the width-height relationships and the prediction direction ranges corresponding to the preset width-height relationships, and determine the angle offset ranges of the reference blocks based on the preset angle sampling points; and determine the actual angular modes corresponding to the reference blocks represented by the relative angle numbers according to the starting angles and the angle offset ranges, such that the intra prediction device can select the actual angular modes within the angle offset ranges with the relative angle numbers starting from the starting angles.
In some implementations of the present disclosure, the relative angle numbers can be used to represent angles, thus, the intra prediction device can determine relative angle starting numbers of the starting angles of the angular modes of the reference blocks according to the width-height relationships and the prediction direction ranges corresponding to the preset width-height relationships; determine the angle offset ranges of the reference blocks to be [0−the number of the preset angle sampling points−1] based on the preset angle sampling points; and determine the actual angular modes corresponding to the reference blocks represented successively by the consecutive numbers in the range of relative angle numbers [2′-66] according to the starting numbers and the angle offset ranges.
Illustratively, in the implementation of the present disclosure, taking the number of the preset angle of sampling points, i.e., angle directions, being 65 as an example, the preset width-height relationships, meanings of starting angular modes represented by 2′, the relative angle numbers, and meanings representing the actual angular modes are shown in Table 2.
It can be understood that in such a representation method, absolute numbers of all angular modes are represented in a range [−14, 80] with reference to the representation method for the actual angular modes in
It should be noted that in the representation method using the relative angle numbers provided in the implementation of the present disclosure, the relative angle numbers are put into a bitstream for transmission in a form of syntax elements during coding, and a decoder appoints the meanings of the relative angle numbers with different width to height ratios during decoding, so the actual angular modes corresponding to the relative angle numbers can be analyzed through the received relative angle numbers.
It can be understood that the relative angle numbers are transmitted as the syntax elements of the angular mode indices in the bitstream, thus the acquisition of the real prediction angles by a codec in the wide angular modes is simplified, which is conducive to improvement of the efficiency of encoding and decoding.
In some implementations of the present disclosure, when the sampling is performed at 33 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 34′, the actual angular modes corresponding to the relative angle numbers are 33 consecutive actual angular modes in a range from −7 to 41, and selection of the 33 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one in sequence.
Illustratively, in the implementation of the present disclosure, taking the number of the preset angle sampling points, i.e., angle directions, being 33 as an example, the preset width-height relationships, meanings of starting angular modes represented by 2′, the relative angle numbers, and meanings representing the actual angular modes are shown in Table 3.
In some implementations of the present disclosure, when the sampling is performed at 129 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 130′, the actual angular modes corresponding to the relative angle numbers are 129 consecutive actual angular modes in a range from −28 to 158, and selection of the 129 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one in sequence.
Illustratively, in the implementation of the present disclosure, taking the number of the preset angle sampling points, i.e., angle directions, being 129 as an example, the preset width-height relationships, meanings of starting angular modes represented by 2′, the relative angle numbers, and meanings representing the actual angular modes are shown in Table 4.
In the implementation of the present disclosure, the number of the preset angle sampling points is not limited.
In S103, the intra prediction device obtains the angular prediction modes corresponding to the reference blocks based on the actual angular modes corresponding to the reference blocks.
After obtaining the actual angular modes corresponding to the reference blocks, the intra prediction device can use the actual angular modes to represent the angular prediction modes (e.g., dirA, dirB) corresponding to the reference blocks.
That is to say, in the implementation of the present disclosure, the angular prediction modes corresponding to the reference blocks obtained by the intra prediction device represent a one-to-one correspondence between the actual angles and the actual angular modes.
In S104, the intra prediction device may perform the intra prediction for the current block based on the angular prediction modes. Here, the intra prediction device needs to construct a prediction mode list of the current block based on the angular prediction modes; and then implement the intra prediction for the current block using the prediction mode list.
In the implementation of the present disclosure, for the luma intra prediction, the prediction mode list is an MPM list.
For the chroma intra prediction, the prediction mode list is a DM list or a MDMS list.
For example, in the luma intra prediction, as shown in
Candidate prediction directions of the MPM list are divided into three groups: neighbouring prediction modes, derived prediction modes and default prediction modes. First, the neighbouring prediction modes are added to the MPM list. Each intra prediction mode can only be added to the MPM list once, that is, the MPM list cannot contain repeated prediction modes. If the number of the prediction modes contained in the MPM list after the adding of the neighbouring prediction modes is completed is less than 6, the derived intra prediction modes are added to the MPM list. If the number of the prediction modes contained in the MPM list after the adding of the neighbouring prediction modes is completed is still less than 6, the default intra prediction modes are added to the MPM list until the MPM list containing six most possible intra prediction modes is derived.
When entropy coding is performed for the intra prediction modes of each luma block, the MPM list of the luma block is obtained firstly, and it is determined whether a selected intra prediction mode of the luma block is in the MPM list. If the selected intra prediction mode of the luma block is in the MPM list, a truncated binary code is used to binarize an index number of the prediction mode in the MPM list. The smaller the index number is, the smaller the resulting truncated binary code is. Then the truncated binary code is encoded by an arithmetic encoder, thus bit overhead can be saved. If the selected intra prediction mode of the luma block is one of the remaining 61 prediction modes that are not in the MPM list, the 61 prediction modes are re-numbered from 0, and 16 prediction modes whose numbers can be exactly divisible by 4 are used as the selected modes. If the intra prediction mode is in the selected modes, bypass coding for it is performed using a fixed 4-bit length. If the intra prediction mode is in the remaining 45 non-selected modes, it is re-numbered and binarized using the truncated binary code, to generate a bit string with a length of 5 or 6 bits according to the number size, and then the bypass coding for it is performed.
Because a 6MPM list in JEM is complicated, a scheme using a simplified 3MPM list is proposed later. However, since the 3MPM list contains fewer prediction modes, and the resulting prediction effects are not accurate enough, a scheme using a simplified 6MPM list (which is also a method used in VTM3.0 at present) is proposed. For example, new candidate prediction modes of the current block are constructed based on prediction modes corresponding to block A (upper) and prediction modes corresponding to block L (left) in
A reference line index used by the current block is 0.
MPM={dirL, Planar/DC, HOR 18th, VER 50th, VER-4, VER+4}; Planar corresponds to 0, DC corresponds to 1, and there must be 6 modes in the MPM list. A neighbouring mode is plus or minus 1.
MPM={dirL, Planar/DC, dirL−1, dirL+1, dirL−2, dirL+2};
MPM={dirL, dirA, Planar/DC, max(dirL, dirA)−1, max(dirL, dirA)+1, max(dirL, dirA)−2};
MPM={dirL, dirA, Planar/DC, dirL−1, dirL+1, dirL−2};
MPM={dirL, dirA, HOR, VER, HOR−4, HOR+4};
When the reference line index is 1 or 3:
MPM={VER, HOR, 2, DIA, VDIA, 26};
MPM={dirL, dirA, min(dirL, dirA)−1, min(dirL, dirA)+1, max(dirL, dirA)−1, max(dirL, dirA)+1, . . . };
MPM={dir, dir−1, dir+1, dir−1, dir+2, dir−3}.
In the implementation of the present disclosure, related description of a method for constructing chroma intra prediction directions of DM and VVC draft 3 in the chroma intra prediction is as follows, as shown in Table 5:
Illustratively, as shown in
It can be determined, in conjunction with the contents shown in table 5 and
MDMS for chroma intra prediction is a more complex method of constructing chroma intra prediction directions. As shown in table 6, compared with DM, there is a code rate saving of 0.2%, but it has not been applied to VVC because of excessive complexity.
As shown in
That is to say, in the implementation of the present disclosure, the intra prediction device can perform both luma intra prediction and chroma intra prediction.
It can be understood that it is proposed in the present disclosure that the angular modes of all rectangular blocks represent the actual angular modes corresponding to the starting angles plus the angle offset ranges in turn using the relative angle numbers, and are unified an internal within [a starting value angle+a lower limit value of an angle offset range, a starting value angle+an upper limit value of the angle offset range]. In the representation method of the implementation of the present disclosure, different starting angles are set according to different reference block width-height relationships, old angular mode indices are used as the meaning representing new angular modes, and the angular modes correspond to the angles one to one, thus the angular mode indices of the rectangular blocks with various width-height relationships are unified, and the angle directions represented in the luma MPM list and chroma prediction DM are consistent with the actual angles, thereby simplifying angle conversion in the wide angular modes, unifying an angle value meaning represented by each mode, and eliminating deviation, which is conducive to improvement of the accuracy of intra prediction.
Further, the relative angle numbers are transmitted as the syntax elements of the angular mode indices in the bitstream, thus the acquisition of the real prediction angles by a codec in the wide angular modes is simplified, which is conducive to improvement of the efficiency of encoding and decoding.
Based on the implementation of the foregoing implementations, an implementation of the present disclosure provides an intra prediction device including:
In some implementations of the present disclosure, when the sampling is performed at 65 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 66′, the actual angular modes corresponding to the relative angle numbers are 65 consecutive actual angular modes in a range from −14 to 80, and selection of the 65 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one in sequence.
In some implementations of the present disclosure, when the sampling is performed at 33 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 34′, the actual angular modes corresponding to the relative angle numbers are 33 consecutive actual angular modes in a range from −7 to 41, and selection of the 33 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers corresponding to the actual angular modes one to one in sequence.
In some implementations of the present disclosure, when the sampling is performed at 129 preset angle sampling points, the relative angle numbers are consecutive numbers in a range from 2′ to 130′, the actual angular modes corresponding to the relative angle numbers are 129 consecutive actual angular modes in a range from −28 to 158, and selection of the 129 actual angular modes is determined by the width-height relationships, wherein the relative angle numbers correspond to the actual angular modes one to one in sequence.
On the basis of the implementation of the foregoing implementations, as shown in
The acquisition part 10 is further configured to obtain angular prediction modes corresponding to the reference blocks based on the actual angular modes corresponding to the reference blocks.
The intra prediction device 1 includes an intra prediction part 12 configured to perform intra prediction for the current block based on the angular prediction modes.
In some implementations of the present disclosure, the determining part 11 is specifically configured to determine the starting angles of the angular modes of the reference blocks according to the width-height relationships and the prediction direction ranges corresponding to the preset width-height relationships, and determine angle offset ranges of the reference blocks based on the preset angle sampling points; and determine the actual angular modes corresponding to the reference blocks represented by the relative angle numbers according to the starting angles and the angle offset ranges.
In some implementations of the present disclosure, the intra prediction part 12 is specifically configured to construct a prediction mode list of the current block based on the angular prediction modes; and implement the intra prediction for the current block using the prediction mode list.
In some implementations of the present disclosure, the intra prediction includes at least one of luma intra prediction and chroma intra prediction.
As shown in
In the implementations of the present disclosure, the processor 13 described above may be at least one of an application specific integrated circuit (ASIC), a digital signal processor (DSP), a digital signal processing device (DSPD), a programmable logic device (PLD), a field programmable gate array (FPGA), a central processing unit (CPU), a controller, a microcontroller and a microprocessor. It can be understood that for different devices, electronic devices for implementing the above processor functions may also be other devices, which are not limited specifically in the implementations of the present disclosure. The intra prediction device may further include a memory 14, which may be connected to the processor 13, wherein the memory 14 is used for storing executable program codes including computer operation instructions. The memory 14 may be a volatile memory, such as a random-access memory (RAM); or a non-volatile memory, such as a read-only memory (ROM), a flash memory, a hard disk drive (HDD) or a solid-state drive (SSD); or a combination of the above types of memories, and provides instructions and data to the processor 13.
In the implementations of the present disclosure, the communication bus 15 is used for connecting the processor 13 and the memory 14 and intercommunication between these devices.
In addition, various functional modules in the implementations may be integrated into one processing unit, or various units may be physically presented separately, or two or more than two units may be integrated into one unit. The integrated units may be implemented in a form of hardware, or may be implemented in a form of a software functional module.
The integrated units, if implemented in a form of a software functional module and not sold or used as an independent product, may be stored in a computer-readable storage medium. Based on such understanding, the technical scheme of the implementations, in essence, or the part contributing to the prior art, or all or part of the technical scheme, may be embodied in a form of a software product, which is stored in a storage medium, and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) or a processor to perform all or part of steps of the methods in accordance with the implementations. The aforementioned storage medium includes various media, such as a U disk, a mobile hard disk, a read-only memory, a random access memory, a magnetic disk, or an optical disk, which are capable of storing program codes.
An implementation of the present disclosure provides a computer-readable storage medium having intra prediction instructions stored therein, wherein the intra prediction instructions, when executed by a processor, implement the intra prediction methods described above.
Specifically, the intra prediction instructions corresponding to the intra prediction method in the implementation can be stored on a storage medium such as a CD, a hard disk, a U disk, etc. When the intra prediction instructions corresponding to the intra prediction method in the storage medium is read or executed by an electronic device, the intra prediction method includes the following steps: acquiring width-height relationships of reference blocks of a current block, prediction direction ranges corresponding to preset width-height relationships and preset angle sampling points; determining actual angular modes corresponding to the reference blocks represented by relative angle numbers according to the width-height relationships, the prediction direction ranges corresponding to the preset width-height relationships and the preset angle sampling points, to cause actual angles to correspond to the actual angular modes one to one; obtaining angular prediction modes corresponding to the reference blocks based on the actual angular modes corresponding to the reference blocks; and performing intra prediction for the current block based on the angular prediction modes.
It can be understood that the intra prediction device can use a unified approach for the actual angular modes to process the reference blocks with different width-height relationships in the intra prediction process, to cause the actual angles to correspond to the actual angular modes one to one, such that when a certain angle is represented, the angular modes of blocks of each shape can be specifically determined according to length to width ratios whether in a luma prediction process or in a chroma prediction process, to simplify angle conversion in wide angular modes, unify an angle value meaning represented by each mode, and eliminate a deviation, thereby improving the accuracy of the intra prediction effectively while improving the efficiency of coding and decoding.
It should be understood by a person skilled in the art that the implementations of the present disclosure may be provided as methods, systems, or computer program products. Therefore, the present disclosure may use the form of a hardware implementation, a software implementation, or an implementation combining software and hardware. Moreover, the present disclosure may use the form of a computer program product implemented on one or more computer usable storage media (including, but not limited to, a magnetic disk memory, an optical memory, etc.) containing computer usable program codes.
The present disclosure is described with reference to implementation flowcharts and/or block diagrams of the methods, devices (systems), and computer program products in accordance with the implementations of the present disclosure. It should be understood that each flow and/or block in the flowcharts and/or the block diagrams, and combinations of flows and/or blocks in the flowcharts and/or the block diagrams may be implemented by computer program instructions. These computer program instructions may be provided to a processor of a general purpose computer, a special purpose computer, an embedded processing machine or other programmable data processing devices to generate a machine, such that instructions which are executed by a processor of a computer or another programmable data processing devices generate an apparatus for implementing functions specified in one or more flows in the implementation flowcharts and/or one or more blocks in the block diagrams.
These computer program instructions may also be stored in a computer-readable memory that can instruct a computer or other programmable data processing devices to operate in a particular manner, such that the instructions stored in the computer-readable memory generate an article of manufacture including an instruction apparatus, wherein the instruction apparatus implements functions specified in one or more flows in the implementation flowcharts and/or one or more blocks in the block diagrams.
These computer program instructions may also be loaded onto a computer or other programmable data processing devices to cause a series of operational steps to be performed on a computer or other programmable devices to generate computer-implemented processing, such that the instructions executed on the computer or the other programmable device provide steps for implementing functions specified in one or more flows in the implementation flowcharts and/or one or more blocks in the block diagrams.
What are described above are merely preferred implementations of the present disclosure and are not intended to limit the protection scope of the present disclosure.
The implementations of the present disclosure provide an intra prediction method and device, and a computer-readable storage medium. The intra prediction device can use a unified approach for the actual angular modes to process the reference blocks with different width-height relationships in the intra prediction process, to cause the actual angles to correspond to the actual angular modes one to one, such that when a certain angle is represented, the angular modes of blocks of each shape can be specifically determined according to length to width ratios whether in a luma prediction process or in a chroma prediction process, to simplify angle conversion in wide angular modes, unify an angle value meaning represented by each mode, and eliminate a deviation, thereby improving the accuracy of the intra prediction effectively while improving the efficiency of coding and decoding.
This application is a continuation application of U.S. application Ser. No. 17/355,032 filed on Jun. 22, 2021, which is a continuation application of International PCT Application No. PCT/CN2019/070153 filed on Jan. 2, 2019, the entire content of both are hereby incorporated by reference.
| Number | Name | Date | Kind |
|---|---|---|---|
| 11140387 | Lee | Oct 2021 | B2 |
| 20130287094 | Song et al. | Oct 2013 | A1 |
| 20200099925 | Lee | Mar 2020 | A1 |
| 20200099935 | Lee | Mar 2020 | A1 |
| 20200195950 | Zhao et al. | Jun 2020 | A1 |
| 20210006799 | Lee | Jan 2021 | A1 |
| 20210144402 | Lee | May 2021 | A1 |
| Number | Date | Country |
|---|---|---|
| 107113438 | Aug 2017 | CN |
| 108235008 | Jun 2018 | CN |
| 108605122 | Sep 2018 | CN |
| 108702515 | Oct 2018 | CN |
| 109040754 | Dec 2018 | CN |
| 3349448 | Jul 2018 | EP |
| 2018117894 | Jun 2018 | WO |
| 2018127624 | Jul 2018 | WO |
| 2019083284 | May 2019 | WO |
| 2019232023 | Dec 2019 | WO |
| Entry |
|---|
| Second Office Action for Chinese Application No. 202110708337.8 dated Nov. 11, 2022. 19 pages with English translation. |
| Summons to Oral Proceedings for European Application No. 19906725.7 issued Jan. 20, 2023. 5 pages. |
| International Search Report dated Sep. 26, 2019 of PCT/CN2019/070153 (4 pages). |
| Extended European Search Report for European Application No. 19906725.7 dated Nov. 25, 2021. 11 pages. |
| Bross, B. et al. “Versatile Video Coding (Draft 2)” JVET-K1001-v7; JVET of ITU-T SG 16 WP3 and ISO/IEC JTC 1/SC 29 WG 11; Ljubljana, SI; Jul. 10-18, 2018, 139 pages. |
| Rath, G. et al. “CE3-related: Block Shape Adaptive Intra Prediction Directions” JVET-K0169; JVET of ITU-T SG 16 WP3 and ISO/IEC JTC 1/SC 29 WG 11; Ljubljana, SI; Jul. 10-18, 2018. 6 pages. |
| Lainema, J. et al. “4.4 Intra Mode Coding” in High Efficiency Video Coding (HEVC); Sep. 3, 2014, Springer International Publishing; pp. 103-106. |
| Zhao, L. et al. “CE3-related: Unification of angular intra prediction for square and non-square blocks” JVET-L0279; JVET of ITU-T SG 16 WP3 and ISO/IEC JTC 1/SC 29 WG 11; Macao, CN; Oct. 3-12, 2018, 10 pages. |
| Examination Report for Indian Application No. 202117029572 dated Mar. 10, 2022. 6 pages with English translation. |
| EP Examination Report dated Jul. 13, 2022 In Application No. 19906725.7. |
| China First Office Action dated Jun. 23, 2022 In Application No. 202110708337.8, 20 pages. |
| Japan Office Action dated Jun. 24, 2022 In Application No. 2021-538683, 13 pages. |
| Seregin et al., Neighbor based intra most probable modes list derivation, Joint Video Exploration Team (JVET) of ITU-T SG 16 WP 3 and ISO/IEC JTC I/SC 29/WG 11, 3rd Meeting: Geneva, CH, May 26-Jun. 1, 2016, Document: JVET-C0055, 5 pages. |
| Rath et al., CE3-related: Block Shape Adaptive Intra Prediction Directions, Joint Video Experts Team (JVET) of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11, 11th Meeting: Ljubljana, SI, Jul. 10-18, 2018, Document: JVET-K0169_r1, 6 pages. |
| Zhao et al., CE3-related: Wide angular intra prediction for non-square blocks, Joint Video Exploration Team (JVET) of ITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11, 11th Meeting: Ljubljana, SI, Jul. 10-18, 2018, Document: JVET-K0289_v2, 4 pages. |
| Racape et al., CE3-related: Wide-angle intra prediction for non-square blocks, Joint Video Experts Team (JVET) ofITU-T SG 16 WP 3 and ISO/IEC JTC 1/SC 29/WG 11, 11th Meeting: Ljubljana, SI, Jul. 10-18, 2018, Document: JVET-K0500_r4, 13 pages. |
| Bross et al., Versatile Video Coding (Draft 2), Joint Video Experts Team (JVET), ofITU-T SG 16 WP 3 and 1S0/IEC JTC 1/SC 29/WG 11, 11th Meeting: Ljubljana, SI, Jul. 10-18, 2018, Document: NET-K1001-v6, 12 pages. |
| Non-final Office Action dated Mar. 16, 2023 in U.S. Appl. No. 17/715,390. |
| EP Brief Communication dated Jun. 2, 2023 in EP App. 19906725.7, 5 pages. |
| Non Final Rejection of U.S. Appl. No. 17/715,468 dated Jul. 28, 2023, 20 Pages. |
| Decision to Refuse of European Application No. 19906725.7 dated Jun. 28, 2023, 10 Pages. |
| First Office Action of Indonesian Application No. P00202105542 dated Aug. 29, 2023. 5 Pages with English Translation. |
| Notice of Allowability of U.S. Appl. No. 17/715,390 dated Sep. 13, 2023, 6 Pages. |
| First Office Action of the Japanese application No. 2022-184105, dated Sep. 22, 2023. 4 pages with English translation. |
| Number | Date | Country | |
|---|---|---|---|
| 20220232250 A1 | Jul 2022 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 17355032 | Jun 2021 | US |
| Child | 17715318 | US | |
| Parent | PCT/CN2019/070153 | Jan 2019 | US |
| Child | 17355032 | US |
