WO2012126340A1

WO2012126340A1 - Method and device for determining weight factor, and method and device for intra-frame weighted prediction

Info

Publication number: WO2012126340A1
Application number: PCT/CN2012/072518
Authority: WO
Inventors: 杨海涛; 周建同
Original assignee: 华为技术有限公司
Priority date: 2011-03-20
Filing date: 2012-03-19
Publication date: 2012-09-27
Also published as: CN102695061A; CN102695061B

Abstract

Disclosed are a method and device for determining a weight factor, and a method and device for intra-frame weighted prediction, comprising: determining, according to a difference between an abscissa component and an ordinate component of a current pixel in a currently processed image block, a second value; and determining, according to a first value and the second value, two weight factors of the current pixel, one weight factor being a sum of the first value and the second value, the other weight factor being a difference between the first value and the second value, and a sum of the two weight factors being double of the first value. The first value is determined according to an image block size of the currently processed image block or is preset. The processing speed of a codec is improved, and the processing power consumption thereof is reduced.

Description

Method and apparatus for determining a weighting factor,

And an intra-frame weighted prediction method and apparatus. The present application claims a method and apparatus for determining a weighting factor, and an intra-frame weighted prediction method, which is filed on March 20, 2011, with the application number of 201110067323.9. The priority of the Chinese Patent Application, the entire disclosure of which is incorporated herein by reference.

Technical field

The present invention relates to the field of communications technologies, and in particular, to a method, apparatus, and system for weighted prediction in video image intra prediction encoding and decoding.

BACKGROUND OF THE INVENTION Intra prediction refers to the use of surrounding reconstructed pixel values to predict current processed image block pixel values, and is one of the important modules in a video image codec. To ensure that the encoder and decoder get the same reconstructed video image, the encoder is identical to the intra prediction module in the decoder. Specifically, the module inputs, the algorithm flow within the module, and the module output are identical. Intra prediction techniques usually include two types of techniques: Directional Prediction and DC Prediction. When there is a directional texture in the video image content, the directional prediction technique along the direction can accurately predict the texture of the current processing block, and obtain a higher coding efficiency; when the video image content does not have a significant directional texture, Higher coding efficiency can be achieved using DC prediction techniques. The following uses the intra prediction technique in the High-efficiency Video Coding (HEVC) architecture as an example.

In the HEVC reference software HM, taking 4x4 image blocks as an example, there are 34 intra prediction modes, including one DC prediction mode and 33 directional prediction modes. An example of the prediction direction corresponding to the prediction mode in 33 is given in Fig. 1.

As shown in Figure 1, the two prediction directions VER+8 and HOR+8 fall within the same line, but the directions are opposite. Specifically, the VER+8 direction prediction mode uses the reconstructed pixel value above the current processing block to extrapolate the predicted value in the lower left 45 degrees; and the HOR+8 direction prediction mode uses the reconstructed pixel value on the left side of the current processing block along the upper right side. Extrapolation in the 45 degree direction to obtain the predicted value.

Obviously, the purpose of these two prediction modes is to handle texture content in the same direction. Therefore, current A pixel at any position in the processing block can find two reference pixels at the same time. In this case, the pixel values of the two reference points can be weighted averaged to obtain a predicted value of the currently processed pixel. For example, the reference point pixel values found in the two prediction directions VER+8 and HOR+8 of the point P(x, y) in Fig. 1 are Ref 1 and Ref2, respectively, and the P(x, y) positions are to Refl and Ref2. The distances of the reference points are d2 and d2, respectively, and the weighted average of Ref 1 and Ref2 can be calculated as the predicted value of P(x, y).

P(x,y)=(Refl · d2+Ref2 · dl)/(dl+d2)

The principle of the above weighted average calculation is that the importance of the reference pixel is inversely proportional to its distance from the currently processed pixel. This weighted prediction method combines the advantages of the prediction modes in the two directions of VER+8 and HOR+8, and the extrapolation operation in two independent prediction modes is changed to a linear interpolation operation, which can improve the prediction accuracy and thus improve the intra prediction. Coding efficiency.

A disadvantage of the above weighted prediction method is that the computational complexity is too high. The above calculation formula includes a division operation, and the divisor ( ^{dl + d2} ) may be any integer within a certain range. This division operation will bring a large hardware implementation cost in practical applications, and at the same time, it will reduce the codec processing speed and increase the power consumption of the codec device. SUMMARY OF THE INVENTION Embodiments of the present invention provide a weighted prediction method, apparatus, and system for intra-frame prediction encoding and decoding of video images, which improve processing speed of a codec and reduce processing power consumption.

An embodiment of the present invention provides a method for determining two weighting factors used in pairing, including: determining a second value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of a current pixel in a currently processed image block, where the horizontal coordinate component The coordinate system referenced by the ordinate component is the pixel point of the upper left corner of the currently processed image block as the coordinate origin, the horizontal axis direction is horizontally to the right, and the ordinate direction is vertically downward; the current pixel is determined according to the first value and the second value. Two weighting factors of the point, one of which is the sum of the first value and the second value, and the other weighting factor is the difference between the first value and the second value, and the sum of the two weighting factors is twice the first value The first value is determined or preset according to an image block size of the currently processed image block, where the image block size is a side length N of the NxN image block, and the NxN image block is a square including NxN pixel points. Image block.

An embodiment of the present invention provides an intra-frame weighted prediction method, including: And determining a second value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of the current pixel in the currently processed image block, where the coordinate system referenced by the horizontal coordinate component and the ordinate component is a coordinate point of the upper left corner of the currently processed image block. The origin, the horizontal axis direction is horizontally to the right, and the ordinate direction is vertically downward; determining two weighting factors of the current pixel point according to the first value and the second value, wherein one weighting factor is the first value and the second value And another weighting factor is a difference between the first value and the second value, the sum of the two weighting factors being twice the first value; wherein the first value is determined according to the image block size of the currently processed image block or is Setting, the image block size is an edge length N of the NxN image block, and the NxN image block is a square image block including NxN pixel points;

A weighted prediction value of the current pixel point is calculated based on the two weighting factors and the two predicted values, the predicted values being predicted values obtained in two specified intra prediction modes, respectively.

An embodiment of the present invention provides an intra-frame weighted prediction method based on a table lookup, including:

Determining an index value according to a difference between an abscissa component and a ordinate component of the current pixel in the currently processed image block, wherein the coordinate system referenced by the abscissa component and the ordinate component is a pixel origin of the upper left corner of the currently processed image block , the horizontal axis direction is horizontally to the right, and the ordinate direction is vertically downward; according to the cable I value, two weighting factors are obtained from the weighting factor table, where the two weighting factors are corresponding to the weighting factor table. Two weighting factors for the index value;

An embodiment of the present invention provides a determining apparatus for two weighting factors used in pairing, including: a first determining module, configured to determine, according to a difference between a horizontal coordinate component and a longitudinal coordinate component of a current pixel point in a currently processed image block, Numerical values, the coordinate system referenced by the abscissa component and the ordinate component is taken as the coordinate origin of the pixel in the upper left corner of the currently processed image block, the horizontal direction is horizontal to the right, and the ordinate direction is vertically downward;

a second determining module, configured to determine, according to the first value and the second value, two weighting factors of the current pixel, where one weighting factor is a sum of the first value and the second value, and another weighting factor is the first value a difference from the second value, the sum of the two weighting factors being twice the first value; wherein the first value is determined or preset according to an image block size of the currently processed image block, the image block size is NxN The side length N of the image block, the NxN image block is a square containing NxN pixel points Image block.

An embodiment of the present invention provides an intra-frame weighting prediction apparatus, including:

a third determining module, configured to determine a second value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of the current pixel in the currently processed image block, where the horizontal coordinate component and the coordinate component referenced by the ordinate component are currently processed images The pixel in the upper left corner of the block is taken as the coordinate origin, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward;

a fourth determining module, configured to determine, according to the first value and the second value, two weighting factors of the current pixel, where one weighting factor is a sum of the first value and the second value, and another weighting factor is the first value a difference from the second value, the sum of the two weighting factors being twice the first value; wherein the first value is determined or preset according to an image block size of the currently processed image block, the image block size is NxN An edge length N of the image block, the NxN image block being a square image block including NxN pixel points;

And a first prediction module, configured to calculate a weighted prediction value of the current pixel point according to the two weighting factors and the two prediction values, where the two prediction values are predicted values respectively obtained by using two specified intra prediction modes.

An embodiment of the present invention provides an intra-frame weighting prediction apparatus based on a look-up table, including:

a fifth determining module, configured to determine an index value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of the current pixel in the currently processed image block, where the horizontal coordinate component and the coordinate component referenced by the ordinate component are currently processed image blocks The pixel in the upper left corner is used as the coordinate origin, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward;

a sixth determining module, configured to obtain two weight factors from the weighting factor table according to the index value, where the two weighting factors are two weighting factors corresponding to the index value in the weighting factor table;

And a second prediction module, configured to calculate a weighted prediction value of the current pixel point according to the two weighting factors and the two prediction values, where the two prediction values are predicted values respectively obtained by using two specified intra prediction modes.

According to the technical solution provided by the embodiment of the present invention, the second value is determined according to the difference between the abscissa component and the ordinate component of the current pixel in the currently processed image block, according to the first value and the second value. Determining two weighting factors of the current pixel point, wherein one weighting factor is a sum of the first value and the second value, and another weighting factor is a difference between the first value and the second value, and the sum of the two weighting factors is A technical technique that doubles the value, thereby increasing the processing speed of the codec and reducing the processing power consumption. DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be described in detail below. It is obvious that the drawings in the following description are only some embodiments of the present invention. It will be apparent to those skilled in the art that other drawings may be obtained from these drawings without the inventive labor.

1 is a schematic diagram of a prediction direction in the prior art;

2 is a flow chart of a method for determining two weighting factors used in pairing according to an embodiment of the present invention;

FIG. 3 is a flowchart of an intra-frame weighting prediction method according to an embodiment of the present invention;

4 is a schematic diagram of two intra prediction modes according to an embodiment of the present invention;

5 is a flowchart of an intra-frame weighting prediction method based on table look-up according to an embodiment of the present invention; FIG. 6 is a flowchart of a method for constructing a weighting factor table according to an embodiment of the present invention;

FIG. 7 is a schematic diagram of a determining apparatus for two weighting factors used in pairing according to an embodiment of the present invention;

FIG. 8 is a schematic diagram of an intra-frame weighting prediction apparatus according to an embodiment of the present invention;

FIG. 9 is a schematic diagram of an intra-frame weighting prediction apparatus based on table lookup according to an embodiment of the present invention;

The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings in the embodiments of the present invention. It is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments obtained by a person of ordinary skill in the art based on the embodiments of the present invention without creative efforts are within the scope of the present invention.

As shown in FIG. 2, an embodiment of the present invention provides a method for determining two weighting factors used in pairing, including: S101: Determine a second value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of the current pixel in the currently processed image block, where the horizontal coordinate component and the coordinate coordinate referenced by the vertical coordinate component are the upper left pixel pixel of the currently processed image block. As the origin of the coordinate, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward;

S102: Determine, according to the first value and the second value, two weighting factors of the current pixel, where one weighting factor is a sum of the first value and the second value, and another weighting factor is a first value and a second value. Poor, the sum of the two weighting factors is twice the first value; wherein the first value is determined or preset according to the image block size of the currently processed image block, the image block size being the side length N of the NxN image block The NxN image block is a square image block including NxN pixels.

According to the technical solution provided by the embodiment of the present invention, the second value is determined according to the difference between the abscissa component and the ordinate component of the current pixel in the currently processed image block, and the current pixel is determined according to the first value and the second value. Two weighting factors, one of which is the sum of the first value and the second value, and the other weighting factor is the difference between the first value and the second value, and the sum of the two weighting factors is twice the first value Means, thereby improving the processing speed of the codec and reducing the processing power consumption.

A method for determining two weighting factors for pairing according to an embodiment of the present invention is further described below with reference to FIG. 2 as follows:

S101. Determine a second value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of a current pixel in the currently processed image block, where the horizontal coordinate component and the coordinate coordinate referenced by the ordinate component are the pixel points in the upper left corner of the currently processed image block. As the origin of the coordinates, the horizontal axis direction is horizontal to the right and the ordinate direction is vertically downward.

The second value in the embodiment of the present invention may be a position offset of the current pixel point position and the reference line. The current pixel position can be written as (x, y) where _X , y=0..BlkSize-l. The reference line refers to a line connecting the pixel positions of the current processing image block with the same abscissa component and the ordinate component, and can be recorded as #文(x, y) where x, y = O..BlkSize-l, x= y. The positional offset may be recorded as PosDev y], which may be determined by the abscissa component and the ordinate component of the current pixel point, and the determination method may be various. For example, when the first value is determined as the image block size BlkSize of the current processing block, the position offset may be determined according to the following method. PosDev[x, y] = x— y, where x, y = O..BlkSize-l

For another example, when the first value is determined to be a fixed value FixedBlkSize, the position offset can be determined according to the following method.

PosDev[x, y] = (FixedBlkSize - PosDev[x, y])/ BlkSize, where x, y = 0.. BlkSize- 1

Or equivalent, change the integer multiplication and division in the above equation to the left and right shift operations, PosDev[x, y] = (PosDev[x, y] « log ₂ FixedBlkSize) » log ₂ BlkSize, where x y = 0.. BlkSize-1 is visible. The method for determining the second value according to the difference between the abscissa component and the ordinate component of the current pixel in the currently processed image block may be to compare the abscissa component of the current pixel point with The difference of the ordinate component is used as the second value, or the difference is operated as the second value, and the operation of the difference may be scaling or taking an absolute value or the like.

Because log ₂ FixedBlkSize can be pre-calculated, there is no need to repeat calculation for each processing block, and log ₂ BlkSize can be obtained by table lookup, so the second calculation method can calculate the position by only two shift operations. Offset, with the advantage of low computational complexity. In some embodiments, PosDev^ y] may be a negative integer, and the shift operation should be adjusted according to the representation method of the negative integer in the processor used to ensure that the shift operation and the corresponding multiplication or division can be obtained. the result of.

S102. Determine two weighting factors of the current pixel point according to the first value and the second value, where one weighting factor is a sum of the first value and the second value, and another weighting factor is a first value and a second value. The difference between the two weighting factors is twice the first value; wherein the first value is determined or preset according to the image block size of the currently processed image block, and the image block size is the edge of the NxN image block. Long N, the NxN image block is a square image block containing NxN pixels.

The first value may be determined according to an image block size of the current processing block. If the currently processed image block size is BlkSize, BlkSize can be used as the first value.

The first value can also be preset. For example, a fixed value of FixedBlkSize can be used as the first value.

The two weighting factors of the current pixel point can be written as wl(x, y) and _W2 (x, y). The sum of the two weighting factors can be recorded as WeightSum, and the base-2 logarithm can be recorded as Log2WeightSum.

When the first value is determined as the image block size BlkSize of the current processing block, the two weights are The child can be calculated as follows.

Wl(x,y)=BlkSize+PosDev[x,y], where x, y = O..BlkSize-l

W2(x,y)=BlkSize - PosDev[x,y], where x, y = O..BlkSize-l

The sum of the two weighting factors is

WeightSum = wl(x,y) + w2(x,y) = 2-BlkSize

The base 2 logarithm of the sum of the two weighting factors is

Log2WeightSum = log ₂ WeightSum = log ₂ (2 · BlkSize) = log ₂ BlkSize+l

When the first value is determined to be a fixed value FixedBlkSize, the two weighting factors can be calculated as follows.

Wl(x,y)=FixedBlkSize+PosDev[x,y], where y = 0.. BlkSize- 1

W2(x,y)=FixedBlkSize - PosDev[x,y], where x, y = O..BlkSize-l

The sum of the two weighting factors is

WeightSum = wl(x,y) + w2(x,y) = 2 · FixedBlkSize

The base 2 logarithm of the sum of the two weighting factors is

Log2WeightSum = log ₂ WeightSum = log ₂ (2 · FixedBlkSize) = log ₂ FixedBlkSize+ 1 Because of the block size BlkSize or preset fixed block size in video encoding and decoding

FixedBlkSize can always be expressed as 2 ^N , where N is a positive integer, so the sum of the two weighting factors can always be expressed as 2 ^{Ν +1} , and the base 2 logarithm of the sum of the two weighting factors Log2WeightSum is N +l.

The embodiment of the present invention adopts a specific weighting factor acquisition method, so that the sum of the two weighting factors can always be expressed as 2 ^Ν+1 , so the shift operation can be used instead of using the two weighting factors for weighted prediction. The division operation reduces the computational complexity, thereby speeding up the running speed of the encoder and the decoder, and saving the power consumption of the encoding and decoding devices. The embodiment of the invention can avoid storing the weighting factor lookup table in the encoding and decoding chips, saving the encoding and decoding. The storage unit is solidified, thereby reducing the device cost. The embodiment of the invention can flexibly adjust the weighted prediction precision, can reduce the bit width required for the weighted prediction operation, save bandwidth in the codec processor, thereby increasing parallel processing capability and reducing access. The power required for operation.

An intra-frame weighted prediction method provided by an embodiment of the present invention is described below with reference to FIG. 3: S201. Determine a second value according to a difference between an abscissa component and a ordinate component of the current pixel in the currently processed image block, where the coordinate system referenced by the abscissa component and the ordinate component is the pixel point in the upper left corner of the currently processed image block. As the origin of the coordinates, the horizontal axis direction is horizontal to the right and the ordinate direction is vertically downward.

The second value in the embodiment of the present invention may be a position offset of the current pixel point position and the reference line. The current pixel position can be written as (x, y) where _X , y=0..BlkSize-l. The reference line refers to a line in which the abscissa component and the ordinate component of the currently processed image block are connected by the same pixel position, and can be recorded as #文( y) where x, y = 0.. BlkSize- l and x = y. The position offset may be recorded as PosDev^y], which may be determined by the abscissa component and the ordinate component of the current pixel point, and the determination method may be various. For example, when the first value is determined as the image block size BlkSize of the current processing block, the position offset may be determined according to the following method.

PosDev[x, y] = X— y, where X, y = O..BlkSize-l

For another example, when the first value is determined to be a fixed value FixedBlkSize, the position offset can be determined by normalizing the following method.

PosDev[x, y] = (FixedBlkSize - PosDev[x, y])/ BlkSize, where x, y = 0.. BlkSize-1 or equivalent, change the integer multiplication and division in the above formula to the left With right shift operation, PosDev[x, y] = (PosDev[x, y] « log ₂ FixedBlkSize) » log ₂ BlkSize, where x, y = 0.. BlkSize- 1 because log ₂ FixedBlkSize can be pre-computed, no need The calculation is repeated for each processing block, and the log ₂ BlkSize can be obtained by looking up the table. Therefore, the second calculation method can calculate the position offset only by two shift operations, and has the advantage of low computational complexity. . In some embodiments, PosDev^ y] may be a negative integer, and the shift operation should be adjusted according to the representation method of the negative integer in the processor used to ensure that the shift operation and the corresponding multiplication or division can be obtained. the result of.

S202. Determine, according to the first value and the second value, two weighting factors of the current pixel, where one weighting factor is a sum of the first value and the second value, and another weighting factor is a first value and a second value. Poor, the sum of the two weighting factors is twice the first value; wherein the first value is determined or preset according to the image block size of the currently processed image block, and the image block size is the side length of the NxN image block N, the NxN image block is a square image block containing NxN pixels. The first value may be determined according to an image block size of the current processing block. If the currently processed image block size is BlkSize, BlkSize can be used as the first value.

The first value can also be preset. For example, a fixed value of FixedBlkSize can be used as the first value. The two weighting factors of the current pixel point can be written as wl(x, y) and w2(x, y). The sum of the two weighting factors can be written as WeightSum, and its base 2 logarithm can be recorded as

Log2WeightSum.

When the first value is determined as the image block size BlkSize of the current processing block, the two weight factors can be calculated according to the following method.

Wl(x,y)=BlkSize+PosDev[x,y], where x, y = 0.. BlkSize- 1

W2(x,y)=BlkSize - PosDev[x,y], where x, y = 0..BlkSize- 1

The sum of the two weighting factors is

WeightSum = wl(x,y) + w2(x,y) = 2-BlkSize

The base 2 logarithm of the sum of the two weighting factors is

Log2WeightSum = log ₂ WeightSum = log ₂ (2 · BlkSize) = log ₂ BlkSize+l

Wl(x,y)=FixedBlkSize+PosDev[x,y], where y = 0.. BlkSize- 1

W2(x,y)=FixedBlkSize - PosDev[x,y], where x, y = O..BlkSize-l

The sum of the two weighting factors is

WeightSum = wl(x,y) + w2(x,y) = 2 · FixedBlkSize

The base 2 logarithm of the sum of the two weighting factors is

FixedBlkSize can always be expressed as ^2N , where N is a positive integer, so the sum of the two weighting factors can always be expressed as 2 ^{Ν +1} , and the base 2 logarithm of the sum of the two weighting factors Log2WeightSum is N+ l.

S203. Calculate, according to the two weighting factors and the corresponding two predicted values, a weighted prediction value of the current pixel point, where the two prediction values are predicted values respectively obtained by using two specified intra prediction modes. The two specified intra prediction modes may be any two different intra prediction modes. The method adopted by the embodiment of the present invention is particularly applicable to the two intra prediction modes shown in FIG. 4, that is, FIG. 4(a) The first mode shown is the second mode shown in Figure 4(b). The first mode is an intra prediction mode along a lower left diagonal direction, and the second mode is an intra prediction mode along an upper right diagonal direction.

The two predicted values are input of the method adopted by the embodiment of the present invention, and the method for obtaining the two predicted values is introduced below. According to the prediction method of the first mode, the first predicted value of each pixel position in the current block is calculated, and is recorded as

The second predicted value of each pixel in the current block can be calculated according to the prediction method of the second mode, and is denoted as PredSample2[x, y] where y = O..BlkSize-l. There are a variety of prediction methods in the industry that enable intra prediction along the prediction direction shown in Figure 4, which is a well-known technique. For example, according to the directional intra prediction method described in the HEVC coding framework, the first predicted value predicted by the pixel in the current intra-block position (x, y) in the first prediction mode is

PredSamplel[x, y] = p(x+ y + 1, — 1) where x, y = O..BlkSize-l

The second predicted value predicted by the pixel in the current block position (x, y) according to the second prediction mode is PredSample2[x, y] = p(-l, x+ y+1) where y = O..BlkSize-l

Where p(x,y) where x,y=0..2. BlkSize-l is the reconstructed pixel value around the current processing block, or the pixel value of the reconstructed pixel value around the current processing block after smoothing and so on.

There may be multiple correspondences between the two weighting factors and the two predicted values in the embodiment of the present invention, which are illustrated below. For convenience of description, the first weighting factor is specified to correspond to the first predicted value, the second weighting factor is corresponding to the second predicted value, the first weighting factor, the second weighting factor, the first predicted value, and the second predicted value may be as follows select. One of the two weighting factors is designated as the first weighting factor, and the other weighting factor is the second weighting factor. The first weighting factor and the second weighting factor specifying method and the first and second predicted value determining methods, or the second numerical value determining method, or the current pixel point position, or any of the above three influencing factors The combination of factors affects the impact. For example, according to the first and second predicted values specified above, and calculating the second value according to the method in S201, the weighting factor wl(x, y) obtained in S202 may be used as the first weighting factor, and the weighting factor w2 ( x, y) as the second weighting factor. For another example, if PredSamplel^y] is used as the second predicted value and PredSample x^y] is used as the first predicted value, wl(x, y) is correspondingly used as the second weight. The weight factor, with w2(x, y) as the first weighting factor. For example, if X - y in the calculation formula for obtaining the second value in S201 is replaced by |χ- y|, it is necessary to further determine the relative size of the X component and the y component of the current pixel position, if the X component is greater than or greater than Equal to the y component, w1(x, y) is taken as the first weighting factor, w2(x, y) is taken as the second weighting factor, and if the X component is less than or equal to or less than the y component, wl(x, y) is taken as The second weighting factor, using _W 2 (x, y) as the first weighting factor, and the first predicted value is

PredSamplel[x, y] , the second predicted value is PredSampleZI^ y].

To facilitate the description of the method of the embodiment of the present invention, after determining the first weighting factor and the second weighting factor of the pixel at the (x, y) position according to the method, adjusting the representation of the symbol, uniformly using the symbol w1 to represent the first weighting factor , using the symbol w2 to represent the second weighting factor; after determining the first predicted value and the second predicted value of the pixel located at the (x, y) position according to the method, adjusting the representation of the symbol, uniformly using the symbol

PredSamplel represents the first predicted value, and the second predicted value is represented by the symbol PredSample2.

The calculating the weighted prediction value of the current pixel point according to the two weighting factors and the corresponding two prediction values includes calculating the weighted prediction value by using a division operation or calculating the weighted prediction value by using a shift operation.

The weighted prediction value of the current pixel point can be calculated by the division operation according to the following formula: PredSample PredS ample = (wl- PredSamplel + w2. PredSample2 + (wl+w2) 12^ /(wl+w2) ,

Since (wl+w2) can be expressed equivalently using its base 2 logarithm Log2WeightSum, the above equation can be expressed equivalently by using a shift operation.

PredSample = ( wl · PredSamplel + w2 · PredSample2 + (1 « (Log2WeightSum-l))) » Log2WeightSum The equivalent calculation method uses the shift operation instead of the division operation in the original calculation method to obtain the same calculation result.

The two calculation methods respectively introduce (wl+w2)/2 and Log2WeightSum»l to compensate for the truncation error introduced by integer division or integer division implemented by shift.

The embodiment of the present invention reduces the operation complexity by using a shift operation instead of the division operation in the weighted prediction, thereby speeding up the running speed of the encoder and the decoder, and saving the power consumption of the encoding and decoding devices. The weighting factor lookup table is stored in the decoding chip, which saves the firmware of the codec and the decoder, thereby reducing the device cost. The embodiment of the invention can flexibly adjust the weighted prediction precision, can reduce the bit width required for the weighted prediction operation, and save the codec processor. Bandwidth, so Increase parallel processing power and reduce the power consumption required for access operations.

The following describes an intra-frame weighted prediction method based on table lookup according to an embodiment of the present invention with reference to FIG. 5:

S301. Determine an index value according to a difference between a horizontal coordinate component and a ordinate component of a current pixel in the currently processed image block, where the coordinate system referenced by the horizontal coordinate component and the ordinate component is taken as a pixel point in the upper left corner of the currently processed image block. The origin of the coordinates, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward.

The second value may be determined according to the difference between the abscissa component and the ordinate component of the current pixel in the currently processed image block according to the method in S101, and the index value Index is determined according to the second value PosDev y] according to the following formula [ x, y]

Index[x, y] = |PosDev[x, y]|, where x, y = O..BlkSize-l

The method for determining the index value may first obtain the second value according to the method in S101, and then determine the index value according to the second value; or directly from the difference between the abscissa component and the ordinate component of the current pixel point in the currently processed image block. The value is determined according to the following formula

Index[x, y] = | χ _ y|, where χ, y = O..BlkSize-l

Or

Index[x, y] = (FixedBlkSize-|x- y|)/BlkSize, where x, y = O..BlkSize-l

S302: Obtain two weighting factors from the weighting factor table according to the index value, where the two weighting factors are two weighting factors corresponding to the index value in the weighting factor table.

In the embodiment of the present invention, the weighting factor table may be two one-dimensional lookup tables WeightLUT1 and WeightLUT2. Certainly, different weight structures may be used to represent the weighting factor table, and the corresponding modification effect may be obtained by performing corresponding modification in the searching operation, which belongs to the one-dimensional lookup table Wei ghtLUT 1 and WeightLUT2 of the embodiment of the present invention. The merge is represented as an equivalent two-dimensional lookup table WeightLUT. Other similar situations are not here - enumerated.

Two weighting factors are obtained from the weighting factor table according to the cable I value. If PosDev^ y] > 0 or PosDev^ y] ≥ 0 or equivalent x > y or x ≥ y, use the cable 1 value Index[x, y] to find the first weighting factor table WeightLUT1 to obtain the first weighting factor wl (x,y), use the cable bow | value Index[x, y] to find the first The second weighting factor table WeightLUT2 obtains the second weighting factor w2(x, y); if PosDev^ y] ≤ 0 or

PosDev[x, y] < 0^^ 0^ x < y ^ x < y , then use the bow | value Index[x, y] to find the first weighting factor table WeightLUTl to obtain the second weighting factor w2(x, y) The second weighting factor table WeightLUT2 is used to find the first weighting factor wl(x, y) using the value of the index [x, y].

The method for obtaining two weight factors from the weighting factor table according to the index value may perform corresponding change according to the data structure of the weighting factor table, and the change may obtain the same first for the pixel located at the (x, y) position. The weighting factor and the second weighting factor. Therefore, the change is a table lookup in the embodiment of the present invention.

S303. Calculate a weighted prediction value of the current pixel point according to the two weighting factors and the two prediction values, where the two prediction values are predicted values respectively obtained by using two specified intra prediction modes.

The first predicted value PredSamplel[x, y] may be determined according to the method in S203, where x, y = O..BlkSize-l and the second predicted value PredSample2[x, y] where x, y = O..BlkSize-l.

The weighted prediction value PredSample may be calculated according to the first weighting factor w1, the second weighting factor w2, the first predicted value PredSamplel, and the second predicted value PredSample2 of each pixel according to the method in S203. Here, the first weighting factor wl and the second weighting factor w2 can be determined by the method described in S302.

In the embodiment of the present invention, as shown in FIG. 6, the method for constructing the weighting factor table can be implemented as follows:

S401: Determine an index value of the weighting factor table according to the first value, where the index value includes a positive integer smaller than the first value and 0; wherein, the first value is determined according to an image block size of the currently processed image block or is preset The image block size is an edge length N of the NxN image block, and the NxN image block is a square image block including NxN pixel points.

Assuming that the first value determined is M, then the index is all integers in the set [0, M-1], including 0 and M1. S402. Determine, according to each index value, two weighting factors corresponding thereto, thereby generating a weighting factor table. The method for determining two weighting factors corresponding to each of the indexing values includes: determining, according to the current index value a current second value, determining, according to the first value and the second value, two weighting factors corresponding to the index value, wherein one weighting factor is a sum of a first value and a second value, and another weighting factor is The difference between the first value and the second value, the sum of the two weight factors being twice the first value.

The method of determining the current second value based on the current index value includes directly using the value of the cable as the second value.

Determining, according to the first value and the second value, two weight factors corresponding to the index value, and determining, according to the first value and the second value, two of the current pixel points in S102 The weighting factors are the same.

The embodiment of the present invention reduces the operation complexity by using a shift operation instead of the division operation in the weighted prediction, thereby speeding up the running speed of the encoder and the decoder, and saving the power consumption of the encoding and decoding devices. The weighting factor lookup table is stored in the decoding chip, which saves the firmware of the codec and the decoder, thereby reducing the device cost. The embodiment of the invention can flexibly adjust the weighted prediction precision, can reduce the bit width required for the weighted prediction operation, and save the codec processor. Bandwidth, which increases parallel processing power and reduces the power required for access operations.

As shown in FIG. 7, an embodiment of the present invention provides a determining device for two weighting factors used in pairing, including:

a first determining module 701, configured to determine a second value according to a difference between an abscissa component and a ordinate component of the current pixel in the currently processed image block, where the coordinate system referenced by the abscissa component and the ordinate component is currently processed The pixel in the upper left corner of the image block is used as the coordinate origin, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward;

a second determining module 702, configured to determine, according to the first value and the second value, two weighting factors of the current pixel, where one weighting factor is a sum of the first value and the second value, and another weighting factor is the first a difference between the value and the second value, wherein the sum of the two weighting factors is twice the first value; wherein the first value is determined or preset according to an image block size of the currently processed image block, and the image block size is The side length N of the NxN image block, the NxN image block is positive including NxN pixel points Square image block.

As shown in FIG. 8, an embodiment of the present invention provides an intra-frame weighting prediction apparatus, including:

a third determining module 801, configured to determine a second value according to a difference between an abscissa component and a ordinate component of the current pixel in the currently processed image block, where the coordinate system referenced by the abscissa component and the ordinate component is currently processed The pixel in the upper left corner of the image block is used as the coordinate origin, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward;

a fourth determining module 802, configured to determine, according to the first value and the second value, two weighting factors of the current pixel, where one weighting factor is a sum of the first value and the second value, and another weighting factor is the first a difference between the value and the second value, wherein the sum of the two weighting factors is twice the first value; wherein the first value is determined or preset according to an image block size of the currently processed image block, and the image block size is An edge length N of the NxN image block, the NxN image block being a square image block including NxN pixel points;

The first prediction module 803 is configured to calculate a weighted prediction value of the current pixel point according to the two weighting factors and the two prediction values, where the two prediction values are predicted values respectively obtained by using two specified intra prediction modes.

As shown in FIG. 9, an embodiment of the present invention provides an intra-frame weighting prediction apparatus based on a look-up table, including: a fifth determining module 901, configured to determine an index value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of a current pixel in the currently processed image block, where the horizontal coordinate component and the coordinate system referenced by the ordinate component are currently processed images The pixel in the upper left corner of the block is taken as the coordinate origin, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward;

a sixth determining module 902, configured to acquire two weight factors from the weighting factor table according to the index value, where the two weighting factors are two weighting factors corresponding to the index value in the weighting factor table;

The second prediction module 903 is configured to calculate a weighted prediction value of the current pixel point according to the two weighting factors and the two prediction values, where the two prediction values are prediction values respectively acquired in two specified intra prediction modes.

In some embodiments, the apparatus further includes:

The weighting factor table construction module 904 is configured to determine, according to the first value, an index value of the weighting factor table, where the index value includes a positive integer smaller than the first value and 0; wherein the first value is based on the currently processed image block The image block size is determined or preset, the image block size is the side length N of the NxN image block, and the NxN image block is a square image block including NxN pixel points; corresponding to each index value is determined Two weighting factors, thereby generating a weighting factor table; the method for determining two weighting factors corresponding to each index value includes: determining a current second value according to the current index value, according to the first value and the The second value determines two weighting factors corresponding to the index value, wherein one weighting factor is a sum of the first value and the second value, and the other weighting factor is a difference between the first value and the second value, the two The sum of the weighting factors is twice the first value.

The apparatus provided in this embodiment may be used to implement the method in the method embodiment, and the implementation principle and the technical effect are similar, and details are not described herein again.

The technology provided by the embodiments of the present invention can be applied to the field of digital signal processing, and is implemented by an encoder and a decoder. Video encoders, decoders are widely used in a variety of communication equipment or electronic devices, such as: digital TV, set-top box, media gateway, mobile phone, wireless device, personal data assistant (PDA), handheld or portable computer, GPS receiver / navigator, camera, video player, camcorder, video recorder, surveillance equipment, video conferencing and video telephony equipment, etc. Such equipment It includes the processor, memory, and interface for transferring data. The video codec can be implemented directly by a digital circuit or chip such as a DSP (digital signal processor), or by software code that drives a processor to execute the flow in the software code.

A person skilled in the art can understand that all or part of the steps of implementing the above method embodiments may be completed by using hardware related to program instructions, and the foregoing program may be stored in a computer readable storage medium, and the program is executed when executed. The foregoing steps include the steps of the foregoing method embodiments; and the foregoing storage medium includes: a medium that can store program codes, such as a ROM, a RAM, a magnetic disk, or an optical disk.

It should be noted that the above embodiments are only for explaining the technical solutions of the present invention, and are not intended to be limiting; although the present invention has been described in detail with reference to the foregoing embodiments, it will be understood by those skilled in the art that: The technical solutions described in the foregoing embodiments are modified, or some of the technical features are equivalently replaced. The modifications and substitutions do not depart from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

Claim

A method for determining two weighting factors used in pairing, comprising: determining a second value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of a current pixel in a currently processed image block, the horizontal coordinate The coordinate system referenced by the component and the ordinate component is taken as the coordinate origin of the pixel in the upper left corner of the currently processed image block, the horizontal direction is horizontally rightward, and the ordinate direction is vertically downward; the current value is determined according to the first value and the second value. Two weighting factors of the pixel, wherein one weighting factor is the sum of the first numerical value and the second numerical value, and the other weighting factor is the difference between the first numerical value and the second numerical value, and the sum of the two weighting factors is the first numerical value The first value is determined or preset according to an image block size of the currently processed image block, the image block size is a side length N of the NxN image block, and the NxN image block is a NxN pixel point. Square image block.

2. An intra-frame weighted prediction method, comprising:

And determining a second value according to a difference between a horizontal coordinate component and a longitudinal coordinate component of the current pixel in the currently processed image block, where the coordinate system referenced by the horizontal coordinate component and the ordinate component is a coordinate point of the upper left corner of the currently processed image block. The origin, the horizontal axis direction is horizontally to the right, and the ordinate direction is vertically downward; determining two weighting factors of the current pixel point according to the first value and the second value, wherein one weighting factor is the first value and the second value And another weighting factor is a difference between the first value and the second value, the sum of the two weighting factors being twice the first value; wherein the first value is determined according to the image block size of the currently processed image block or is Setting, the image block size is an edge length N of the NxN image block, and the NxN image block is a square image block including NxN pixel points;

3. An intra-frame weighted prediction method based on table lookup, comprising:

Determining an index value according to a difference between an abscissa component and a ordinate component of the current pixel in the currently processed image block, wherein the coordinate system referenced by the abscissa component and the ordinate component is a pixel origin of the upper left corner of the currently processed image block , the horizontal axis direction is horizontally to the right, and the ordinate direction is vertically downward; according to the cable I value, two weighting factors are obtained from the weighting factor table, where the two weighting factors are corresponding to the weighting factor table. Two weighting factors for the index value; The weighted prediction values of the current pixel points are calculated according to the two weighting factors and the two predicted values, and the two predicted values are predicted values respectively acquired in two specified intra prediction modes.

The method according to claim 3, wherein the method for constructing the weighting factor table comprises:

Determining, according to the first value, an index value of the weighting factor table, where the index value includes a positive integer smaller than the first value and 0; wherein, the first value is determined or preset according to an image block size of the currently processed image block. The image block size is an edge length N of the NxN image block, and the NxN image block is a square image block including NxN pixel points;

Determining two weighting factors corresponding to each index value, thereby generating a weighting factor table; the method for determining two weighting factors corresponding to each of the indexing values includes: determining the current number according to the current index value a second value, determining, according to the first value and the second value, two weighting factors corresponding to the index value, wherein one weighting factor is a sum of a first value and a second value, and another weighting factor is first The difference between the value and the second value, the sum of the two weighting factors being twice the first value.

A device for determining two weighting factors for pairing, comprising: a first determining module, configured to determine, according to a difference between an abscissa component and a ordinate component of a current pixel point in a currently processed image block; The two numerical values, the coordinate system referenced by the abscissa component and the ordinate component is the pixel origin of the upper left corner of the currently processed image block, the horizontal axis direction is rightward, and the ordinate direction is vertically downward;

a second determining module, configured to determine, according to the first value and the second value, two weighting factors of the current pixel, where one weighting factor is a sum of the first value and the second value, and another weighting factor is the first value a difference from the second value, the sum of the two weighting factors being twice the first value; wherein the first value is determined or preset according to an image block size of the currently processed image block, the image block size is NxN The side length N of the image block, the NxN image block is a square image block containing NxN pixels.

6. An intra-frame weighted prediction apparatus, comprising:

a third determining module, configured to determine a second value according to a difference between an abscissa component and a ordinate component of the current pixel in the currently processed image block, where the abscissa component and the ordinate component are referenced The target system uses the pixel point in the upper left corner of the currently processed image block as the coordinate origin, the horizontal axis direction is horizontal to the right, and the ordinate direction is vertically downward;

7. An intra-frame weighting prediction apparatus based on table lookup, comprising:

8. The device according to claim 7, wherein the device further comprises:

Weighting factor table construction module, for

Determining, according to the first value, an index value of the weighting factor table, where the index value includes a positive integer smaller than the first value and 0; wherein, the first value is determined or preset according to an image block size of the currently processed image block. The image block size is an edge length N of the NxN image block, and the NxN image block is a square image block including NxN pixel points; two corresponding to each index value are determined a weighting factor, thereby generating a weighting factor table; the method for determining two weighting factors corresponding to each index value includes: determining a current second value according to the current index value, according to the first value and the first The two values determine two weighting factors corresponding to the index value, wherein one weighting factor is the sum of the first value and the second value, and the other weighting factor is the difference between the first value and the second value, the two weights The sum of the factors is twice the first value.