WO2013000437A1

WO2013000437A1 - Intra-frame prediction-based decoding method and decoding device

Info

Publication number: WO2013000437A1
Application number: PCT/CN2012/078039
Authority: WO
Inventors: 杨海涛; 赖昌材
Original assignee: 华为技术有限公司
Priority date: 2011-06-30
Filing date: 2012-07-02
Publication date: 2013-01-03
Also published as: CN102857763B; CN102857763A

Abstract

The embodiments of the present invention provide an intra-frame prediction-based decoding method and decoding device. The decoding method comprises: acquiring for a current prediction unit one or multiple split flags, a split shape flag, and a split direction flag from a compressed video image stream; splitting the current prediction unit ultimately into multiple square transform units or multiple rectangular transform units according to the acquired one or multiple split flags, split shape flag, and split direction flag; executing decoding process on each square transform unit of the ultimately split multiple square transform units or on each rectangular transform unit of the ultimately split multiple rectangular transform units.

Description

One by one

Decoding method and decoding device based on intra prediction

The present application claims priority to Chinese Patent Application No. 201110182388.8, entitled "Decoding Method and Decoding Device Based on Intra Prediction", filed on June 30, 2011, the entire disclosure of which is incorporated by reference. Combined in this application. Technical field

Embodiments of the present invention relate to the field of image processing processing, and more particularly, to a decoding method and a decoding apparatus based on intra prediction. Background technique

In order to save the bandwidth of the network transmission, the video image is usually compressed and encoded by the encoder at the transmitting end, and then transmitted to the receiving end through the network, and the receiving end decompresses the decoder to recover the video image. The transmitting end or receiving end can be a mobile phone, a portable computer, a handheld computer, a video camera, a video monitoring device, and the like.

In the decoding scheme of the video image compressed code stream, a prediction module and a transform module for the image block signal are usually included. The prediction module predicts the current image block signal based on the decoded information to obtain a prediction signal. The transform module performs an inverse transform operation on the received transform coefficients to obtain a reconstructed difference signal. The decoding end adds the prediction signal to the reconstructed difference signal to obtain a reconstructed image block signal.

The prediction is divided into two types: inter prediction and intra prediction. The inter prediction obtains a prediction signal based on temporal proximity of the current image block to the encoded information, and the intra prediction obtains the prediction signal based on spatially adjacent encoded information of the current image block.

In the intraframe coding and decoding scheme, the coding unit is an image block that is operated when encoding or decoding is performed at the encoding end or the decoding end. The prediction unit is an image block having an independent prediction mode in the coding unit. The prediction block is an image block in which the coding unit performs a prediction operation, and one prediction unit may include a plurality of prediction blocks, and the prediction blocks may perform a prediction operation using a common prediction mode (ie, a prediction mode of a prediction unit in which the prediction unit is located). The transform unit is an image block in which a transform operation is performed in the coding unit. The prediction block size is generally the same as the transform unit size because the correlation of the prediction difference signals of the boundary portions of adjacent intra prediction blocks is weak, so the transform unit should not cross the prediction block boundary. At the same time, considering that the correlation of the difference signal inside the prediction block is strong, and the large block transformation has higher energy concentration performance than the small block transformation, it is usually selected and pre-selected. - - Measure the same size of the transform unit.

In a square block-based intraframe codec solution, a quadtree partitioning flag of a current prediction unit may be obtained from a code stream, and the current prediction unit is iteratively divided into square transform units of different sizes according to the label. A quadtree partitioning flag of 1 indicates that further segmentation is required, and a flag of 0 indicates that segmentation is stopped. The prior art uses a split-transform-flag as a split flag to indicate whether the current block is further split. The split-transform-flag can be encoded as a separate syntax element or jointly encoded with other syntax elements. The decoding end needs to perform the following operations on each transform unit in sequence: performing an intra prediction operation on the intra prediction mode of the prediction unit corresponding to the transform unit to obtain a prediction signal; and performing inverse transform on the received transform coefficient to obtain a reconstructed prediction difference signal; The predicted signal is added to the reconstructed predicted difference signal to obtain a reconstructed image signal.

Since the technical solution only uses the square division method to obtain the square transformation unit, the strip texture which frequently appears is not considered. Therefore, this technique has a low compression efficiency when a strip texture is present. Summary of the invention

Embodiments of the present invention provide a decoding method and a decoding apparatus based on intra prediction, which can improve video image compression efficiency.

An embodiment of the present invention provides a decoding method based on intra prediction, which includes: acquiring, for a current prediction unit, one or more segmentation marks, a segmentation shape mark, and a segmentation direction from a compressed code stream of a video image. a mark, a split mark corresponding to the current block among the one or more split marks is used to indicate whether to divide the current block into a plurality of blocks, the split shape mark indicates that the current block is divided into multiple square blocks or divided into multiple a rectangular block, the dividing direction mark indicating that the current block is divided into a plurality of rectangular blocks in the horizontal direction or divided into a plurality of rectangular blocks in the vertical direction, and each of the plurality of square blocks has the same width and height, and a plurality of Each rectangular block in the rectangular block is not equal in width and height; the current prediction unit is finally divided into a plurality of square transformation units or a plurality according to the acquired one or more segmentation marks, the segmentation shape flag, and the segmentation direction flag. a rectangular transform unit; each of the plurality of square transform units obtained by the final split or finally Each of the plurality of rectangular transform units obtained by the division performs a decoding process.

According to another embodiment of the present invention, there is provided an apparatus for decoding based on intra prediction, comprising: a label acquisition module, configured to acquire one or more from a compressed code stream of a video image for a current prediction unit - a segmentation marker, a segmentation shape marker, and a segmentation direction marker, wherein the segmentation marker corresponding to the current block of the one or more segmentation markers is used to indicate whether to divide the current block into a plurality of blocks, the segmentation shape marker indication The current block is divided into a plurality of square blocks or divided into a plurality of rectangular blocks, and the division direction mark indicates that the current block is divided into a plurality of rectangular blocks in the horizontal direction or is divided into a plurality of rectangular blocks in the vertical direction, and the plurality of square blocks are Each square block has the same width and height, and each of the plurality of rectangular blocks has a width and height that are not equal; the segmentation module is configured to: according to the acquired one or more segmentation marks, the segmentation shape mark, and the segmentation The direction marker finally divides the current prediction unit into a plurality of square transformation units or a plurality of rectangular transformation units; and a decoding module, configured to use each of the plurality of square transformation units obtained by the final division or the final division Each rectangular transform unit in the rectangular transform unit performs a decoding process.

The embodiment of the present invention may process the strip by using an optional transform unit (ie, a rectangular transform unit) that adapts the strip texture of the image frame by selecting the prediction unit into a square transform unit or a rectangular transform unit in the same prediction unit. Shaped texture, which can improve the compression efficiency of video images. DRAWINGS

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the embodiments or the prior art description will be briefly described below. Obviously, the drawings in the following description are only some of the present invention. For the embodiments, those skilled in the art can obtain other drawings according to the drawings without any creative labor.

1A to 1C are diagrams showing a coding unit, a prediction unit, and a transformation unit according to an embodiment of the present invention.

2A and 2B are diagrams showing a coding unit, a prediction unit, and a transformation unit according to another embodiment of the present invention.

FIG. 3 is a schematic flow chart showing a decoding method based on intra prediction according to an embodiment of the present invention.

4 is a schematic flow chart showing a decoding method based on intra prediction according to another embodiment of the present invention.

FIG. 5 shows a schematic structural diagram of a decoding apparatus based on intra prediction according to an embodiment of the present invention. - - detailed description

BRIEF DESCRIPTION OF THE DRAWINGS The technical solutions in the embodiments of the present invention are clearly and completely described in the following with reference to the accompanying drawings. 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.

Video encoders or decoders are widely used in a variety of electronic devices, such as mobile phones, laptops, MP3/MP4, handheld computers, video cameras, video surveillance devices. The video encoder or decoder can be implemented by a digital processing circuit, such as a DSP chip, or by a software (such as a CPU) calling software code.

In the new generation video codec technology, an intra prediction mode of rectangular block - short range intra prediction mode (SDIP) is proposed. Unlike the square block-based intra prediction method, the prediction unit of SDIP is a rectangular block of M x N, where M and N are not equal. When decoding each intra-coded unit, a 1-bit SDIP_flag flag is first obtained to determine whether the prediction mode is employed. If the SDIP_flag is 1, then the 1-bit SDIP_direction flag is obtained to judge whether the prediction unit is strip-shaped in the horizontal direction or in the vertical direction. The decoding end needs to perform the following operations on each transform unit in sequence: performing an intra prediction operation by using an intra prediction mode of a prediction unit corresponding to the transform unit to obtain a prediction signal; and performing inverse transformation by inversely transforming the received transform coefficients Predicting the difference signal; adding the predicted signal to the reconstructed predicted difference signal to obtain a reconstructed image signal.

When the current coding unit adopts the SDIP predictive coding mode, each prediction unit needs to independently select its own prediction mode. Therefore, the coding end divides the prediction unit by multiple division methods, which greatly increases the coding complexity.

1A to 1C are schematic views of a coding unit, a prediction unit, and a transformation unit, according to an embodiment of the present invention.

As shown in FIG. 1A, the coding unit 100 according to an embodiment of the present invention may be the same size as the prediction unit 110, that is, the current coding unit may be used as one prediction unit.

As shown in FIG. 1B, the prediction unit 110 is first divided into four rectangular blocks of equal size, such as a rectangular block 130. The rectangular block located at the top of FIG. 1B is further divided into four equal-sized rectangular blocks, such as rectangular block 120. Since the rectangular block 130 and the rectangular block 120 are not intermittently divided, they are two rectangular transform units among the plurality of rectangular transform units finally obtained by the segmentation prediction unit 110. Similarly, can The prediction unit 110 is divided into a plurality of levels by a rectangular division method to obtain a plurality of rectangular transformation units.

As shown in Fig. 1C, the prediction unit 110 is first divided into four equal-sized square blocks, such as a square block 140. The square block located at the lower left of Fig. 1C is further divided into two to obtain a square block 150. Since the square block 140 and the square block 150 are not further divided, they are two square transform units among the plurality of square transform units finally obtained by the segmentation prediction unit 110. Similarly, the prediction unit 110 may be divided into a plurality of levels by a square division method to obtain a plurality of square transformation units.

It should be noted that the rectangular block or the rectangular transform unit described herein may represent a quadrangular block or transform unit whose unequal horizontal and vertical directions are different, and the square or square transform unit described herein may represent horizontal direction and vertical direction. A quadrilateral block or transform unit of equal size.

As shown in FIG. 2A, according to another embodiment of the present invention, the current coding unit 200 may be divided into four equal-sized prediction units 210, where the width and height of each prediction unit 210 are respectively obtained by the current coding unit 200. One-half the width. Those skilled in the art will know how to divide a coding unit into four equal-sized prediction units, which will not be described in detail herein.

As shown in FIG. 2B, according to another embodiment of the present invention, each prediction unit 210 may be further divided into a plurality of transform units. For example, the prediction unit 210 located at the upper left of FIG. 2B adopts a rectangular division manner. After two divisions, the rectangular transformation unit 220 is obtained. The prediction unit 210 located at the upper right of FIG. 2B adopts a square division method, and after two divisions, The square transform unit 230 is located at the lower right of FIG. 2B and is divided by a rectangle. After one division, the transform unit 240 is obtained. The prediction unit 210 located at the lower left of FIG. 2B is not divided and directly serves as the square transform unit 250. In other words, each prediction unit 210 is finally divided into a plurality of rectangular or square transformation units as shown in Fig. 2B.

Since the prediction unit division manner (i.e., the prediction unit division manner without using a rectangle) is reduced according to an embodiment of the present invention, the coding side computational complexity can be reduced.

As shown in FIG. 3, in 310, one or more split marks, a split shape mark, and a split direction mark are acquired from the compressed code stream of the video image for the current prediction unit, and the one or more points are divided. - a split flag corresponding to the current block in the cut mark is used to indicate whether the current block is divided into a plurality of blocks, the split shape mark indicates that the current block is divided into a plurality of square blocks or divided into a plurality of rectangular blocks, the split The direction mark indicates that the current block is divided into a plurality of rectangular blocks in the horizontal direction or is divided into a plurality of rectangular blocks in the vertical direction, and each of the plurality of square blocks has the same width and height, and each of the plurality of rectangular blocks The width of the rectangular block is not equal to the height. In 320, the current prediction unit is finally divided into a plurality of square transformation units or a plurality of rectangular transformation units according to the acquired one or more segmentation markers, the segmentation shape marker, and the segmentation direction marker. In 330, a decoding process is performed on each of the plurality of square transform units obtained by the final split or each of the plurality of rectangular transform units obtained by the final split.

In the decoding process based on intra prediction, in order to obtain a reconstructed difference signal, the coding unit is divided into transform units. According to an embodiment of the present invention, for example, a coding unit may be divided into a plurality of prediction units (e.g., square prediction units) or a coding unit may be used as one prediction unit. Then, one or more segmentation markers can be acquired for each prediction unit from the code stream, wherein each segmentation marker is used to indicate whether to segment a certain image block corresponding to the segmentation marker. In the case where it is determined that the prediction unit is divided, a split shape mark and a split direction mark may be acquired from the code stream, and the prediction unit may be divided into a rectangular transform unit or a square transform unit according to the split shape mark. In the case where it is determined that the prediction unit is divided into rectangular transformation units, the prediction unit is divided in the horizontal direction or in the vertical direction according to the division direction mark. After the prediction unit is subjected to one or more divisions to obtain the final transformation unit, the decoding process is performed for each transformation unit. The embodiment according to the present invention is not limited thereto, and an instant decoding operation may be performed on a transform unit that is no longer divided. Since the same prediction unit uses a split shape mark and a split direction mark, the split direction and the split shape of the transform unit are identical in the same prediction unit.

According to the decoding method based on intra prediction provided by the embodiment of the present invention, by selecting the prediction unit into a square transformation unit or a rectangular transformation unit in the same prediction unit, a strip texture adapted to the image frame may be adopted. Selecting a transform unit (ie, a rectangular transform unit) to process the strip texture can improve video image compression efficiency. Since the prediction unit division mode (i.e., the prediction unit division manner without using a rectangle) is reduced according to an embodiment of the present invention, the encoding side computational complexity can be further reduced.

The following description of FIG. 3 will be described with reference to FIGS. 1A to 1C and FIGS. 2A and 2B and a specific implementation method. - - Decoding method.

At 310, one or more segmentation marks, a segmentation shape marker, and a segmentation direction marker are acquired from the compressed code stream of the video image for the current prediction unit, and the segmentation marker corresponding to the current block among the one or more segmentation markers For indicating whether to divide the current block into a plurality of blocks, the split shape flag indicates that the current block is divided into a plurality of square blocks or divided into a plurality of rectangular blocks, and the split direction mark indicates that the current block is divided into multiples in the horizontal direction. The rectangular block is divided into a plurality of rectangular blocks in the vertical direction, and each of the plurality of square blocks has the same width and height, and each of the plurality of rectangular blocks has a width and a height that are not equal.

According to an embodiment of the present invention, the current block may be an image block of the same size as the current prediction unit. According to another embodiment of the present invention, one of the plurality of blocks obtained by the division is used as the current block of the next division. In other words, the segmentation of the prediction unit may be a hierarchical iterative process until the segmentation flag corresponding to the current block indicates that the current block is no longer segmented, and the image block obtained at this time is a transformation unit, as shown in FIG. Figure 1 C and Figures 2A and 2C.

According to an embodiment of the present invention, each of the plurality of rectangular blocks divided in the horizontal direction (for example, the rectangular transformation unit 220 in FIG. 2B) has a width greater than a height, and is divided into a plurality of segments obtained in the vertical direction. The width of each rectangular block in the rectangular block (for example, the rectangular transformation unit 240 in Fig. 2B) is smaller than the height.

According to an embodiment of the present invention, in a case where the first one of the one or more segmentation marks indicates that the current block is divided into a plurality of blocks, after acquiring the first segmentation flag, the compressed code stream is acquired. A split shape mark; and in the case where one split shape mark indicates that the current block is divided into a plurality of rectangular blocks, after the split shape mark is acquired, one split direction mark is acquired from the compressed code stream. In other words, a corresponding segmentation flag can be acquired for an image block at a different segmentation level of one prediction unit, and a prediction unit can acquire a segmentation shape marker and a segmentation direction marker. In addition, after acquiring the first divided shape mark and then acquiring the split shape mark and the split direction mark in order, the efficiency of acquiring the mark can be improved. Of course, embodiments in accordance with the present invention are not limited thereto, and these indicia may be acquired for the same prediction unit in other orders.

According to an embodiment of the present invention, for example, in a specific implementation, syntax elements may be added to the structure of the compressed code stream: non-square-transform-flag and non-square-transform-direction. The former indicates whether the current prediction unit is divided into multiple rectangular transformation units, and the latter indicates whether horizontal or vertical is used. - - Strip division to obtain multiple rectangular transformation units. The semantics of the values of the two syntax elements can be explained as follows: non-square-transform-flag is 1 to divide the current prediction unit into several rectangular transformation units, and non-square_transform_flag is 0 to indicate the current prediction unit. Divided into several square transform units. Non-square-transform-direction is 1 to indicate that the rectangular transform unit is determined by the horizontal strip division method. The so-called transform unit determined by the horizontal strip division method means that the transform unit has a width greater than a height. Non-square-transform-direction is 0 to indicate that the rectangular transform unit is determined in a vertical stripe manner. The rectangular transformation unit determined by the vertical strip division means that the width of the transformation unit is smaller than the height. The intra prediction of luminance is described as an example in accordance with an embodiment of the present invention.

The decoding end may perform the following operations to obtain a non-square_transform_flag and a non-square_transform_direction: determining whether a non-square_transform-flag is obtained from the code stream, and if so, performing the corresponding coding according to a specified coding manner. The parsing and decoding operations obtain non-square_transform_flag from the code stream, otherwise the non-square_transform_flag is set to the default value of 0; if non-square-transform-flag is 1, the shell 'J is encoded according to the specified The mode performs a corresponding parsing and decoding operation to obtain a non-square-transform-direction from the code stream.

The non-square_transform_flag may be obtained from the code stream based on at least one of the following conditions: 1) The split-transform-flag obtained is the first split-transform-flag in the current prediction unit, 2) The obtained split_transform_flag is 1, which means that the current block is divided into multiple transform units, and 3) the current block size satisfies the given constraint condition. Of course, embodiments according to the invention are not limited thereto. In the embodiment of the present invention, it is assumed that only the above items 1) and 2) are considered as the judgment conditions.

The pseudo code syntax table corresponding to the above decoding operation is shown in Table 1:

- -

The flag split-transform_flag indicates the first split_transform_flag in the current prediction unit. The value of the flag split_transform_flag can be determined in accordance with conventional methods in the art. In the table, u(l) indicates that the code is encoded as a 1-bit unsigned integer variable, that is, directly written to the code stream; ae(v) represents a binary entropy coding mode based on the context model.

It should be noted that the semantics of the syntax elements non-square-transform-flag and non-square-transform-direction according to embodiments of the present invention may also be used in other interpretation methods. For example, a non-square_transform_flag of 1 indicates that the current prediction unit is divided into a plurality of square transformation units, and a non-square_transform_flag of 0 indicates that the current prediction unit is divided into a plurality of rectangular transformation units. The non-square_transform_direction is 1 to determine the rectangular transform unit in a vertical strip division manner, and the non-square_transform_direction 0 indicates that the rectangular transform unit is determined in a horizontal strip division manner. When the semantic interpretation method is different, the decoding process needs to be changed accordingly. A person skilled in the art knows how to make such a change according to an embodiment of the present invention, and details are not described herein again.

In 320, the current prediction unit is finally divided into a plurality of square transformation units or a plurality of rectangular transformation units according to the acquired one or more segmentation marks, the segmentation shape marker, and the segmentation direction marker.

According to an embodiment of the present invention, in a case where the split flag corresponding to the current block in the one or more split marks indicates that the current block is divided into a plurality of blocks, the current block is divided into the split direction corresponding to the split direction mark into A plurality of square blocks or rectangular blocks corresponding to the divided shape marks.

Specifically, in the case where the split shape flag indicates that the current block is divided into a plurality of square blocks, - - The current block is divided into four equal-sized square blocks, and each of the four equal-sized square blocks has a width or height that is one-half of the width or height of the current block, respectively.

In the case where the split shape flag indicates that the current block is divided into a plurality of rectangular blocks, the current block is further divided into four equal-sized rectangular blocks according to the split direction mark. In the case where the split direction flag indicates that the current block is divided into a plurality of rectangular blocks in the horizontal direction, the current block is divided into four equal-sized rectangular blocks in the horizontal direction, wherein each of the four equal-sized rectangular blocks The width of the rectangular block is the same as the width of the current block, and the height of each of the four equally sized rectangular blocks is one quarter of the height of the current block.

Further, in the case where the split direction flag indicates that the current block is divided into a plurality of rectangular blocks in the vertical direction, the current block is divided into four equal-sized rectangular blocks in the vertical direction, wherein four equal-sized rectangular blocks are divided. The height of each of the rectangular blocks is the same as the height of the current block, and the width of each of the four equally sized rectangular blocks is one quarter of the width of the current block.

According to an embodiment of the present invention, in a specific implementation, for example, before performing the decoding process, all transform unit splitting flags in the current prediction unit are split-transform-flag, rectangular transform unit-labeled non-square-transform-flag, and rectangular transform The unit split direction label non-square-transform-direction has been obtained from the code stream and can be used in the decoding process. The current block is the current prediction unit before the current prediction unit is split. After segmenting the current prediction unit, the current block can be split to get any image block.

According to an embodiment of the present invention, each input parameter may be first acquired: a brightness position (xB, yB) indicating the displacement of the pixel in the upper left corner of the current block relative to the pixel in the upper left corner of the current image; the logarithm of the current block width indicates log2TrafoWidth and height The logarithm represents log2TrafoHeight; the depth of the current block relative to the current coding unit trafoDepth; the intra prediction mode intraPredMode used by the current block; the color component index cldx of the current block, cldx is 0 for the luma component, cldx is 1 or 2 Chromatic component. When non-square_transform_flag is 0, log2Trafo Width is equal to log2TrafoHeight, and these two variables are recorded as log2TrafoSize. Here, log2TrafoWidth represents the logarithm of the width from the base 2, for example, 4 of 4 blocks (the value of ^21⁄4&\\3⁄41«1 is 2, and the log2TrafoWidth of 8 X 8 blocks is 3. log2TrafoSize means base 2 Find the logarithm of the side length of the square block.

According to the embodiment of the present invention, when the current block split_transform_flag is 1, that is, when the current block needs to be split, the following processing is performed. - - First, set the position information of the four blocks obtained by the division. Dividing the four positions of the upper left corner of the luminance position is divided into 'J' (xB, yB), (xBl, yBl), (xB2, yB2) and (xB3, yB3). When non-square_transform_flag is 1 and non-square-transform-direction is 1, χΒ1, xB2, xB3 are both set to xB, yBl is set to yB+((l«log2TrafoHeight)»2), yB2 is set For yBl+((l«log2TrafoHeight)»2), set yB3 to yB2+(( 1 «log2TrafoHeight)»2); when non-square—transform—flag is 1 and non-square—transform—direction is 0, yB 1 , yB2, yB3 are set to yB; ^ xBl is set to xB+((l«log2TrafoWidth)»2), χΒ2 is set to xB 1 +(( 1 «log2TrafoWidth)»2) , χΒ33⁄4 is set to χΒ2+(( 1 «log2TrafoWidth)»2); When non-square-transform-flag is 0, xB2 is set to xB, ^xx1 and xB3 are set to xB+(( 1 «log2TrafoSize)» 1 ), yBl is set to yB, yB2 and yB3 are set to yB+((l«log2TrafoSize)»l). Here, l«x refers to shifting the binary number 1 to the left by x bits. For example, 1<<2 means binary 100, and x>>2 means shifting the binary number X to the right by 2 bits. For example, 100>>2 means binary number. 1.

Next, the four blocks obtained by the division are processed. To process the first of the four blocks first, you need to set the following input parameters: Brightness position (xB, yB) is (xB, yB) and block width logarithm indicates log2Trafo Width and height logarithm represent log2TrafoHeight. The specific setting method is as follows: when non-square_transform_flag is 1 and non-square_transform_direction is 1, log2TrafoHeight is set to log2TrafoHeight-2, log2Trafo Width remains unchanged; in non-square-transform-flag is 1 and non-square-transform-direction is 0, log2Trafo Width is set to log2TrafoWidth-2, log2TrafoHeight remains unchanged; when non-square_transform_flag is 0, log2Trafo Width and log2TrafoHeight are set to log2TrafoWidth-l; Set the hierarchical depth trafoDepth of the current block to trafoDepth+1; set the intra prediction mode intraPredMode to intraPredMode (ie, keep the intra prediction mode unchanged); set the color component index cldx to cldx (ie, keep the color component index unchanged).

Processing the second of the four blocks is similar to the method of processing the first of the four blocks described above, except that the luminance position (xB, yB) is set to (xB1, yB1). Processing the third of the four blocks is similar to the method of processing the first of the four blocks described above, except that the luminance position (xB, yB) is set to (xB2, yB2). Process the fourth of the four blocks. Method of processing the first of the four blocks with the above - - Similar, except that the brightness position (xB, yB) is set to (xB3, yB3).

Using the above input parameters, the steps of setting the position information of the four blocks and the steps of processing the four blocks can be iteratively performed.

In 330, a decoding process is performed on each of the plurality of square transform units obtained by the final split or each of the plurality of rectangular transform units obtained by the final split.

According to an embodiment of the present invention, the split flag corresponding to the current block among the one or more split marks indicates each square transform unit or rectangular transform unit obtained in the final split without further splitting the current block.

According to an embodiment of the present invention, the prediction signals of each square transform unit or rectangular transform unit are acquired according to the position, width and height of each square transform unit or rectangular transform unit and the intra prediction mode determined by the prediction unit in which it is located, Obtaining a reconstructed prediction difference signal of each square transform unit or a rectangular transform unit according to the position of each square transform unit or rectangular transform unit, the width and height of each square transform unit or rectangular transform unit, and the depth of the partition hierarchy. And, based on the prediction signal and the reconstructed prediction difference signal, a reconstruction signal of each square transformation unit or rectangular transformation unit is obtained.

According to an embodiment of the present invention, for example, when the split_transform_flag corresponding to the current block is 0, the current block is determined to be the final transform unit obtained by the split, and the following decoding process is performed. Here, the brightness is still taken as an example for description.

First set the size of the current block. At this time, the current block can be a square transform unit or a rectangular transform unit. The size of the current block can be expressed using (nW)x(nH), where nW represents the number of samples of the luminance component of the current block in the horizontal direction, and nH represents the number of samples of the luminance component of the current block along the vertical direction. When the current block is a square transform unit, nW is equal to nH. When the current block is a rectangular transform unit, nW and nH are not equal. nW and nH can be set by the following method: When non-square_transform_flag is 0, set nW and nH to (l«log2TrafoSize); otherwise, set nW to l«log2TrafoWidth, and set nH to l«log2TrafoHeight .

Second, get the prediction signal of the current block. The brightness position (xB, yB) of the current block, the intra prediction mode intraPredMode of the current block, the width nW and the height nH of the current block, and the color component cldx of the current block are taken as inputs, and the intra prediction module is called to acquire (nW)x. (nH) size prediction signal matrix.

Then, the current block reconstruction prediction difference signal is obtained. Current block brightness position (xB, yB), current - The block level depth trafoDepth, the current block color component cldx, the current block width nW and the height nH are taken as inputs, and the scaling and transform module is called to acquire a (nW)x(nH) size reconstructed difference signal matrix.

Next, get the current block reconstruction signal. A (nW)x(nH)-sized prediction signal matrix is added to the (nW)x(nH)-sized reconstructed difference signal matrix to obtain a reconstructed signal matrix of (nW)x(nH) size.

Finally, the current block reconstruction signal is placed in the corresponding position in the reconstructed image. The reconstructed signal matrix of (nW)x(nH) size is placed in the corresponding position marked by the luminance position (xB, yB) in the reconstructed image.

It should be noted that there are a number of ways to achieve the same execution results as the above decoding process, which are not enumerated here. In addition, the embodiments of the present invention can be used to process image luminance components, and can also be used to process image chrominance components, and can also be used to process image luminance components and chrominance components simultaneously.

It can be seen that the syntax of the intra prediction based decoding method can be modified according to an embodiment of the present invention, that is, the code stream structure is modified: the syntax element non-square_transform_flag and non-square-transform are added. — direction , and JU' falsified the corresponding decoding side parsing operation. In the decoding process, according to the non-square-transform-flag and non-square-transform-direction value selection processing, the corresponding processing flow can be realized not only to realize the rectangular transform unit, but also compatible with the quad-tree transform segmentation mode, and can be regarded as a quadtree A change in the way the transformation is split. Therefore, it is much simpler in terms of grammar description methods and implementation.

According to the intra prediction based method provided by the embodiment of the present invention, an adaptive map is adopted by selecting a prediction unit into a square transform unit or a rectangular transform unit in the same prediction unit.

The prediction unit division method of the rectangle is not used, so the computational complexity of the encoding end can be further reduced.

4 is a schematic flow chart showing a decoding method based on intra prediction according to another embodiment of the present invention. According to an embodiment of the present invention, a coding unit may be divided into a plurality of prediction units, and then a similar operation is performed for each prediction unit iteration.

As shown in FIG. 4, in 405, the first segmentation flag is obtained from the compressed code stream of the view image for the current prediction unit.

At 410, it is determined whether the first split flag is one. If yes, execute 415, otherwise, execute 460. - In 415, a split shape marker and a split direction marker are obtained from the compressed code stream of the view image for the current prediction unit.

In 420, it is determined whether the split shape flag indicates whether the current block is divided into a rectangular block or a square block. If it is split into rectangles, execute 425. If it is divided into square blocks, execute 440.

In 425, it is judged whether the division direction flag indicates that the current block is divided in the horizontal direction or the vertical direction. If it is divided in the horizontal direction, 430 is performed. If it is divided in the vertical direction, 435 is performed.

In 430, the current block is divided into rectangular blocks in the horizontal direction, and execution continues to 450.

In 435, the current block is divided into rectangular blocks in the vertical direction, and execution continues with 455.

In 440, the current block is split into square blocks and execution continues with 445.

In 445, it is judged whether or not the division flag corresponding to the divided square block is 1. If yes, execute

465, otherwise execute 460.

At 450, it is judged whether or not the division target corresponding to the rectangular block divided in the horizontal direction is 1. If yes, execute 470, otherwise, execute 460.

In 455, it is judged whether or not the division flag corresponding to the rectangular block divided in the vertical direction is 1. If yes, execute 470, otherwise, execute 460.

In 460, decoding is performed on the current block and execution continues at 475.

In 465, the divided square block is taken as the current block, and execution proceeds to 420.

In 470, the divided rectangular block is taken as the current block, and execution proceeds to 420.

In 475, it is determined whether the current block is the last one of the prediction units. If yes, execute 485, otherwise, execute 480.

In 480, the next image block is taken as the front block, and execution continues with 420.

The next prediction unit is taken as the current prediction unit in 485, and step 405 is continued. This decoding process can be performed on each prediction unit in turn. Certainly, the embodiment of the present invention is not limited thereto, and the foregoing decoding process may be performed on multiple prediction units at the same time. In this case, when the determination result of 475 is that the current block is the last one of the prediction units, the current prediction is ended. The decoding process of the unit.

It should be noted that the implementation according to the present invention is not limited to the above-described execution order. For example, the decoding operation may be performed simultaneously after all the transform units are obtained by dividing the prediction unit. In addition, in the case where the size of the prediction unit is the same as the size of the coding unit, 485 can be omitted.

In an embodiment in accordance with the invention, the prediction unit is segmented into two by selecting within the same prediction unit - A square transform unit or a rectangular transform unit, whereby an optional transform unit (ie, a rectangular transform unit) adapted to the strip texture of the image frame is used to process the strip texture, which can improve the video image compression efficiency. Since the prediction unit division manner (i.e., the prediction unit division manner without using a rectangle) is reduced according to an embodiment of the present invention, the coding side computational complexity can be further reduced.

FIG. 5 is a schematic configuration diagram showing a decoding apparatus 500 based on intra prediction according to an embodiment of the present invention. The apparatus 500 of FIG. 5 includes a tag acquisition module, a segmentation module 520, and a decoding module 530.

The tag obtaining module 510 is configured to obtain, for the current prediction unit, one or more segmentation marks, a segmentation shape marker, and a segmentation direction marker from the compressed code stream of the video image, where the one or more segmentation markers correspond to the current block. The split flag is used to indicate whether to divide the current block into a plurality of blocks, the split shape flag indicates that the current block is divided into a plurality of square blocks or is divided into a plurality of rectangular blocks, and the split direction mark indicates that the current block is divided into horizontal blocks into A plurality of rectangular blocks are divided into a plurality of rectangular blocks in a vertical direction, and each of the plurality of square blocks has a width equal to a height, and each of the plurality of rectangular blocks has a width and a height that are not equal.

The segmentation module 520 is configured to finally divide the current prediction unit into a plurality of square transformation units or a plurality of rectangular transformation units according to the acquired one or more segmentation markers, the segmentation shape markers, and the segmentation direction markers.

The decoding module 530 is configured to perform a decoding process on each of the plurality of square transform units or the plurality of rectangular transform units that are finally divided.

According to an embodiment of the invention, the current block is an image block of the same size as the current prediction unit. According to an embodiment of the present invention, one of the plurality of divided blocks is taken as the current block to be divided next.

According to an embodiment of the present invention, the mark obtaining module 510, after the first one of the one or more split marks indicates that the current block is divided into a plurality of blocks, after acquiring the first split mark, Obtaining a split shape mark in the compressed code stream; and in the case where the split shape mark indicates that the current block is divided into a plurality of rectangular blocks, after the split shape mark is acquired, one split direction mark is acquired from the compressed code stream.

According to an embodiment of the present invention, the segmentation module 520 marks the segmentation direction according to the segmentation direction when the segmentation flag corresponding to the current block in the one or more segmentation marks indicates that the current block is divided into a plurality of blocks. - - The corresponding split direction divides the current block into a plurality of square blocks or rectangular blocks corresponding to the split shape marks.

According to an embodiment of the present invention, the decoding module 530, in the case that the segmentation flag corresponding to the current block in the one or more segmentation marks indicates that the current block is not further segmented, each square transformation unit obtained by the final segmentation Or a rectangular transformation unit.

According to an embodiment of the present invention, the decoding unit 530 acquires each square transform unit or rectangular transform unit according to the position, width and height of each square transform unit or rectangular transform unit and the intra prediction mode determined by the prediction unit in which it is located. The prediction signal is obtained according to the position of each square transformation unit or rectangular transformation unit, the width and height of each square transformation unit or rectangular transformation unit, and the depth of the division level, and the reconstruction prediction of each square transformation unit or rectangular transformation unit is obtained. And a difference signal, and obtaining a reconstructed signal of each square transform unit or rectangular transform unit according to the predicted signal and the reconstructed predicted difference signal.

The above and other operations and/or functions of the tag acquisition module 510, the segmentation module 520, and the decoding unit 530 of the acknowledgment device 500 provided by the embodiment of the present invention may refer to the methods 310, 320, and 330 of FIG. 3 above, in order to avoid duplication, This will not be repeated here.

According to the decoding apparatus based on the intra prediction provided by the embodiment of the present invention, by selecting the prediction unit into a square transformation unit or a rectangular transformation unit in the same prediction unit,

(That is, the prediction unit division method of the rectangle is not used), so the calculation end complexity can be further reduced.

Those of ordinary skill in the art will appreciate that the elements and algorithm steps of the various examples described in connection with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of both, for clarity of hardware and software. Interchangeability, the composition and steps of the various examples have been generally described in terms of function in the above description. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the solution. A person skilled in the art can use different methods for implementing the described functions for each particular application, but such implementation should not be considered to be beyond the scope of the present invention.

A person skilled in the art can clearly understand that for the convenience and brevity of the description, the specific working process of the system, the device and the unit described above can refer to the corresponding process in the foregoing method embodiment. - - I won't go into details here.

In the several embodiments provided herein, it should be understood that the disclosed systems, devices, and methods may be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the segmentation of the cells is only a logical function segmentation. In actual implementation, there may be another segmentation manner, for example, multiple cells or components may be combined or Can be integrated into another system, or some features can be ignored, or not executed. In addition, the coupling or direct coupling or communication connection shown or discussed may be an indirect coupling or communication connection through some interface, device or unit, and may be electrical, mechanical or otherwise.

The units described as separate components may or may not be physically separate, and the components displayed as the units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the objectives of the solution of the embodiment.

In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

The integrated unit, if implemented in the form of a software functional unit and sold or used as a standalone product, may be stored in a computer readable storage medium. Based on such understanding, the technical solution of the present invention may contribute to the prior art or all or part of the technical solution may be embodied in the form of a software product stored in a storage medium. A number of instructions are included to cause a computer device (which may be a personal computer, server, or network device, etc.) to perform all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: a U disk, a removable hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and the like, which can store program codes. .

The above is only a specific embodiment of the present invention, but the scope of the present invention is not limited thereto, and any person skilled in the art can easily think of changes or substitutions within the technical scope of the present invention. It should be covered by the scope of the present invention. Therefore, the scope of the invention should be determined by the scope of the claims. .

Claims

Rights request

A decoding method based on intra prediction, comprising:

Obtaining, for the current prediction unit, one or more segmentation marks, a segmentation shape marker, and a segmentation direction marker from the compressed code stream of the video image, wherein the segmentation marker corresponding to the current block of the one or more segmentation markers is used to indicate Whether to divide the current block into a plurality of blocks, the one split shape mark indicates dividing the current block into a plurality of square blocks or dividing into a plurality of rectangular blocks, the one split direction mark indicating that the current block is to be Dividing into the plurality of rectangular blocks in a horizontal direction or dividing into a plurality of rectangular blocks in a vertical direction, each of the plurality of square blocks having a width equal to a height, wherein the plurality of rectangular blocks are Each rectangular block is not equal in width and height;

And dividing the current prediction unit into a plurality of square transformation units or a plurality of rectangular transformation units according to the acquired one or more segmentation marks, the one segmentation shape marker, and the one segmentation direction marker;

A decoding process is performed on each of the plurality of square transform units obtained in the final split or each of the plurality of rectangular transform units finally divided.

2. The intra prediction based decoding method according to claim 1, wherein the current block is an image block of the same size as the current prediction unit.

3. The intra prediction based decoding method according to claim 1, wherein one of the plurality of blocks obtained by the segmentation is used as a current block to be divided next.

The intra prediction-based decoding method according to claim 1, wherein each of the plurality of rectangular blocks obtained by dividing in a horizontal direction has a width greater than a height, and the edge The width of each of the plurality of rectangular blocks obtained by dividing in the vertical direction is smaller than the height.

5. The intra prediction based decoding method according to claim 1, wherein: in a case where the first one of the one or more divided marks indicates that the current block is divided into a plurality of blocks After obtaining the first segmentation flag, acquiring the one segmentation shape marker from the compressed code stream;

In a case where the one divided shape flag indicates that the current block is divided into a plurality of rectangular blocks, the one divided direction mark is acquired from the compressed code stream after acquiring the one divided shape mark.

The intra prediction based decoding method according to claim 1, wherein the one or more segmentation marks, the one segmentation shape flag, and the one segment are obtained according to the acquired Cutting the direction mark to divide the current prediction unit into a plurality of square transformation units or a plurality of rectangular transformation units includes:

And in a case where the split flag corresponding to the current block in the one or more split marks indicates that the current block is divided into multiple blocks, the current block is performed according to a split direction corresponding to the split direction mark It is divided into a plurality of square blocks or rectangular blocks corresponding to the divided shape marks.

The intra prediction based decoding method according to claim 6, wherein the current block is divided into a plurality of square blocks corresponding to the split shape mark according to the split direction corresponding to the split direction mark Or rectangular blocks include:

In a case where the one split shape flag indicates that the current block is divided into the plurality of square blocks, the current block is divided into four equal-sized square blocks, wherein the four equal-sized square blocks are The width or height of each square block is respectively two-half of the width or height of the current block, in the case where the one split shape mark indicates that the current block is divided into the plurality of rectangular blocks, according to the A split direction mark further divides the current block into four equal-sized rectangular blocks.

The intra prediction-based decoding method according to claim 7, wherein, in a case where the one divided shape flag indicates that the current block is divided into a plurality of rectangular blocks, according to the one segmentation The direction marker divides the current block into four equal-sized rectangular blocks including:

In a case where the one partitioning direction flag indicates that the current block is divided into the plurality of rectangular blocks in the horizontal direction, the current block is divided into four equal-sized rectangular blocks in a horizontal direction, wherein the four The width of each of the equal-sized rectangular blocks is the same as the width of the current block, and the height of each of the four equally-sized rectangular blocks is the height of the current block. One of the points;

In a case where the one split direction flag indicates that the current block is divided into the plurality of rectangular blocks in the vertical direction, the current block is divided into four equal-sized rectangular blocks in the vertical direction, where The height of each of the four equal-sized rectangular blocks is the same as the height of the current block, and the width of each of the four equally-sized rectangular blocks is the width of the current block One quarter of the.

The intra prediction-based decoding method according to claim 1, wherein each of the plurality of square transform units obtained by the final splitting or the final splitting is obtained. Each of the plurality of rectangular transform units performs a decoding process:

And performing, in the case where the segmentation flag corresponding to the current block in the one or more segmentation marks indicates that the current block is not further segmented, performing a decoding process on each of the square transformation units or the rectangular transformation unit obtained by the final segmentation .

10. The intra prediction-based decoding method according to claim 9, wherein the frame is determined according to the position, the width and the height of each of the square transform units or the rectangular transform unit, and the prediction unit determined by the prediction unit in which it is located. a mode, obtaining a prediction signal of each of the square transformation units or the rectangular transformation unit, and acquiring each of the square transformations according to a position, a width, and a height of each of the square transformation units or the rectangular transformation units and a depth of the division level Reconstructing the difference signal of the unit or the rectangular transform unit, and obtaining a reconstructed signal of each of the square transform units or the rectangular transform unit according to the prediction signal and the reconstructed prediction difference signal.

11. A decoding apparatus based on intra prediction, comprising:

a mark obtaining module, configured to acquire one or more segmentation marks, a segmentation shape mark, and a segmentation direction mark from the compressed code stream of the video image for the current prediction unit, where the one or more segmentation marks correspond to the current block The segmentation flag is used to indicate whether the current block is divided into a plurality of blocks, and the one segmentation shape flag indicates that the current block is divided into a plurality of square blocks or divided into a plurality of rectangular blocks, and the one segmentation direction flag is Instructing to divide the current block into the plurality of rectangular blocks in a horizontal direction or to divide the plurality of rectangular blocks in a vertical direction, and each of the plurality of square blocks has a width equal to a height, Each of the plurality of rectangular blocks is unequal in width and height; the dividing module is configured to: according to the acquired one or more divided marks, the one divided shape mark, and the one divided direction mark The current prediction unit is finally divided into a plurality of square transformation units or a plurality of rectangular transformation units;

And a decoding module, configured to perform a decoding process on each of the plurality of square transform units obtained by the final splitting or each of the plurality of rectangular transform units obtained by the final splitting.

12. The intra prediction based decoding apparatus according to claim 11, wherein the current block is an image block of the same size as the current prediction unit.

The intra prediction-based decoding apparatus according to claim 11, wherein one of the plurality of blocks obtained by the division is used as a current block to be divided next.

The intra prediction-based decoding apparatus according to claim 11, wherein the first acquisition flag of the one or more segmentation marks by the mark acquisition module indicates that the In a case where the current block is divided into a plurality of blocks, after the first segmentation flag is acquired, the one segmentation shape flag is acquired from the compressed code stream; and the one segmentation shape flag indicates that the current In the case where the block is divided into a plurality of rectangular blocks, the one divided direction mark is acquired from the compressed code stream after the one divided shape mark is acquired.

The intra prediction-based decoding apparatus according to claim 14, wherein the segmentation module indicates, in the one or more segmentation marks, a segmentation mark corresponding to the current block to indicate the current When the block is divided into a plurality of blocks, the current block is divided into a plurality of square blocks or rectangular blocks corresponding to the divided shape marks in accordance with the division direction corresponding to the division direction marks.

The intra prediction-based decoding apparatus according to claim 11, wherein the decoding module indicates, in the one or more split marks, a split flag corresponding to the current block, that the current In the case where the block is further divided, the decoding process is performed for each square transform unit or rectangular transform unit finally obtained.

The intra prediction-based decoding apparatus according to claim 16, wherein the position, the width and the height of each of the square transform units or the rectangular transform units and the intra prediction determined by the prediction unit in which the unit is located are a mode, obtaining a prediction signal of each of the square transformation units or the rectangular transformation unit, and acquiring each of the square transformations according to a position, a width, and a height of each of the square transformation units or the rectangular transformation units and a depth of the division level Reconstructing the difference signal of the unit or the rectangular transform unit, and obtaining a reconstructed signal of each of the square transform units or the rectangular transform unit according to the prediction signal and the reconstructed prediction difference signal.