JP2018107692A - Moving image decoder, moving image decoding method, moving image encoder, moving image encoding method and computer readable recording medium - Google Patents

Abstract

PROBLEM TO BE SOLVED: To improve prediction accuracy of intra-prediction.SOLUTION: A moving image decoder 102 comprises: an entropy decoding section 602 for decoding a bit stream showing a moving image including a block encoded by using intra-prediction; an inverse quantization/inverse transformation section 604 for inversely-quantizing and inversely transforming a level value of the block, which is obtained as a result of decoding, and generating a differential value of the block; an intra-prediction section 610 for generating a prediction value of the block by intra-prediction based on a reconstructed block around the block; and a reconstruction section 606 for reconstructing the block on the basis of the generated differential value and the generated prediction value. The intra-prediction section generates the prediction value of a color difference component of the block by using a non-linear model between a luminance component and the color difference component of the block.SELECTED DRAWING: Figure 6

Description

  The present invention relates to a video decoding device, a video decoding method, a video encoding device, a video encoding method, and a computer-readable recording medium.

  A moving picture coding scheme using intra prediction or inter prediction, residual transform, and entropy coding has been proposed. By using intra prediction, the spatial redundancy of the image can be reduced. The basic mechanism of intra prediction is as follows. First, a reference line adjacent to the prediction block (upper and left) is generated, and the original sample in the prediction block is projected onto the reference line along the selected prediction direction. Thereafter, the difference (residual error) between the sample of the reference line and the original sample projected thereon is transformed and quantized.

  In general, intra prediction is performed on each of the luminance component and the color difference component of the block to be processed. Patent Document 1 discloses a technique for performing intra prediction of a color difference component using a linear model between a luminance component and a color difference component.

JP 2014-195142 A

Siwei Ma, Shiqi Wang, Shanshe Wang, Liang Zhao, Qin Yu, Wen Gao, "Low Complexity Rate Distortion Optimization for HEVC", 2013 Data Compression Conference.

  According to the technique described in Patent Document 1, it is possible to realize good intra prediction. However, it is desired to further improve the accuracy of intra prediction without being limited to this.

  The present invention has been made in view of these problems, and an object thereof is to provide a technique capable of improving the accuracy of intra prediction.

  One embodiment of the present invention relates to a moving picture decoding apparatus. This moving image decoding apparatus includes a decoding unit that decodes a bit stream representing a moving image including a block encoded using intra prediction, and performs inverse quantization and inverse conversion on a block level value obtained as a result of decoding. Inverse processing means for generating a difference value of a block, intra prediction means for generating a prediction value of a block by intra prediction based on a reconstructed block around the block, and based on the generated difference value and the generated prediction value Reconstructing means for reconstructing the block. The intra prediction unit generates a predicted value of the color difference component of the block using a nonlinear model between the luminance component and the color difference component of the block.

  Another aspect of the present invention is a video encoding apparatus. The moving image encoding device generates intra block prediction values based on intra prediction based on encoded blocks around a moving image block, and generates block difference values based on the generated prediction values. Difference generating means, processing means for converting and quantizing the generated difference value to generate a level value, and encoding means for encoding the generated level value to generate a bitstream. The intra prediction unit generates a predicted value of the color difference component of the block using a nonlinear model between the luminance component and the color difference component of the block.

  It should be noted that any combination of the above-described constituent elements, or those obtained by replacing the constituent elements and expressions of the present invention with each other between apparatuses, methods, systems, computer programs, recording media storing computer programs, and the like are also included in the present invention. It is effective as an embodiment of

  According to the present invention, the accuracy of intra prediction can be increased.

It is a schematic diagram which shows the structure of the delivery system which concerns on embodiment. It is a block diagram which shows the function and structure of the moving image encoder of FIG. It is a flowchart which shows the flow of a series of processes in the intra estimation part of FIG. It is a flowchart which shows the flow of a series of processes in CCNM mode. FIGS. 5A and 5B are schematic diagrams showing the processing target Y block and the processing target Cb block, respectively. It is a block diagram which shows the function and structure of the moving image decoding apparatus of FIG. It is a flowchart which shows the flow of a series of processes in the intra estimation part of FIG. FIGS. 8A and 8B are schematic diagrams illustrating reference lines according to the first modification selected for the processing target Y block and the processing target Cb block. FIGS. 9A and 9B are schematic diagrams for explaining weighting when the reference line extends over a plurality of blocks. FIGS. 10A and 10B are schematic diagrams for explaining weighting when a plurality of reference lines are selected from one block. It is a flowchart which shows the flow of a series of processes in the intra estimation part which concerns on a 2nd modification. 12A and 12B are explanatory diagrams of a prediction model according to the fourth modification.

  Hereinafter, the same or equivalent components, members, and processes shown in the drawings are denoted by the same reference numerals, and repeated description is appropriately omitted. In addition, in the drawings, some of the members that are not important for explanation are omitted.

  In the embodiment, in intra-prediction (intra-prediction) of video encoding / decoding, a luminance component (luma component) corresponding to a predicted value of a chroma component (chroma component) is used. Calculated from a nonlinear model as input. At this time, the parameters of the nonlinear model are calculated based on the reference lines of the luminance component and chrominance component of the encoded / reconstructed block around the prediction target block. Accordingly, it is possible to realize intra prediction of the color difference component with higher accuracy while suppressing an increase in the amount of data to be signaled.

  FIG. 1 is a schematic diagram illustrating a configuration of a distribution system 100 according to an embodiment. The distribution system 100 is a system used in a moving image distribution service that distributes moving images. The moving image distribution service may be VOD (Video On Demand), for example. The distribution system 100 is connected to a moving image decoding apparatus 102 such as a set top box or a personal computer installed at a user site via a network 106 such as the Internet. The moving picture decoding apparatus 102 is connected to a display apparatus 104 such as a television receiver or a monitor. The moving image decoding apparatus 102 and the display apparatus 104 may together constitute a mobile terminal such as a smartphone.

  It should be noted that the distribution system in the moving image distribution service is an example, and that the technical idea according to the present embodiment can be applied to any system or service including encoding or decoding of a moving image is described in this specification. It will be apparent to those skilled in the art.

  The distribution system 100 receives designation of moving image content desired to be viewed from the user via the network 106. The distribution system 100 encodes the data of the specified moving image content to generate a bit stream BS. The distribution system 100 transmits the generated bit stream BS to the video decoding device 102 of the requesting user via the network 106. The moving picture decoding apparatus 102 decodes the received bit stream BS to generate moving picture data, and transmits it to the display apparatus 104. The display device 104 processes the received moving image data and outputs designated moving image content.

  The distribution system 100 includes a moving image DB (database) 108 and a moving image encoding device 110. The moving image DB 108 holds moving image data. The moving image encoding device 110 acquires moving image data corresponding to the specified moving image content from the moving image DB 108, encodes the acquired moving image data, and generates a bit stream BS representing the moving image.

  FIG. 2 is a block diagram showing the function and configuration of the moving picture encoding apparatus 110 of FIG. Each block shown in the block diagram of the present specification can be realized in terms of hardware by an element such as a CPU of a computer or a mechanical device, and in terms of software, it can be realized by a computer program or the like. The functional block realized by those cooperation is drawn. Therefore, those skilled in the art who have touched this specification will understand that these functional blocks can be realized in various forms by a combination of hardware and software. The computer program for realizing the moving image encoding device 110 may be stored in a computer-readable recording medium or distributed via a network.

  The video encoding device 110 includes a frame buffer 202, an in-loop filter 204, an inter prediction unit 206, an intra prediction unit 208, a transform / quantization unit 210, an entropy encoding unit 212, an inverse quantization / An inverse conversion unit 214, a subtraction unit 218, an addition unit 220, and a block division unit 222 are provided.

  The block dividing unit 222 divides a picture to be encoded included in moving image data (video data) from the moving image DB 108 into a plurality of blocks. The block sizes vary, and the plurality of blocks have a quadtree structure. The plurality of blocks may have a binary tree structure. The division of the picture into blocks in the block dividing unit 222 is based on the division of the picture into blocks in High Efficiency Video Coding (HEVC). The block division unit 222 outputs the processing target block to the subtraction unit 218, the inter prediction unit 206, and the intra prediction unit 208.

  A reference picture corresponding to the processing target block is input from the frame buffer 202 to the inter prediction unit 206. The inter prediction unit 206 outputs a motion compensated image of the block to be processed by inter prediction (inter-prediction, inter-screen prediction) based on the input reference picture.

  The intra prediction unit 208 receives image data of an encoded block (a block that has already been encoded) around the processing target block from the addition unit 220. The intra prediction unit 208 generates and outputs a prediction value of the pixel of the processing target block by intra prediction based on the pixel value of the input encoded block. Either the output of the inter prediction unit 206 or the output of the intra prediction unit 208 is output to the subtraction unit 218 depending on whether the inter prediction or the intra prediction is applied to the processing target block.

  The pixel value includes a luminance component and a color difference component. In this embodiment, a case where Y is used as a luminance component, Cb and Cr are used as color difference components, and 4: 2: 0 is used as a color format will be described. However, the technical idea according to the present embodiment can be applied even when other luminance components / chrominance components and other color formats (for example, 4: 4: 4, 4: 2: 2) are adopted. It will be apparent to those skilled in the art who have touched the specification.

  Hereinafter, a block composed of the luminance component (Y) of the processing target block is referred to as a processing target Y block, and a block composed of two color difference components (Cb, Cr) is referred to as a processing target Cb block and a processing target Cr block, respectively. Since the color format is 4: 2: 0, if the size of the processing target Y block is 16 × 16 pixels, the size of the corresponding processing target Cb block (processing target Cr block) becomes 8 × 8 pixels by downsampling. Smaller than the size of the Y block to be processed.

  The subtraction unit 218 performs subtraction between the actual pixel value of the pixel and the corresponding prediction value generated by inter prediction or intra prediction for each of the processing target Y block, processing target Cb block, and processing target Cr block. To generate a pixel difference value (prediction residual signal). The subtraction unit 218 outputs the generated difference value to the transform / quantization unit 210.

  The transform / quantization unit 210 transforms (for example, orthogonal transform) and quantizes the pixel difference value of the processing target block acquired from the subtraction unit 218, and generates a level value. The transform / quantization unit 210 outputs the generated level value to the entropy coding unit 212 and the inverse quantization / inverse transform unit 214.

  The entropy encoding unit 212 entropy-encodes the level value generated by the transform / quantization unit 210 and the side information acquired from the inter prediction unit 206 and the intra prediction unit 208 to generate a bitstream BS. .

  Note that side information is information necessary for reconstruction of pixel values used in the decoding device, a prediction mode indicating whether intra prediction or inter prediction is used, motion information in inter prediction, It includes related information such as quantization parameters and block size.

  As a prediction mode indicating the type of intra prediction of the luminance component, 35 types of intra prediction modes (intra prediction modes) defined by HEVC are adopted. The 35 types of intra prediction modes include a Planar mode, a DC mode, and 33 types of directional modes. Each of the 33 types of directional modes has a different prediction direction.

  As prediction modes indicating the types of intra prediction of color difference components, six types obtained by adding the CCNM (Cross Component Nonlinear Prediction Model) mode according to the present embodiment to the five types of intra prediction modes defined by HEVC are adopted. The five types of existing intra prediction modes include a Planar mode, a DC mode, two types of directional modes (Horizontal, Vertical), and a Delivered / Upright mode.

The inverse quantization / inverse transform unit 214 performs a process opposite to the process in the transform / quantization unit 210 to generate a pixel difference value of the processing target block.
The adder 220 adds the difference value output from the inverse quantization / inverse transform unit 214 and the corresponding prediction value generated by inter prediction or intra prediction to reconstruct the pixel value of the pixel of the processing target block To do. The adding unit 220 outputs the reconstructed block to the intra prediction unit 208 and the in-loop filter 204.

  The in-loop filter 204 generates a local decoded image corresponding to the processing target frame and outputs it to the frame buffer 202. The frame buffer 202 holds a locally decoded image. This locally decoded image is used for inter prediction in the inter prediction unit 206.

  FIG. 3 is a flowchart showing a flow of a series of processes in the intra prediction unit 208 of FIG. The intra prediction unit 208 acquires a processing target block from the block dividing unit 222 (S302). The intra prediction unit 208 determines an intra prediction mode to be applied to the luminance component of the processing target block based on the intra prediction based on the encoded blocks around the acquired processing target block, and the determined intra prediction. A predicted value is generated and output in accordance with the mode (S304). The luminance component intra prediction in step S304 may be realized by HEVC intra prediction. The intra prediction unit 208 generates side information including the luminance component intra prediction mode determined in step S <b> 304 and outputs the side information to the entropy encoding unit 212.

  The intra prediction unit 208 calculates an RD cost when the color difference component of the processing target block is predicted in each of the six types of intra prediction modes including the CCNM mode (S306). The RD cost may be a rate-distortion cost used in the rate-distortion optimization described in Non-Patent Document 1, for example. The intra prediction unit 208 determines an intra prediction mode to be applied to the color difference component of the processing target block based on the calculated RD cost (S308). For example, the intra prediction unit 208 selects an intra prediction mode that minimizes the calculated RD cost as an intra prediction mode to be applied from among six types of intra prediction modes. The intra prediction unit 208 generates and outputs the predicted value of the color difference component of the processing target block according to the determined intra prediction mode (S310). After the luminance component side information is generated as described above, the intra prediction unit 208 generates side information including the color difference component intra prediction mode determined in step S308, and outputs the side information to the entropy encoding unit 212 (S310). ).

  The CCNM mode used in step S306 in FIG. 3 will be described in more detail. In the CCNM mode, the intra prediction unit 208 generates a prediction value of the processing target Cb block using a nonlinear model between the processing target Y block and the processing target Cb block. Since the same applies to the processing target Cr block, generation of the prediction value of the processing target Cb block will be described below, and description of generation of the prediction value of the processing target Cr block will be omitted.

  FIG. 4 is a flowchart showing a flow of a series of processes in the CCNM mode. FIGS. 5A and 5B are schematic diagrams illustrating the processing target Y block 502 and the processing target Cb block 504, respectively. In this example, both the processing target block and the encoded blocks around it are 16 × 16 pixels. Therefore, the processing target Y block 502 has a size of 16 × 16 pixels, and the processing target Cb block 504 has a size of 8 × 8 pixels. It should be noted that symmetrical processing target blocks of other sizes such as 32 × 32 pixels and 8 × 8 pixels, and asymmetric processing such as 16 × 8 pixels, 16 × 4 pixels, 8 × 16 pixels, and 8 × 32 pixels. It is obvious to those skilled in the art who have touched this specification that the technical idea according to the present embodiment can be applied to the target block.

  In the example shown in FIGS. 5A and 5B, there are five encoded blocks around the block to be processed, each of which is a Y block (hereinafter referred to as an encoded Y block) and a Cb block. (Hereinafter referred to as an encoded Cb block).

  The picture is encoded zigzag horizontally from the upper left block. Therefore, around the processing target Y block 502, a second encoded Y block 508 that is in contact with the processing target Y block 502, a fourth encoded Y block 512 that is in contact with the left of the processing target Y block 502, and The first encoded Y block 506 that touches the left of the second encoded Y block 508, the third encoded Y block 510 that touches the right of the second encoded Y block 508, and the fourth encoded Y There is a fifth encoded Y block 514 that borders below the block 512. Around the processing target Cb block 504, there are a second encoded Cb block 518 in contact with the processing target Cb block 504, a fourth encoded Cb block 522 in contact with the left of the processing target Cb block 504, and a second The first encoded Cb block 516 that touches the left of the encoded Cb block 518, the third encoded Cb block 520 that touches the right of the second encoded Cb block 518, and the fourth encoded Cb block 522 And a fifth encoded Cb block 524 that touches the bottom.

  The intra prediction unit 208 selects a reference line from a plurality of lines included in the encoded block around the processing target block (S402). The intra prediction unit 208 includes a line adjacent to the processing target Cb block 504 among lines extending in the vertical direction included in the fourth encoded Cb block 522, and a line of the fifth encoded Cb block 524 connected to the line. , The left adjacent Cb reference line 526 is selected. The intra prediction unit 208 includes a line adjacent to the processing target Cb block 504 among lines extending in the left-right direction included in the second encoded Cb block 518, and a line of the third encoded Cb block 520 connected to the line. , And the adjacent Cb reference line 528 is selected.

  When the fifth encoded Cb block 524 is not usable during intra prediction, the pixel values of the pixels that are not usable in the left adjacent Cb reference line 526 are the lines of the fourth encoded Cb block 522. It is configured by duplicating the pixel value of the pixel at the lower end of. The same applies when the third encoded Cb block 520 is not available.

  The intra-prediction unit 208 generates the left adjacent Y reference line 530 of the encoded Y blocks 506, 508, 510, 512, 514 corresponding to the selected left adjacent Cb reference line 526 and the upper adjacent Cb reference line 528, respectively. An adjacent Y reference line 532 is identified. The left adjacent Y reference line 530 extends over the fourth encoded Y block 512 and the fifth encoded Y block 514, and the upper adjacent Y reference line 532 includes the second encoded Y block 508 and the third encoded Y block 508. A line straddling the encoded Y block 510.

  The intra prediction unit 208 refers to the reconstruction value of the processing target Y block 502 that has already been generated by the intra prediction of the luminance component executed in step S304, and the left adjacent Y reference line 530 and the upper adjacent Y reference specified in step S402. The line 532 is down-sampled (S404). This downsampling is performed at a rate corresponding to the color format of the moving image. For example, in the examples of FIGS. 5A and 5B, downsampling is performed to make 2 × 2 pixels (= 4 pixels) of the Y block one pixel, which corresponds to the color format 4: 2: 0. . In FIG. 5A, a circle 534 hatched with a diagonal line from the upper right to the lower left is one pixel in which the corresponding 2 × 2 pixel region 536 has a pixel value obtained by interpolation of the reconstructed values of those pixels. Indicates down-sampling. A white circle 538 indicates that the corresponding 3 × 2 pixel region 540 is down-sampled to one pixel having a pixel value obtained by interpolation of the reconstruction values of those pixels. In other embodiments, downsampling is not performed when the size of the luminance component and the size of the color difference component are the same.

  The intra prediction unit 208 moves from the upper right to the lower left, the pixel value of the pixel (the white circle 538) of the processing target Y block 502 obtained as a result of the downsampling, the left adjacent Y reference line 530, and the upper adjacent Y reference line 532. A circle 534 hatched with diagonal lines is filtered (S406). The filtering is a process for increasing the degree of matching between the color difference component and the luminance component after downsampling, and may be a smoothing process, for example. The filter used in this smoothing process may be a low-pass filter such as a Gaussian filter. An example of the filter coefficient is (1/4, 1/2, 1/4), but is not limited to this.

  Alternatively, when the intra prediction mode of the processing target Y block 502 is neither the Planar mode nor the DC mode, filtering by a directional filter may be performed. The direction of the directional filter may be determined based on the prediction direction of the directional mode applied to the processing target Y block 502. When the intra prediction mode of the processing target Y block 502 is either the Planar mode or the DC mode, filtering by a non-directional filter (non-directional filter) may be performed. In this way, the filter to be applied may be changed according to the presence / absence of the prediction direction and the angle in the intra prediction of the luminance component.

…（式１）
ここで、

は、多項式モデルのパラメータである。Ｎは３以上の自然数であり、予め定められた同じ値が符号化装置および復号装置の両方に保持される。あるいはまた、符号化装置は適応的にＮの値を決定し、ビットストリームに含めて復号装置に伝送してもよい。なお、Ｎ＝２は線形モデルに対応する。Ｎの値を増やすと予測の確度は向上するが、モデルはより複雑となり、演算量が増大する。したがって、Ｎの値を、例えば予測の確度の向上による影響と、演算量の増大による影響と、が拮抗する値、例えば３に設定してもよい。 The intra prediction unit 208 generates a nonlinear model based on the left adjacent Y reference line 530 and the upper adjacent Y reference line 532 obtained as a result of the filtering, and the left adjacent Cb reference line 526 and the upper adjacent Cb reference line 528 selected in step S402. The parameter is determined (S408). In the present embodiment, the non-linear model is a model in which the color difference component is represented by a corresponding luminance component polynomial. When c is the value of the Cb component and y is the value of the corresponding Y component, c is expressed by the following formula 1 as a polynomial of y.

... (Formula 1)
here,

Are parameters of the polynomial model. N is a natural number of 3 or more, and the same predetermined value is held in both the encoding device and the decoding device. Alternatively, the encoding apparatus may adaptively determine the value of N and include it in the bitstream and transmit it to the decoding apparatus. N = 2 corresponds to a linear model. Increasing the value of N increases the accuracy of prediction, but the model becomes more complex and the amount of computation increases. Therefore, for example, the value of N may be set to a value, for example, 3, in which the influence due to the improvement in the accuracy of prediction and the influence due to the increase in the calculation amount are competing.

…（式２）
ここで、

Ｍはパラメータの評価に用いられるサンプルの総数であり、図５（ａ）、（ｂ）の例では左隣接Ｃｂ参照ライン５２６および上隣接Ｃｂ参照ライン５２８に含まれる画素の総数（＝３２）である。式２を解くことにより、パラメータベクトルＡは以下の式３により求められる。

…（式３）
イントラ予測部２０８は、処理対象Ｃｂブロック５０４の参照ラインおよび処理対象Ｙブロック５０２の参照ラインから得られる再構成値を式３に代入することでパラメータベクトルＡを算出する。その際、処理対象Ｃｂブロック５０４の参照ラインの画素（例えば、丸５４２の画素）の再構成値と、処理対象Ｙブロック５０２の参照ラインの対応するダウンサンプリング後の画素（例えば、丸５３４の画素）の再構成値と、が測定値の組（ｃ_ｊ、ｙ_ｊ）とされる。 In step S408, the intra prediction unit 208 performs linear regression analysis using a least squares method in order to determine the parameter a _i . The above polynomial model can be rewritten by Equation 2 below.

... (Formula 2)
here,

M is the total number of samples used for parameter evaluation. In the examples of FIGS. 5A and 5B, the total number of pixels included in the left adjacent Cb reference line 526 and the upper adjacent Cb reference line 528 (= 32). is there. By solving Equation 2, the parameter vector A is obtained by Equation 3 below.

... (Formula 3)
The intra prediction unit 208 calculates the parameter vector A by substituting the reconstructed values obtained from the reference line of the processing target Cb block 504 and the reference line of the processing target Y block 502 into Equation 3. At that time, the reconstructed value of the pixel of the reference line of the processing target Cb block 504 (for example, the pixel of the circle 542) and the corresponding downsampled pixel (for example, the pixel of the circular 534) of the reference line of the processing target Y block 502 ) Is a set of measurement values (c _j , y _j ).

In the example of FIGS. 5A and 5B, the reconstructed Cb values of 16 pixels included in the left adjacent Cb reference line 526 are c _{0 to} c ₁₅ from the bottom, and the left adjacent Y reference line 530 is displayed. The reconstructed Y values of the 16 pixels included after downsampling and filtering are y ₀ to y ₁₅ from the bottom. The reconstructed Cb values of the 16 pixels included in the upper adjacent Cb reference line 528 are c _{16 to} c ₃₁ from the left, and the 16 pixels after the downsampling and filtering included in the upper adjacent Y reference line 532 are processed. The reconstructed Y value is defined as y ₁₆ to y ₃₁ from the left.

The intra prediction unit 208 generates a predicted value of the processing target Cb block 504 using a nonlinear model (S410). The intra prediction unit 208 applies the parameter a _i determined in step S408 to Equation 1. The intra prediction unit 208 substitutes the reconstructed value of the pixel of the processing target Y block 502 obtained as a result of the filtering in step S <b> 406 into y in Equation 1, thereby obtaining the predicted value c of the corresponding pixel of the processing target Cb block 504. calculate. In the example of FIGS. 5A and 5B, for example, for the predicted value of the pixel 544 of the processing target Cb block 504, the reconstructed value of the corresponding pixel 546 after down-sampling of the processing target Y block 502 is substituted into Equation 1. It is calculated by doing. The intra prediction unit 208 calculates prediction values in the same manner for all pixels of the processing target Cb block 504.

  FIG. 6 is a block diagram showing the function and configuration of the moving picture decoding apparatus 102 of FIG. The video decoding apparatus 102 includes an entropy decoding unit 602, an inverse quantization / inverse transform unit 604, an addition unit 606, an inter prediction unit 608, an intra prediction unit 610, a frame buffer 612, and an in-loop filter 614. . The moving picture decoding apparatus 102 obtains output moving picture data from the bit stream BS basically by a procedure reverse to the procedure performed by the moving picture encoding apparatus 110.

  The entropy decoding unit 602 receives the bit stream BS from the distribution system 100 via the network 106. The entropy decoding unit 602 performs entropy decoding on the received bit stream, and extracts a level value and side information. Note that the process of obtaining side information and level values from a bitstream is referred to as a parse process. Reconstructing the pixel value using the side information and the level value obtained in this way is called a decoding process.

The inverse quantization / inverse transform unit 604 performs inverse quantization and inverse transform on the level value of the processing target block obtained as a result of entropy decoding, and generates a difference value between the pixels of the processing target block.
The adding unit 606 acquires a prediction value generated by the inter prediction unit 608 or the intra prediction unit 610 depending on whether the processing target block is intra predicted or inter predicted. The adding unit 606 adds the difference value generated by the inverse quantization / inverse transform unit 604 and the corresponding predicted value obtained for each component of Y, Cb, and Cr, and adds each component of the processing target block. The pixel value of the pixel is reconstructed. The adding unit 606 outputs each component of the reconfigured processing target block to the intra prediction unit 610 and the in-loop filter 614.

  A reference picture corresponding to the processing target block is input from the frame buffer 612 to the inter prediction unit 608. The inter prediction unit 608 outputs a motion compensated image of the processing target block by inter prediction based on the input reference picture.

  Image data of a reconstructed block (a block that has already been decoded or a block that has already been reconstructed) around the processing target block is input from the adder 606 to the intra prediction unit 610. The intra prediction unit 610 generates and outputs a predicted value of the pixel of the processing target block by intra prediction based on the input pixel value of the reconfigured block.

  The in-loop filter 614 is a deblock filter, for example. The in-loop filter 614 generates a locally decoded image corresponding to the processing target frame and outputs the locally decoded image to the frame buffer 612. This locally decoded image is used for inter prediction in the inter prediction unit 608 and is simultaneously output to the display device 104 as output moving image data.

  FIG. 7 is a flowchart showing a flow of a series of processes in the intra prediction unit 610 of FIG. The intra prediction unit 610 acquires the intra prediction mode for the luminance component of the processing target block from the side information extracted by the entropy decoding unit 602 (S702). Based on the acquired intra prediction mode, the intra prediction unit 610 identifies a reference line from among the lines included in the luminance component of the reconstructed block around the processing target block (S704). The intra prediction unit 610 performs intra prediction of the luminance component using the reference line specified in step S704 in accordance with the intra prediction mode acquired in step S702 (S706). The intra prediction unit 610 outputs a predicted value of the luminance component generated by the intra prediction. The luminance component intra prediction in step S706 may be realized by HEVC intra prediction.

  The intra prediction unit 610 acquires an intra prediction mode for the color difference component of the processing target block from the side information extracted by the entropy decoding unit 602 (S708). Based on the acquired intra prediction mode, the intra prediction unit 610 identifies a reference line from among the lines included in the color difference components of the reconstructed blocks around the processing target block (S710). The intra prediction unit 610 performs intra prediction of the color difference component using the reference line specified in step S710 in accordance with the intra prediction mode acquired in step S708 (S712). The intra prediction unit 610 outputs a predicted value of the color difference component generated by the intra prediction.

  When the intra prediction mode acquired in step S708 is the CCNM mode, processing similar to the processing shown in FIG. 4 is performed in steps S710 and S712. In particular, step S710 corresponds to step S402, and step S712 corresponds to steps S404, S406, S408, and S410. By replacing “encoded block” in the description of FIG. 4 with “reconstructed block”, description here is omitted.

  According to video encoding device 110 or video decoding device 102 according to the present embodiment, intra prediction of color difference components is performed using a nonlinear model between luminance components and color difference components. Compared with, the accuracy of the predicted value can be increased. In particular, the nonlinear model according to the present embodiment is effective when an intra picture has a complex texture and cannot be handled by a linear model.

  Further, in the present embodiment, the reference line used in intra prediction of the color difference component spans a plurality of encoded / reconstructed blocks. Therefore, since the number of samples is increased compared to the case where only the line of the adjacent encoded / reconstructed block is used as the reference line, the accuracy of the parameter evaluation of the nonlinear model can be increased, or the parameter The robustness of the evaluation can be increased. In addition, since the reference line extends to the upper right and / or lower left, even when there is a directional texture from the lower left to the upper right, highly accurate parameter evaluation is realized.

  In the present embodiment, filtering is performed after downsampling of the luminance component. Accordingly, it is possible to suppress the deviation between the luminance component and the color difference component that may be generated by downsampling. Further, even if the color difference component is generated by downsampling as in the color format 4: 2: 0 adopted in the present embodiment, the reconstructed value of the luminance component is caused by quantization noise or the like. A noise component may be included. According to the present embodiment, such a noise component can be reduced or removed from both the reference line of the luminance component and the reconstructed value of the luminance component by filtering.

Also, in the CCNM mode according to the present embodiment, the nonlinear model parameter a _i is determined from the luminance component reference line and the chrominance component reference line in each of the moving image decoding apparatus 102 and the moving image encoding apparatus 110. . Therefore, since it is not necessary to transmit the parameter a _i by including it in the bit stream BS, it is possible to suppress or eliminate an increase in the code amount.

  The configuration and operation of the embodiment have been described above. This embodiment is an exemplification, and it will be understood by those skilled in the art that various modifications can be made to each component and combination of processes, and such modifications are within the scope of the present invention.

(First modification)
In the embodiment, the case of using a reference line extending over a plurality of blocks has been described, but the present invention is not limited to this. For example, the parameters of the nonlinear model may be determined based on a plurality of reference lines of luminance components and a plurality of reference lines corresponding to chrominance components of encoded / reconstructed blocks around the processing target block. In this case, since the number of samples is increased as compared with the case where one reference line is used, the accuracy of the evaluation of the parameters of the nonlinear model can be increased, or the robustness of the evaluation of the parameters can be increased. .

  FIGS. 8A and 8B are schematic diagrams illustrating reference lines according to the first modification selected for the processing target Y block 502 and the processing target Cb block 504. FIG. The intra prediction unit of the video encoding device according to the first modification includes a line adjacent to the processing target Cb block 504 among lines extending in the vertical direction included in the fourth encoded Cb block 522, and the left of the line. The left adjacent Cb reference line 826 is selected by combining the adjacent lines. The intra prediction unit combines the line adjacent to the processing target Cb block 504 among the lines extending in the left-right direction included in the second encoded Cb block 518 and the upper adjacent Cb that is adjacent to the line above the line. A reference line 828 is selected.

  The intra prediction unit identifies the left adjacent Y reference line 830 and the upper adjacent Y reference line 832 of the encoded Y blocks 508 and 512 corresponding to the selected left adjacent Cb reference line 826 and the upper adjacent Cb reference line 828, respectively. To do. The left adjacent Y reference line 830 is two lines included in the fourth encoded Y block 512 (after downsampling), and the upper adjacent Y reference line 832 is included in the second encoded Y block 508 (down Two lines (after sampling). The intra prediction unit of the video decoding device performs the same processing as described above.

(Second modification)
In the embodiments, the case where the parameters of the nonlinear model are determined using the pixel values included in the reference line has been described, but the present invention is not limited to this. For example, the parameters of the nonlinear model may be determined based on the weight assigned to each of the plurality of reference lines. In this case, since the reliability of each reference line can be considered individually, more accurate parameter evaluation is realized. The weight may be determined according to the block size and the intra prediction mode of the luminance component.

FIGS. 9A and 9B are schematic diagrams for explaining weighting when the reference line extends over a plurality of blocks. FIGS. 10A and 10B are schematic diagrams for explaining weighting when a plurality of reference lines are selected from one block. The intra prediction unit of the video encoding device according to the second modification classifies the pixels constituting the reference line into four zones. A weight w _AB = 0.6 is assigned to zone A (ZONEA) and zone B (ZONEB), and a weight w _CD = 0.4 is assigned to zone C (ZONEC) and zone D (ZONED).

イントラ予測部は、ゾーンＣおよびゾーンＤの画素を用いて以下の第２非線形モデルのパラメータａ_ｉ ^ＣＤを決定する。

イントラ予測部は、
Ｃ＝ｗ_ＡＢ＊Ｃ_ＡＢ＋ｗ_ＣＤ＊Ｃ_ＣＤ
を演算することにより色差成分の予測値を生成する。動画像復号装置のイントラ予測部は上記と同様の処理を行う。本例では重みを０．６：０．４としたが、これは一例であって他の重みが用いられてもよい。 The intra prediction unit determines the following parameters a _i ^AB of the first nonlinear model using the pixels of zone A and zone B.

The intra prediction unit determines a parameter a _i ^CD of the second nonlinear model below using the pixels of zone C and zone D.

The intra prediction unit
C = w _AB * C _AB + w _CD * C _CD
Is used to generate a predicted value of the color difference component. The intra prediction unit of the video decoding device performs the same processing as described above. In this example, the weight is set to 0.6: 0.4, but this is an example, and other weights may be used.

FIG. 11 is a flowchart showing a flow of a series of processes in the intra prediction unit according to the second modification. The intra prediction unit classifies the pixels of the reference line included in the encoded / reconstructed block around the processing target block into a plurality of groups, and assigns a weight to each group (S152). The intra prediction unit determines parameters of a plurality of intra prediction models based on the reconstructed values of the pixels of the reference line (S154). In the above example, the parameter a _i ^AB of the first nonlinear model and the parameter a _i ^CD of the second nonlinear model are determined. When the error e at the time of parameter evaluation is smaller than the predetermined threshold value δ or when the number of repetitions T in step S154 is larger than the predetermined threshold th_iter (YES in S156), the intra prediction unit at that time Holds the parameters. The intra prediction unit performs intra prediction of chrominance components using a plurality of models composed of the held parameters (S158). If e ≧ δ and T ≦ th_iter in step S156, the process returns to step S152.

(Third Modification)
In the second modified example, first, the pixels of the reference line are classified into a plurality of groups, and a weight is assigned to each group. The final model is expressed as a weighted sum of polynomial models derived from each group. In the third modification, a plurality of prediction models, for example, a linear model and a nonlinear model are prepared, and each parameter is determined using all the pixels of the reference line. Specifically, a linear model and a nonlinear model are adopted as the plurality of models in step S154 in FIG. 11, and parameters of each model are determined using all the pixels of the reference line.

イントラ予測部は、参照ラインの画素を用いて以下の非線形モデルのパラメータａ_０、ａ_１、ａ_２を決定する。

イントラ予測部は、
Ｃ＝ｗ_Ｌ＊Ｃ_Ｌ＋ｗ_Ｎ＊Ｃ_Ｎ
を演算することにより色差成分の予測値を生成する。ｗ_Ｌおよびｗ_Ｎはそれぞれ線形モデル、非線形モデルに割り当てられた重みである。動画像復号装置のイントラ予測部は上記と同様の処理を行う。 The intra prediction unit of the video encoding apparatus according to the third modification determines parameters a and b of the following linear model using the pixels of the reference line.

The intra prediction unit determines the following parameters a ₀ , a ₁ , and a ₂ of the nonlinear model using the pixels of the reference line.

The intra prediction unit
C = w _L * C _L + w _N * C _N
Is used to generate a predicted value of the color difference component. w _L and w _N are weights assigned to the linear model and the nonlinear model, respectively. The intra prediction unit of the video decoding device performs the same processing as described above.

(Fourth modification)
In the second modified example and the third modified example, the case where the predicted value of each pixel of the color difference component is calculated using a plurality of prediction models has been described. In the fourth modification, the intra prediction unit first divides the processing target Cb block 504 into a plurality of subgroups. The intra prediction unit sets a prediction model individually for each subgroup. That is, the type of prediction model (linear model / nonlinear model) and the prediction parameter may differ for each subgroup. The intra prediction unit calculates the prediction value of the color difference component for each subgroup using the prediction model set for the subgroup.

(5th modification)
In the embodiment, the case where the intra prediction values of the Cb component and the Cr component are calculated from the reconstructed value of the Y component in the CCNM mode has been described. However, the present invention is not limited to this. 12A and 12B are explanatory diagrams of a prediction model according to the fifth modification. In FIG. 12A, the predicted value of the Cb component is generated using the first prediction model 180 between the Y component and the Cb component. A predicted value of the Cr component is generated using the second prediction model 182 between the Y component and the Cr component and the third prediction model 184 between the Cb component and the Cr component. At least one of the first prediction model 180, the second prediction model 182 and the third prediction model 184 is a non-linear model. For example, the predicted value of the Cr component is
Cr = α * Y + (1-α) * Cb
Is calculated by Here, α is a weighting coefficient of 0 or more and 1 or less.

In FIG. 12B, the predicted value of the Cr component is generated using the fourth prediction model 186 between the Y component and the Cr component. A predicted value of the Cb component is generated using a fifth prediction model 188 between the Y component and the Cb component and a sixth prediction model 190 between the Cr component and the Cb component. At least one of the fourth prediction model 186, the fifth prediction model 188, and the sixth prediction model 190 is a non-linear model. For example, the predicted value of the Cb component is
Cb = β * Y + (1-β) * Cr
Is calculated by Here, β is a weighting coefficient of 0 or more and 1 or less.

  In the embodiment, the case where the polynomial model is adopted as the nonlinear model has been described.

  In the embodiment, the case where the reconstructed value of the luminance component corresponding to the nonlinear model is input when the predicted value of the color difference component is calculated from the nonlinear model is not limited to this. A predicted value may be input.

  In the embodiment, the case where both the reconstruction value of the Y block and the reference line of the Y block are filtered in step S406 in FIG. 4 is not limited to this. For example, the reference line of the Y block is filtered in step S406. The reconstruction value of the Y block may be filtered before step S408. Alternatively, filtering may not be performed.

  DESCRIPTION OF SYMBOLS 100 distribution system, 102 moving image decoding apparatus, 104 display apparatus, 106 network, 110 moving image encoding apparatus.

Claims (13)

Decoding means for decoding a bitstream representing a moving image including a block encoded using intra prediction;
Inverse processing means for generating a difference value of the block by dequantizing and inversely transforming the level value of the block obtained as a result of decoding;
Intra prediction means for generating a prediction value of the block by intra prediction based on a reconstructed block around the block;
Reconstructing means for reconstructing the block based on the generated difference value and the generated predicted value;
The intra-prediction unit is a moving picture decoding apparatus that generates a prediction value of a color difference component of a block using a nonlinear model between a luminance component and a color difference component of the block.   The moving image decoding apparatus according to claim 1, wherein the non-linear model is a model that represents a color difference component by a polynomial of a corresponding luminance component.   3. The moving image according to claim 1, wherein the intra prediction unit determines the parameters of the nonlinear model based on a reference line of a luminance component and a corresponding reference line of a color difference component of a reconstructed block around the block. Decoding device. The reconstructed block around the block is adjacent to the first reconstructed block in the direction in which the first reference line of the luminance component of the first reconstructed block and the first reconstructed block adjacent to the block extend. A second reconstructed block to
The video decoding device according to claim 3, wherein the intra prediction unit determines the parameter of the nonlinear model based on the first reference line and a second reference line of a luminance component of the second reconstructed block.   The said intra prediction means determines the parameter of the said nonlinear model based on the some reference line of the luminance component of the reconfigure | reconstructed block around the said block, and the some reference line corresponding to a colour-difference component. Video decoding device.   The moving picture decoding apparatus according to claim 4 or 5, wherein the intra prediction means determines the parameter of the nonlinear model based on a weight assigned to each reference line. The size of the color difference component of the block is smaller than the size of the luminance component of the block,
The intra prediction means includes
Means for downsampling the luminance component of the block and a reference line of luminance components of the reconstructed block around the block;
Means for filtering the luminance component and the reference line obtained as a result of downsampling;
Means for determining parameters of the nonlinear model based on a reference line obtained as a result of filtering and a corresponding reference line of the color difference component;
7. The moving picture decoding apparatus according to claim 3, further comprising: a unit that generates a predicted value of the corresponding color difference component using the determined parameter and the luminance component obtained as a result of filtering.   The moving picture decoding apparatus according to claim 7, wherein the filtering unit is a directional filter corresponding to the direction of the directional mode used for intra prediction of the luminance component of the block. The color difference component of the block includes a first color difference component and a second color difference component,
The intra prediction unit generates a predicted value of the first color difference component of the block using a first prediction model between the luminance component of the block and the first color difference component, and Generating a predicted value of the second color difference component of the block using a second prediction model between the two color difference components and a third prediction model between the first color difference component and the second color difference component of the block;
9. The moving picture decoding apparatus according to claim 1, wherein at least one of the first prediction model, the second prediction model, and the third prediction model is a non-linear model. A decoding step of decoding a bitstream representing a moving image including a block encoded using intra prediction;
An inverse processing step of generating a difference value of the block by inverse-quantizing and inverse-transforming the level value of the block obtained as a result of decoding;
An intra prediction step of generating a prediction value of the block by intra prediction based on a reconstructed block around the block;
Reconstructing the block based on the generated difference value and the generated predicted value, and
The intra prediction step includes a step of generating a prediction value of the color difference component of the block using a nonlinear model between the luminance component and color difference component of the block. Intra prediction means for generating a prediction value of the block by intra prediction based on an encoded block around a block of a moving image;
Difference generation means for generating a difference value of the block based on the generated predicted value;
Processing means for converting and quantizing the generated difference value to generate a level value;
Encoding means for encoding the generated level value to generate a bitstream,
The intra-prediction unit is a moving picture coding apparatus that generates a prediction value of a color difference component of a block using a nonlinear model between a luminance component and a color difference component of the block. An intra prediction step of generating a prediction value of the block by intra prediction based on an encoded block around a block of a moving image;
A difference generation step of generating a difference value of the block based on the generated predicted value;
Processing and transforming and quantizing the generated difference value to generate a level value;
Encoding a generated level value to generate a bitstream; and
The intra prediction step includes a step of generating a predicted value of the color difference component of the block using a nonlinear model between the luminance component and color difference component of the block.   A computer-readable recording medium comprising a program that causes a computer to function as the moving picture decoding apparatus according to any one of claims 1 to 9 or the moving picture encoding apparatus according to claim 11.

Images