WO2018212582A1 - Intra prediction encoding or decoding method and device - Google Patents

Intra prediction encoding or decoding method and device Download PDF

Info

Publication number
WO2018212582A1
WO2018212582A1 PCT/KR2018/005596 KR2018005596W WO2018212582A1 WO 2018212582 A1 WO2018212582 A1 WO 2018212582A1 KR 2018005596 W KR2018005596 W KR 2018005596W WO 2018212582 A1 WO2018212582 A1 WO 2018212582A1
Authority
WO
WIPO (PCT)
Prior art keywords
value
pixel
current block
neighboring pixel
neighboring
Prior art date
Application number
PCT/KR2018/005596
Other languages
French (fr)
Korean (ko)
Inventor
임정연
김효성
김형덕
손세훈
신재섭
이경택
이선영
Original Assignee
에스케이텔레콤 주식회사
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Priority claimed from KR1020170088703A external-priority patent/KR102356317B1/en
Application filed by 에스케이텔레콤 주식회사 filed Critical 에스케이텔레콤 주식회사
Publication of WO2018212582A1 publication Critical patent/WO2018212582A1/en

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/102—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the element, parameter or selection affected or controlled by the adaptive coding
    • H04N19/103—Selection of coding mode or of prediction mode
    • H04N19/11—Selection of coding mode or of prediction mode among a plurality of spatial predictive coding modes
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/10—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding
    • H04N19/169—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding
    • H04N19/182—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using adaptive coding characterised by the coding unit, i.e. the structural portion or semantic portion of the video signal being the object or the subject of the adaptive coding the unit being a pixel
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/50—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding
    • H04N19/593—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using predictive coding involving spatial prediction techniques
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
    • H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
    • H04N19/70—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards

Definitions

  • the present invention relates to image encoding or decoding for improving the accuracy of prediction in intra prediction.
  • Non-directional prediction modes include a planar prediction mode and a DC prediction mode.
  • the directional prediction mode is effective for a region having a constant directionality in an image, but the non-directional prediction mode is more effective for a flat region having continuity with respect to surrounding reference signals.
  • the DC prediction mode is provided as an alternative to the directional prediction mode as a prediction method based on the average value of neighboring pixels in the current block, but the prediction accuracy is poor in that it is only an approximation. Accordingly, the plane prediction mode is newly introduced in HEVC.
  • Planar prediction mode is a prediction method for generating a two-dimensional plane reflecting the distance according to the position of the adjacent pixels can generate a more precise prediction value.
  • 1 is an exemplary diagram for describing a planar prediction mode according to the prior art.
  • a prediction value of the current block is generated by using adjacent pixels located in the upper row and the left column of the current block as reference pixels.
  • neighboring pixels used as reference pixels are pixels reconstructed by prediction and reconstruction. Since coding units (CUs) in a coding tree unit (CTU) are processed in a Z-scan order, adjacent pixels in the right column and the lower row of the current block cannot be used as reference pixels in the intra prediction process.
  • CUs coding units
  • CTU coding tree unit
  • the values of all adjacent pixels located in the right column of the current block 100 are set to be the same as the values of the adjacent pixels TR located in the upper right corner.
  • the value of all adjacent pixels located in the lower row of (100) is set equal to the values of the adjacent pixels BL and 116 located in the lower left corner. Accordingly, when the values of all adjacent pixels surrounding the current block 100 are determined, bilinear interpolation is performed in the horizontal direction and the vertical direction with respect to the current pixel 110.
  • a first prediction value of the current pixel 100 is generated by linear interpolation using adjacent pixels 114 and 120 positioned in the same row as the current pixel 110, and FIG. 1.
  • a second prediction value of the current pixel 100 is generated by linear interpolation using adjacent pixels 118 and 122 positioned in the same column as the current pixel 110.
  • the final prediction value of the current pixel 100 is determined as an average value of the first prediction value and the second prediction value.
  • the conventional planar prediction mode sets the values of all adjacent pixels located in the right column of the current block 100 and the values of all adjacent pixels located in the lower row, respectively, to the same value, the lower right corner of the current block 100. As the direction increases, the accuracy of the pixel prediction value tends to decrease.
  • the present invention provides an image encoding or decoding technique for increasing the precision of intra prediction.
  • the method comprising: setting a value of the first neighboring pixel by decoding information indicating a first neighboring pixel located at a lower right corner of the current block from a bitstream; The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; Obtaining a value of a lower adjacent pixel and a value of a right adjacent pixel of the pixel; And obtaining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel. It provides a video decoding method comprising the step.
  • the current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block;
  • the decoding unit for decoding the information indicating the first adjacent pixel located in the lower right corner of the current block from the bitstream;
  • a first neighboring pixel setting unit configured to set a value of the first neighboring pixel by using information indicating the first neighboring pixel;
  • the current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block;
  • a neighboring pixel predictor configured to determine a value of a lower neighboring pixel and a value of a right neighboring pixel of the pixel; And determining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel.
  • the first adjacent pixel setting unit for setting the value of the first adjacent pixel located in the lower right corner of the current block based on the reference pixels of the current block;
  • the current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block;
  • a neighboring pixel predictor configured to determine a value of a lower neighboring pixel and a value of a right neighboring pixel of the pixel; And determining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel.
  • 1 is an exemplary diagram for describing a planar prediction mode according to the prior art.
  • FIG. 2 is a block diagram of an image encoding apparatus according to an embodiment of the present invention.
  • FIG. 3 is a block diagram of an image decoding apparatus according to an embodiment of the present invention.
  • FIG. 4 is a block diagram of an apparatus for performing intra prediction in a planar prediction mode according to an embodiment of the present invention.
  • FIG. 5 is an exemplary diagram for describing a planar prediction mode according to an embodiment of the present invention.
  • 6 through 8 are exemplary diagrams of a planar prediction mode using information decoded from a bitstream according to an embodiment of the present invention.
  • 9 to 11 are exemplary views of planar prediction mode using a value calculated based on neighboring pixels of the current block according to another embodiment of the present invention.
  • FIGS. 12 to 15 are exemplary diagrams of a planar prediction mode in a 360 image according to another embodiment of the present invention.
  • 16 is an exemplary diagram illustrating a method of setting values of neighboring pixels of a current block according to another embodiment of the present invention.
  • 17 is a flowchart illustrating an intra prediction prediction method according to an embodiment of the present invention.
  • FIG. 18 is a flowchart illustrating an intra prediction prediction method according to another embodiment of the present invention.
  • the current block may refer to a coding unit (CU) or a prediction unit (PU) to be encoded or decoded.
  • the current pixel may refer to a pixel in a current block for which a prediction value is to be generated according to an intra prediction mode.
  • the adjacent pixel may refer to a pixel immediately adjacent to the current block.
  • the adjacent pixel may be a reconstructed pixel value or may be a value predicted using another pixel value.
  • the peripheral pixel may refer to the reconstructed pixel that is not adjacent to the current block but located in the peripheral.
  • the reference pixel may refer to a reconstructed pixel that is referred to to generate a prediction value of the current pixel, and may include not only pixels (adjacent pixels) adjacent to the current block but also non-adjacent pixels (main pixel). Meanwhile, the term 'pixel' is interchangeable with terms such as pixel and sample.
  • FIG. 2 is a block diagram of an image encoding apparatus 200 according to an embodiment of the present invention.
  • the image encoding apparatus 200 may include a block divider 210, a predictor 220, a subtractor 230, a transformer 240, a quantizer 245, an encoder 250, an inverse quantizer 260, An inverse transform unit 265, an adder 270, a filter unit 280, and a memory 290 are included.
  • Each component of the encoding apparatus 200 may be implemented by a hardware chip, or may be implemented by software and a microprocessor to execute a function of software corresponding to each component.
  • the block dividing unit 210 After dividing each picture constituting the image into a plurality of coding tree units (CTUs), the block dividing unit 210 recursively divides the CTUs using a tree structure.
  • a leaf node in the tree structure becomes a coding unit (CU) which is a basic unit of coding.
  • CU coding unit
  • QT QuadTree
  • QTBT QuadTree
  • BT binaryTree
  • BinaryTree BinaryTree
  • the prediction unit 220 generates a prediction block by predicting the current block.
  • the predictor 220 includes an intra predictor 222 and an inter predictor 224.
  • the current block is a basic unit of encoding corresponding to a leaf node in the tree structure, and means a CU to be currently encoded.
  • the current block may be one subblock of the plurality of subblocks divided from the CU.
  • the intra predictor 222 predicts pixels in the current block by using pixels (reference pixels) positioned in the vicinity of the current block in the current picture including the current block.
  • the intra predictor 222 may determine an intra prediction mode to be used to encode the current block.
  • intra prediction unit 222 may encode the current block using several intra prediction modes and select an appropriate intra prediction mode to use from the tested modes. For example, intra predictor 222 calculates rate distortion values using rate-distortion analysis for several tested intra prediction modes, and has the best rate distortion characteristics among the tested modes. Intra prediction mode may be selected.
  • the plurality of intra prediction modes according to the embodiment of the present invention may include two non-directional modes (plane mode and DC mode) and 65 directional modes. A detailed description of the planar mode according to embodiments of the present invention will be described later with reference to other drawings.
  • the intra prediction unit 222 selects one intra prediction mode from among the plurality of intra prediction modes, and predicts the current block by using a reference pixel and an operation formula determined according to the selected intra prediction mode. Information on the selected intra prediction mode is encoded by the encoder 250 and transmitted to the image decoding apparatus.
  • the inter prediction unit 224 searches for the block most similar to the current block in the coded and decoded reference picture before the current picture, and generates a prediction block for the current block using the searched block.
  • a motion vector (MV) corresponding to a displacement between the current block in the current picture and the prediction block in the reference picture is generated.
  • Motion information including information about a reference picture and information about a motion vector used to predict the current block is encoded by the encoder 250 and transmitted to the image decoding apparatus.
  • the subtractor 230 subtracts the prediction block generated by the intra predictor 222 or the inter predictor 224 from the current block to generate a residual block.
  • the converter 240 converts the residual signal in the residual block having pixel values of the spatial domain into a transform coefficient of the frequency domain.
  • the transform unit 240 may convert the residual signals in the residual block using the size of the current block as a conversion unit, or divide the residual block into a plurality of smaller subblocks and convert the residual signals in a subblock-sized transform unit. You can also convert. There may be various ways of dividing the residual block into smaller subblocks. For example, the subblock may be divided into sub-blocks having the same size, or may be divided by a quadtree (QT) method using the residual block as a root node.
  • QT quadtree
  • the quantization unit 245 quantizes the transform coefficients output from the transform unit 240, and outputs the quantized transform coefficients to the encoder 250.
  • the encoder 250 generates a bitstream by encoding the quantized transform coefficients by using an encoding method such as CABAC.
  • the encoder 250 encodes information about the size of the CTU located in the highest layer of the tree structure and split information for dividing the block into the tree structure from the CTU, so that the decoding apparatus divides the block in the same way as the encoding apparatus. Do it.
  • QT splitting QT splitting information indicating whether a block of an upper layer is divided into four blocks of a lower layer is encoded.
  • BT partitioning BT partitioning information indicating whether each block is divided into two blocks and the type of partitioning is encoded, starting from the block corresponding to the leaf node of the QT.
  • the encoder 250 encodes information about a prediction type indicating whether the current block is encoded by intra prediction or inter prediction, and encodes intra prediction information or inter prediction information according to the prediction type.
  • the inverse quantizer 260 inversely quantizes the quantized transform coefficients output from the quantizer 245 to generate transform coefficients.
  • the inverse transformer 265 restores the residual block by converting the transform coefficients output from the inverse quantizer 260 from the frequency domain to the spatial domain.
  • the adder 270 reconstructs the current block by adding the reconstructed residual block and the prediction block generated by the prediction unit 220.
  • the pixels in the reconstructed current block are used as reference pixels when intra prediction of blocks of the next order.
  • the filter unit 280 deblocks and filters the boundary between the reconstructed blocks in order to remove blocking artifacts that occur due to encoding / decoding of blocks. When all the blocks in a picture are reconstructed, the reconstructed picture is used as a reference picture for inter prediction of a block in a picture to be encoded later.
  • FIG. 3 is a block diagram of an image decoding apparatus 300 according to an embodiment of the present invention.
  • the image decoding apparatus 300 includes a decoder 310, an inverse quantizer 320, an inverse transformer 330, a predictor 340, an adder 350, a filter 360, and a memory 370.
  • the components shown in FIG. 3 may be implemented in a hardware chip or may be implemented in software and implemented so that the microprocessor executes a function of software corresponding to each component.
  • the decoder 310 decodes the bitstream received from the image encoding apparatus, extracts information related to block division, determines a current block to be decoded, and includes prediction information and residual signal information necessary for reconstructing the current block. Extract
  • the decoder 310 extracts information on the CTU size from a Sequence Parameter Set (SPS) or Picture Parameter Set (PPS) to determine the size of the CTU, and divides the picture into a CTU of the determined size.
  • the CTU is determined as the highest layer of the tree structure, that is, the root node, and the CTU is partitioned using the tree structure by extracting partition information about the CTU. For example, when splitting a CTU using a QTBT structure, first, a first flag (QT_split_flag) related to splitting of QT is extracted, and each node is divided into four nodes of a lower layer. For the node corresponding to the leaf node of the QT, the second flag BT_split_flag and the split type information related to the splitting of the BT are extracted to split the corresponding leaf node into the BT structure.
  • SPS Sequence Parameter Set
  • PPS Picture Parameter Set
  • the decoder 310 determines the current block (current block) to be decoded by splitting the tree structure, the decoder 310 extracts information about a prediction type indicating whether the current block is intra predicted or inter predicted.
  • the decoder 310 extracts a syntax element for intra prediction information (eg, intra prediction mode, information on a reference pixel, etc.) of the current block.
  • a syntax element for intra prediction information eg, intra prediction mode, information on a reference pixel, etc.
  • the decoder 310 extracts a syntax element for inter prediction information (eg, inter prediction mode).
  • the decoder 310 extracts information on the quantized transform coefficients of the current block as information on the residual signal.
  • the inverse quantizer 320 inversely quantizes the quantized transform coefficients, and the inverse transformer 330 inversely transforms the inverse quantized transform coefficients from the frequency domain to the spatial domain to generate a residual block for the current block. .
  • the predictor 340 includes an intra predictor 342 and an inter predictor 344.
  • the intra predictor 342 is activated when the intra prediction is the prediction type of the current block
  • the inter predictor 344 is activated when the intra prediction is the prediction type of the current block.
  • the intra prediction unit 342 determines the intra prediction mode of the current block among the plurality of intra prediction modes from the syntax elements for the intra prediction mode extracted from the decoder 310, and references the reference pixels around the current block according to the intra prediction mode. Predict the current block using In addition, the intra predictor 342 may set a value of a reference pixel to be used for intra prediction from a syntax element for the reference pixel extracted from the decoder 310.
  • the inter prediction unit 344 determines motion information of the current block using a syntax element of the inter prediction mode extracted from the decoder 310, and predicts the current block using the determined motion information.
  • the adder 350 reconstructs the current block by adding the residual block output from the inverse transformer 330 and the prediction block output from the inter predictor 344 or the intra predictor 342.
  • the pixels in the reconstructed current block are utilized as reference pixels in intra prediction of a block to be subsequently decoded.
  • the filter unit 360 deblocks and filters the boundary between the reconstructed blocks in order to remove blocking artifacts caused by block-by-block decoding, and stores them in the memory 370.
  • the reconstructed picture is used as a reference picture for inter prediction of a block in a picture to be decoded later.
  • the apparatus 400 may include a first neighboring pixel setting unit 410, a neighboring pixel prediction unit 420, and a current pixel prediction unit 430, and may be implemented in the inter prediction unit 344 of FIG. 3.
  • the components shown in FIG. 4 may be implemented as a hardware chip, or may be implemented in software and such that the microprocessor executes the functions of the software corresponding to each component.
  • FIG. 5 is an exemplary diagram for describing a planar prediction mode according to an embodiment of the present invention.
  • the planar prediction mode proposed in the present invention predicts the current pixel 510 by performing double interpolation in the vertical and horizontal directions of the current pixel 510 as shown in FIGS. 5A and 5B.
  • the present invention first sets the value of the adjacent pixel (BR, hereinafter referred to as 'first adjacent pixel') located at the lower right corner of the current block 500, and based on this value, The lower neighboring pixels and the right neighboring pixels of 500 are predicted.
  • the first adjacent pixel value is the value of the adjacent pixel (TR, hereinafter referred to as 'second adjacent pixel') located at the upper right corner, and the adjacent pixel (BL, hereinafter referred to as 'third adjacent pixel') located at the lower left corner. It can be set using a value of).
  • the value of the right adjacent pixel 512 of the current block 500 is set equal to the value of the second adjacent pixel located at the upper right corner. It is set to a value obtained by interpolation using the value of the second neighboring pixel TR and the value of the first neighboring pixel BR.
  • the lower side of the current block 500 is conventionally referred to.
  • the value of the adjacent pixel 516 is set to be the same as the value of the third adjacent pixel located at the lower left corner, but in this embodiment, the value of the third adjacent pixel BL and the value of the first adjacent pixel BR are used. It is set to the value obtained by interpolation.
  • a process of obtaining the values of the right adjacent pixel 512 and the lower neighboring pixel 516 of the current block 500 may be represented by Equation 1.
  • H means the height of the current block
  • W means the width of the current block
  • P (W, y) means the value of the adjacent pixel 512 adjacent to the right boundary of the current block
  • P (x, H) means the value of the adjacent pixel 516 adjacent to the lower boundary of the current block
  • P (W, H) means the value of the first adjacent pixel BR.
  • Pixels denoted as R are reference pixels of the current block
  • R (W, -1) denotes the value of the second adjacent pixel TR
  • R (-1, H) denotes the value of the third adjacent pixel BL.
  • the first prediction value of the current pixel 510 is generated by a value interpolated in the horizontal direction using the value of the left adjacent pixel 514 and the right adjacent pixel 512 of the current block 500.
  • the second prediction value of the current pixel 510 is generated by a value interpolated in the vertical direction by using the value of the lower neighboring pixel 516 and the upper neighboring pixel 518 of the current block 500.
  • the first prediction value and the second prediction value of the current pixel 510 may be expressed by Equation 2 when obtained by using linear interpolation.
  • P 1 (x, y) means the first predicted value of the position (x, y) interpolated in the horizontal direction
  • P 2 (x, y) means the first predicted value of the position (x, y) interpolated in the vertical direction.
  • 2 means a predicted value.
  • H means the height of the current block
  • W means the width of the current block.
  • R (x, -1) means the value of the upper neighboring pixel
  • R (-1, y) means the value of the left neighboring pixel.
  • the final prediction value of the current pixel 510 according to the present embodiment is determined using the first prediction value and the second prediction value. For example, in the case of using the average of the first prediction value and the second prediction value, this is expressed by Equation 3 below.
  • P 1 (x, y) means the first predicted value of the position (x, y) interpolated in the horizontal direction
  • P 2 (x, y) means the first predicted value of the position (x, y) interpolated in the vertical direction.
  • 2 means a predicted value
  • P f (x, y) means the final predicted value of the (x, y) coordinates according to the present embodiment. Equation 3 is an example, and a method of calculating the final prediction value using the first prediction value and the second prediction value is not limited thereto.
  • a weighted average of the first prediction value and the second prediction value may be used.
  • the reference pixel R that is, the left and upper neighboring pixels of the current block
  • the neighboring pixel P that is, the right and lower neighboring pixels of the current block
  • a method of varying weight based on the position of the first predicted value predicted horizontally and the second predicted value predicted vertically is also possible. This is expressed as an equation (4).
  • P 1 (x, y) means the first predicted value of the position (x, y) interpolated in the horizontal direction
  • P 2 (x, y) means the first predicted value of the position (x, y) interpolated in the vertical direction.
  • 2 means a predicted value.
  • P f (x, y) means the final predicted value of the (x, y) coordinates according to the present embodiment.
  • the vertical distance (y + 1) and the horizontal distance (x + 1) are used as weights. As the value of y in the vertical direction increases and the value of x in the horizontal direction, the final prediction value in each direction moves away from the reference pixel R and approaches the adjacent pixel P, so that the two weights are equal to the first prediction value and the second prediction value.
  • the first weight applied to the first prediction value is determined by the distance from the upper reference pixel located in the same column as the current pixel to the current pixel, and the second weight applied to the second prediction value is located on the left in the same row as the current pixel. It is determined by the distance from the reference pixel to the current pixel.
  • the ranges of x and y are (0 to W-1) and (0 to H-1), respectively, so that the same calculation as in Equation 3 is performed when (x, y) is (0, 0).
  • the final predicted value is obtained by giving a greater weight to the predicted value closer to the reference pixel R among the first predicted value and the second predicted value.
  • the first neighboring pixel setting unit 410 sets the value of the first neighboring pixel BR by using the information indicating the first neighboring pixel BR of the current block decoded from the bitstream by the decoder 310. Can be. Alternatively, the first neighboring pixel setting unit 410 may set the value of the first neighboring pixel BR based on the reference pixels of the current block.
  • the neighboring pixel prediction unit 420 may determine the value of the lower neighboring pixel of the current block by interpolation using the value of the first neighboring pixel BR, the value of the second neighboring pixel TR, and the value of the third neighboring pixel BL.
  • the neighboring pixel predictor 420 determines the value of the right neighboring pixel by interpolation using the value of the first neighboring pixel BR and the value of the second neighboring pixel TR, and determines the first neighboring pixel BR.
  • the value of the lower adjacent pixel is determined by interpolation using the value of) and the value of the third adjacent pixel BL.
  • the current pixel prediction unit 430 determines the prediction value of the current pixel in the current block by interpolation using the value of the left neighboring pixel, the upper neighboring pixel, the lower neighboring pixel, and the right neighboring pixel of the current block. Specifically, the current pixel prediction unit 430 determines the first prediction value of the current pixel by interpolation using the value of the left neighboring pixel and the value of the right neighboring pixel, and uses the value of the upper neighboring pixel and the value of the lower neighboring pixel. The second prediction value of the current pixel is determined by interpolation. The current pixel prediction unit 430 determines a final prediction value by using the first prediction value and the second prediction value of the current pixel.
  • the left neighboring pixels and the upper neighboring pixels are pixels that have already been reconstructed by prediction and reconstruction as reference pixels of the current block.
  • An important issue in the method of planar prediction mode of this embodiment is how to set the value of the first adjacent pixel BR located at the lower right corner of the current block.
  • the value of the first neighboring pixel BR is calculated by the same operation as the encoding apparatus using other neighboring pixels and / or neighboring pixels in the decoding apparatus, or for the value of the first neighboring pixel BR determined by the encoding apparatus.
  • Information may be transmitted to the decoding device.
  • the present embodiment may not obtain an improved effect when the residual data reduced according to the planar prediction of the present embodiment is smaller than the data of the first adjacent pixel BR transmitted. Therefore, it is very important to efficiently transmit the value of the first adjacent pixel BR.
  • FIGS. 6 to 8 are exemplary diagrams of a planar prediction mode using information indicating a first neighboring pixel BR decoded from a bitstream according to an embodiment of the present invention.
  • the information indicating the first adjacent pixel BR includes information indicating a position of a pixel having a value most similar to the original value of the first adjacent pixel BR and the original value of the first adjacent pixel BR.
  • FIG. 6 illustrates an embodiment in which the image encoding apparatus directly transmits information about a value of the first adjacent pixel BR to the image decoding apparatus.
  • the image decoding apparatus decodes a syntax element (eg, br_value ) indicating a value of the first neighboring pixel BR from the bitstream and decodes a value indicated by the corresponding syntax element in the first neighboring pixel BR. It can be set to the value of).
  • a syntax element eg, br_value
  • FIG. 7 illustrates an embodiment in which the image encoding apparatus transmits information about a position of a pixel having a value most similar to the original value of the first adjacent pixel BR to the image decoding apparatus.
  • the information about the position of the reference pixel having the value most similar to the original value of the first adjacent pixel BR may be represented by a one-dimensional index.
  • the image decoding apparatus decodes a syntax element (for example, br_idx ) indicating information about a position of a reference pixel from the bitstream to determine a value of a reference pixel at a position indicated by the corresponding syntax element. It may be set to the value of the pixel BR.
  • a syntax element for example, br_idx
  • the br_idx syntax element indicates the position of a pixel having a value most similar to the original value of the first neighboring pixel BR as a one-dimensional offset by setting the position of one pixel among the reference pixels of the current block as (0, 0). can do.
  • the one-dimensional offset is set by setting any one of the pixel TL, the second adjacent pixel TR, and the third adjacent pixel BL positioned at the upper left corner of the current block as (0, 0). Can be determined. 7 illustrates a case where the position of the pixel TL is set to (0, 0).
  • the apparatus for encoding an image may transmit a syntax element (eg, br_direction ) indicating a reference direction so that the apparatus for decoding an image may determine the reference direction horizontally or vertically by decoding the syntax element.
  • FIG. 8 illustrates another embodiment in which the image encoding apparatus transmits information about a position of a pixel having a value most similar to the original value of the first adjacent pixel BR to the image decoding apparatus.
  • the information about the position of the neighboring pixel and / or the adjacent pixel having the value most similar to the original value of the first neighboring pixel BR may be expressed as a two-dimensional vector.
  • the image decoding device further comprises: a flat prediction mode is selected if, syntax elements indicating the information on the location of the pixels having similar values: decoding (for example, u_br, v_br) from the bit stream of the peripheral pixels of the location indicated by the corresponding syntax element The value may be set to the value of the first adjacent pixel BR.
  • the u_br and v_br syntax elements are two-dimensional offsets of the position of the pixel having the value most similar to the original value of the first adjacent pixel BR, with the position of one pixel among the reference pixels of the current block as (0, 0). Can be indicated.
  • a two-dimensional offset is set by setting any one of the pixel TL, the second adjacent pixel TR, and the third adjacent pixel BL positioned at the upper left corner of the current block as (0, 0).
  • Can be determined. 8 illustrates a case where the position of the pixel TL is set to (0, 0).
  • FIGS. 9 through 11 are exemplary views of planar prediction mode using a value calculated based on neighboring pixels and / or neighboring pixels of a current block according to another embodiment of the present invention.
  • FIG. 9 illustrates a first neighboring pixel based on an amount of change of values of neighboring pixels and neighboring pixels reconstructed without separately receiving information on a value of a first neighboring pixel from the image encoding apparatus.
  • An embodiment of a method of interpolating a value of (BR) is shown.
  • the apparatus for decoding an image may be arranged on the same straight line as a reference pixel (eg, TR, TL, BL) or a value of a neighboring pixel of the current block located on a straight line in at least one specific direction extending from the first adjacent pixel BR.
  • the first neighboring pixel (BR) prediction value in the corresponding direction is calculated by summing change amounts with the values of neighboring pixels which are already restored.
  • a value of the first adjacent pixel BR is set as an average of the first neighboring pixel BR predicted values from each direction.
  • the specific direction may be at least one of a horizontal direction, a vertical direction, and a diagonal direction from the first adjacent pixel BR
  • the reference pixel may include the adjacent pixel TL and the second adjacent pixel located at the upper left corner of the current block. TR) and at least one of the third adjacent pixels BL.
  • FIG. 9 illustrates a case of interpolating values of the first neighboring pixel BR based on the amount of change in each direction of three reference pixels TR, TL, and BL located at the edge of the current block.
  • the amount of change between the value of the second adjacent pixel TR located on a straight line in the vertical direction of the first adjacent pixel BR and the value of the neighboring pixel 910 which has already been restored is added to the value of the second adjacent pixel TR to add the first change. 1
  • the vertical prediction value of the adjacent pixel BR is calculated.
  • the amount of change between the value of the adjacent pixel TL positioned on the diagonal line of the first adjacent pixel BR and the value of the neighboring pixel 920 already restored is added to the value of the peripheral pixel TL to add the first adjacent pixel ( Calculate the diagonal prediction value of BR).
  • the amount of change between the value of the third adjacent pixel BL positioned on the horizontal straight line of the first adjacent pixel BR and the value of the neighboring pixel 930 that is already restored is added to the value of the third adjacent pixel BL.
  • the horizontal prediction value of the first adjacent pixel BR is calculated.
  • the value of the first adjacent pixel BR is generated as an average of the predicted values of the second adjacent pixel BR calculated from the adjacent pixels TR, TL, and BL.
  • Equation 5 according to the interpolation method of the first adjacent pixel BR proposed in the present embodiment is shown in Equation 5 below.
  • H means the height of the current block
  • W means the width of the current block.
  • P (W, H) means the value of the first adjacent pixel BR
  • R (-1, -1) means the value of the adjacent pixel TL located at the upper left corner of the current block
  • R ( W and -1 mean a value of the second adjacent pixel TR
  • R (-1, H) means a value of the third adjacent pixel BL.
  • R (-1-W, -1-H), R (W, -1-H), and R (-1-W, H) are the values of the neighboring pixels that have already been restored, separated by the size of the current block in each direction. it means.
  • FIG. 10 illustrates an example of calculating an interpolation value of the first adjacent pixel BR according to the planar prediction mode of FIG. 9.
  • the value of the second adjacent pixel TR is 50
  • the value of the reconstructed peripheral pixel 1010 at a position vertically away from the second adjacent pixel TR is 60
  • the value of the upper left corner pixel TL is 71
  • the left side is 85
  • the value of the reconstructed peripheral pixel 1020 at a position diagonally away from the upper edge pixel TL is 85
  • the value of the third adjacent pixel BL is 50
  • the value of the position is horizontally separated from the third adjacent pixel BL. It is assumed that the value of the restored peripheral pixel 1030 is 60.
  • the interpolation value of the first adjacent pixel BR reflecting the amount of change in the pixel value in each direction is calculated as 46.
  • the amount of change is calculated based on the distance between blocks, but the method of calculating the amount of change is not limited thereto, and various methods such as the distance between pixels may be used.
  • weights may be added to each pixel value used in the interpolation process.
  • Equation 6 the equation in the case where the weight is added to the values of the reference pixels TL, TR, and BL in which each change amount is corrected is expressed by Equation 6.
  • Each weight is determined in the same manner by the image encoding apparatus and the decoding apparatus according to the characteristics of the image, and is not limited to a specific method.
  • FIG. 11 shows a value of a first neighboring pixel BR based on an average value or a weighted average value of reference pixels of a current block without separately receiving information about the value of the first neighboring pixel BR from the image encoding apparatus.
  • An embodiment of a setting method is shown. Examples of the first adjacent pixel value BR calculated according to the present embodiment are as follows.
  • TR is the adjacent pixel located at the upper right corner of the current block (ie, the second adjacent pixel)
  • BL is the adjacent pixel located at the lower left corner (ie, the third adjacent pixel)
  • TL is left It means the adjacent pixel located in the upper corner.
  • TR ⁇ x means pixels located in the left and right directions with respect to TR
  • BL ⁇ y means pixels located in the up and down direction with respect to BL.
  • TR + 1 means a pixel in contact with the right side of TR and TR + 2 means a pixel in contact with the right side of TR + 1.
  • Methods of calculating a value of the first neighboring pixel BR are not limited to the above five examples, and may include various methods of calculating using reference pixels of the current block.
  • the value of the first adjacent pixel BR is calculated by the same method in the image encoding apparatus and the decoding apparatus.
  • the image encoding apparatus and the decoding apparatus may set the same value of the first neighboring pixel BR by using an equation such as Equation (7).
  • H means the height of the current block
  • W means the width of the current block.
  • P (W, H) means the value of the first adjacent pixel BR
  • R (W, -1) means the value of the second adjacent pixel TR
  • R (-1, H) is the zero It means the value of 3 adjacent pixels BL.
  • R (W-1, -1), R (W-2, -1), R (W + 1, -1) and R (W + 2, -1) are adjacent pixels of the second adjacent pixel TR
  • R (-1, H-1), R (-1, H-2), R (-1, H + 1) and R (-1, H + 2) are the third adjacent pixels BL ) Means adjacent pixels.
  • the image encoding apparatus and the decoding apparatus may set the same value of the first neighboring pixel BR by using an equation such as Equation (8).
  • H means the height of the current block
  • W means the width of the current block
  • P (W, H) means the value of the first adjacent pixel BR
  • R (W, -1) means the value of the second adjacent pixel TR
  • R (-1, H) is the zero It means the value of 3 adjacent pixels BL.
  • Each weight is determined by the same method as the image encoding apparatus and the decoding apparatus according to the characteristics of the image. For example, c w may use the width of the current block and c h may use the height of the current block as a weight.
  • the apparatus for encoding an image selects a candidate value having a value most similar to the original value of the first neighboring pixel BR among candidate values of the plurality of predefined first neighboring pixels BR, and is selected.
  • Information indicating a may be transmitted to the image decoding apparatus.
  • the image decoding apparatus decodes information indicating one of the candidate values of the plurality of first adjacent pixels BR from the bitstream, and determines the candidate value of the first adjacent pixel BR indicated by the decoded information. Can be set to a value of (BR).
  • the br_idx syntax element may be used as information indicating a candidate value of the first adjacent pixel BR. Table 1 shows an example of candidate values of br_idx and the first adjacent pixel BR.
  • br_idx Associated candidate 0 Average value of TR and BL One Average of TR, TR + 1, BL, BL + 1 values 2 Average of TR, TR-1, TR + 1, BL, BL-1, BL + 1 values 3 Average value of TR, TR + 1, TR + 2, BL, BL + 1, BL + 2 values
  • the image decoding apparatus When the image encoding apparatus transmits the selected br_idx value to the image decoding apparatus, the image decoding apparatus obtains a candidate value corresponding to the transmitted br_idx and sets it to the value of the first neighboring pixel BR. For example, when the value of br_idx decoded from the bitstream by the image decoding apparatus is 1, the average value of the values of TR, TR + 1, BL, and BL + 1 may be obtained and set as the value of the first adjacent pixel BR. have.
  • the apparatus for encoding an image may predict the first neighboring pixel BR based on the original value of the first neighboring pixel BR and values of reference pixels (eg, TL, TR, BL, etc.).
  • the difference value of can be transmitted.
  • the difference value can be transmitted via the br_residual syntax element.
  • the image decoding apparatus may additionally transmit code information about the difference value in addition to the difference value.
  • the image decoding apparatus decodes information ( br_residual syntax element) representing a difference value between the original value of the first neighboring pixel BR and the predicted value of the first neighboring pixel BR from the bitstream, and obtains the same method as the image encoding apparatus.
  • a value that compensates for the decoded difference value in the predicted value of one first adjacent pixel BR may be set as the value of the first adjacent pixel BR.
  • the image decoding apparatus may separately decode code information about the difference value, and set a value compensated for the decoded code information as a value of the first adjacent pixel BR.
  • the predicted value of the first adjacent pixel BR may be obtained by the method described with reference to FIGS. 9 to 11, but is not necessarily limited thereto.
  • the value of the reconstructed second neighboring pixel TR is 55
  • the value of the reconstructed third neighboring pixel BL is 45
  • the original value of the first neighboring pixel BR is 60.
  • the average value 50 of the values of the second adjacent pixel TR and the third adjacent pixel BL is generated as the predicted value of the first adjacent pixel BR, and then the original value of the first adjacent pixel BR is generated.
  • the difference value 10 between 60 and the predicted value 50 is transmitted.
  • the image encoding apparatus additionally transmits code information (+/ ⁇ ) of the difference value 10.
  • the image decoding apparatus generates the predicted value of the first neighboring pixel BR in the same manner, adds the transmitted difference value 10, compensates the transmitted sign (+/ ⁇ ), and compensates the first neighboring pixel BR. Can produce a value of 60.
  • the predicted value of the first adjacent pixel BR may be set to a reconstructed pixel value of a specific position instead of the calculated value.
  • FIGS. 12 to 15 are exemplary diagrams of a planar prediction mode in a 360 image according to another embodiment of the present invention.
  • a method of setting a value of the first adjacent pixel BR for an image of a special format such as a 360 image will be described.
  • the current block is located on the first surface to be decoded from the 360 image encoded into the 2D image.
  • FIG. 12 (a) shows the result of projecting a 360 image in an equirectangular format.
  • One of the features of the square projection format is that the left and right boundaries of an image abut each other.
  • the values of the reconstructed pixels 1210 adjacent to the left boundary of the image are used as the values of the adjacent pixels 1220 adjacent to the right boundary of the image.
  • the value of the first adjacent pixel BR for the adjacent block may be determined.
  • the image decoding apparatus contacts the right boundary of the current block based on the 360 image and identifies the already decoded block.
  • the image decoding apparatus sets a value of a pixel in contact with the first adjacent pixel BR based on a 360 image among the pixels 1210 in the block adjacent to the left boundary of the identified block to the value of the first adjacent pixel BR. do.
  • the image decoding apparatus may set at least some of the pixels 1210 in the block adjacent to the left boundary of the identified block as part of the reference pixels 1220 of the current block.
  • FIG. 13 illustrates a case where a right boundary of the current block coincides with a right boundary of a 2D image in which a 360 image is projected in a square format.
  • the value of the first adjacent pixels BR and 1300 of the current block is in contact with the right boundary of the current block with respect to the 360 image and the value of the pixel in contact with the first adjacent pixels BR. 1300 among the decoded pixels 1210. Is set.
  • the values of the adjacent pixels 1220 on the right side of the current block are also adjacent to the right boundary of the current block on the basis of 360 images using the characteristics of the square format, and the values of the adjacent pixels 1210 on the left boundary of the already decoded block. Can be set. In this case, it is not necessary to interpolate using the first adjacent pixel BR and the second adjacent pixel TR.
  • FIG. 14 shows a 2D image of a non-compact layout without changing the layout
  • FIG. 14 shows a compact layout in which a layout is changed to a rectangular shape without spaces by rearranging faces.
  • the layout is shown. Top, Bottom, Front, Back, Right, Left represent each projection face when 360 images are projected onto the cube. The boundaries of each plane are adjacent to each other based on the 360 image.
  • each adjacent relationship in the 360 image is represented by a figure. For example, in a non-compact layout, the left border of the front plane is in contact with the right border of the left plane in a 360 image.
  • the image decoding apparatus may determine the first surface (based on 360 images). A second face 1520 that is in contact with the right boundary of 1510 and has already been decrypted is identified. The image decoding apparatus may determine a value of a pixel in contact with the first adjacent pixel BR in the 360 image among the pixels in the second surface 1520 that is in contact with the left boundary of the second surface 1520. Set to. Also, the image decoding apparatus may set at least some of the pixels 1522 in the second surface 1520 adjacent to the left boundary of the second surface 1520 as at least some of the right adjacent pixels 1512 of the current block.
  • FIG. 16 is a diagram illustrating a method of setting values of neighboring pixels adjacent to a current block according to another embodiment of the present invention.
  • the value P (W) of the right adjacent pixels adjacent to the current block is shown.
  • , y) and the values P (x, H) of lower neighboring pixels adjacent to the current block can be set.
  • P (W, y) may be set to the value of each of the reconstructed pixels positioned on the upper right side including the second adjacent pixel TR among the reference pixels of the current block.
  • P (x, H) may be set to the value of each of the reconstructed pixels located at the lower left side including the third adjacent pixel BL among the reference pixels of the current block.
  • FIGS. 17 and 18 are flowchart illustrating an intra prediction prediction method according to an embodiment of the present invention.
  • the image decoding apparatus decodes information indicating the first adjacent pixel located at the lower right corner of the current block from the bitstream and sets a value of the first adjacent pixel BR (S1710).
  • the method according to the aforementioned various embodiments may be applied to the method of setting the value of the first adjacent pixel BR.
  • the information indicating the first adjacent pixel BR may be information indicating a position of a pixel having a value most similar to the original value of the first adjacent pixel BR.
  • the pixel value at the position indicated by the information indicating the position of the pixel having the most similar value among the reference pixels of the current block may be set as the value of the first adjacent pixel BR.
  • the information indicating the position of the pixel is based on the position of any one of adjacent pixels TL located at the upper left corner of the current block, adjacent pixels TR located at the upper right corner, and adjacent pixels BL located at the lower right corner. To indicate the position of the pixel.
  • the apparatus for decoding an image decodes information indicating one of candidate values of a plurality of first neighboring pixels BR configured as an average value or a weighted average value of the reference pixels of the current block from the bitstream, and the decoded information is indicated.
  • the candidate value of one first adjacent pixel BR may be set to the value of the first adjacent pixel BR.
  • the candidate value of one first adjacent pixel indicated by the decoded information is the value most similar to the original value of the first adjacent pixel BR among the candidate values of the plurality of first adjacent pixels BR. It may mean.
  • the image decoding apparatus determines the value of the first neighboring pixel BR, the value of the second neighboring pixel TR located at the upper right corner of the current block, and the lower left of the current block.
  • the interpolation using the value of the third neighboring pixel BL located at the corner is obtained to obtain the predicted value of the lower neighboring pixel and the predicted value of the right neighboring pixel in step S1720.
  • the predicted value of the right neighboring pixel is obtained by interpolation using the value of the first neighboring pixel BR and the value of the second neighboring pixel TR, and the predicted value of the lower neighboring pixel is obtained from the first neighboring pixel BR.
  • Value is obtained by interpolation using the value and the value of the third adjacent pixel BL.
  • the left adjacent pixel and the right adjacent pixel may be positioned in the same row as the current pixel, and the upper adjacent pixel and the lower adjacent pixel may be positioned in the same column as the current pixel.
  • the image decoding apparatus obtains the predicted value of the current pixel in the current block by interpolation using the value of the left neighboring pixel of the current block, the upper neighboring pixel of the current block, the predicted value of the lower neighboring pixel, and the predicted value of the right neighboring pixel. (S1730). Specifically, the image decoding apparatus obtains a first prediction value of the current pixel by interpolation using the value of the left neighboring pixel and the predicted value of the right neighboring pixel, and currently interpolates using the value of the upper neighboring pixel and the prediction value of the lower neighboring pixel. Obtain a second predicted value of the pixel. The final prediction value of the current pixel is obtained using the first prediction value and the second prediction value. For example, the final prediction value may be an average value of the first prediction value and the second prediction value.
  • FIG. 18 is a flowchart illustrating an intra prediction prediction method according to another embodiment of the present invention.
  • the image decoding apparatus sets the value of the first adjacent pixel BR located at the lower right corner of the current block based on the reference pixels of the current block (S1810).
  • the value of the first adjacent pixel BR may be directly calculated by the image decoding apparatus in the same manner as that used by the image encoding apparatus.
  • the method of calculating a value of the first adjacent pixel BR may include methods according to the aforementioned various embodiments.
  • the image decoding apparatus may have already been restored on the same straight line to the value of the reference pixel or the adjacent pixel of the current block located on the straight line in at least one specific direction extending from the first adjacent pixel BR.
  • the first neighboring pixel BR prediction value in the corresponding direction is calculated by summing the amount of change with the value of the pixel, and the value of the first neighboring pixel BR is set as an average or a weighted average of the first neighboring pixel BR prediction values. Can be.
  • the specific direction may be at least one of a linear direction, a vertical direction, and a diagonal direction from the first adjacent pixel BR
  • the reference pixel of the current block may include a peripheral pixel located at the upper left corner of the current block, a second adjacent pixel ( TR) and at least one of the third adjacent pixels BL.
  • the apparatus for decoding an image may set the average value or weighted average value of the reference pixels of the current block to the value of the first adjacent pixel BR.
  • the apparatus for decoding an image may determine the prediction value of the first neighboring pixel BR based on the reference pixels of the current block, and indicate the difference value between the original value and the prediction value of the first neighboring pixel BR from the bitstream. The value of the first neighboring pixel BR may be determined by decoding the sum and adding the decoded difference value to the directly determined prediction value.
  • FIGS. 17 and 18 are described as sequentially executing each process, but is not necessarily limited thereto. In other words, since the process described in FIGS. 17 and 18 may be applied by changing the process or executing one or more processes in parallel, FIGS. 17 and 18 are not limited to the time series order.
  • the image encoding or decoding method according to the present embodiment described in FIGS. 17 and 18 may be implemented in a computer program and recorded on a computer-readable recording medium.
  • a computer program for recording an image encoding or decoding method according to the present embodiment and a computer readable recording medium include all kinds of recording devices that store data that can be read by a computing system.

Landscapes

  • Engineering & Computer Science (AREA)
  • Multimedia (AREA)
  • Signal Processing (AREA)
  • Compression Or Coding Systems Of Tv Signals (AREA)

Abstract

The present invention relates to image encoding or decoding for improving prediction accuracy in intra prediction. According to one aspect of the present invention, provided is an image decoding method comprising the steps of: decoding information, which indicates a first adjacent pixel located at the right lower edge of a current block from a bitstream, so as to set the value of the first adjacent pixel; acquiring the value of a lower adjacent pixel and the value of a right adjacent pixel of the current block by means of interpolation using the value of the first adjacent pixel, the value of a second adjacent pixel located at the right upper edge of the current block and the value of a third adjacent pixel located at the left lower edge of the current block; and acquiring the prediction value of a current pixel in the current block by means of interpolation using the value of a left adjacent pixel of the current block, the value of an upper adjacent pixel of the current block, the value of the lower adjacent pixel and the value of the right adjacent pixel.

Description

화면 내 예측 부호화 또는 복호화 방법 및 장치Intra prediction encoding or decoding method and apparatus
본 발명은 화면 내 예측에 있어서 예측의 정밀도를 개선하기 위한 영상 부호화 또는 복호화에 관한 것이다.The present invention relates to image encoding or decoding for improving the accuracy of prediction in intra prediction.
이 부분에 기술된 내용은 단순히 본 실시예에 대한 배경 정보를 제공할 뿐 종래 기술을 구성하는 것은 아니다.The contents described in this section merely provide background information on the present embodiment and do not constitute a prior art.
비디오 압축 표준인 HEVC(High Efficiency Video Coding)는 화면 내 예측(Intra Prediction) 모드로서 2개의 비방향성 예측 모드와 33개의 방향성 예측 모드를 제공한다. 비방향성 예측 모드로는 평면(planar) 예측 모드와 DC 예측 모드가 있다. 방향성 예측 모드는 영상 내 일정한 방향성을 가진 영역에 대해서는 효과적이지만, 주변의 참조신호에 대해 연속성을 가진 평탄한 영역에 대해서는 비방향성 예측 모드가 더 효과적이다. DC 예측 모드는 현재블록에 인접 화소들의 평균값에 기반한 예측 방법으로 방향성 예측 모드의 대안으로서 제공되지만, 대략적인 근사치에 불과한 점에서 예측 정밀도가 떨어진다. 이에 따라 HEVC에서는 평면 예측 모드가 새로 도입되었다. 평면 예측 모드는 인접 화소들의 위치에 따른 거리를 반영한 2차원 평면을 생성하는 예측 방법으로 보다 정밀한 예측값을 생성할 수 있다.High Efficiency Video Coding (HEVC), a video compression standard, provides two non-directional prediction modes and 33 directional prediction modes as an intra prediction mode. Non-directional prediction modes include a planar prediction mode and a DC prediction mode. The directional prediction mode is effective for a region having a constant directionality in an image, but the non-directional prediction mode is more effective for a flat region having continuity with respect to surrounding reference signals. The DC prediction mode is provided as an alternative to the directional prediction mode as a prediction method based on the average value of neighboring pixels in the current block, but the prediction accuracy is poor in that it is only an approximation. Accordingly, the plane prediction mode is newly introduced in HEVC. Planar prediction mode is a prediction method for generating a two-dimensional plane reflecting the distance according to the position of the adjacent pixels can generate a more precise prediction value.
도 1은 종래기술에 따른 평면 예측 모드를 설명하기 위한 예시도이다.1 is an exemplary diagram for describing a planar prediction mode according to the prior art.
일반적으로 화면 내 예측 과정에서는 현재블록의 상측 행과 좌측 열에 위치하는 인접화소들을 참조화소로 이용하여 현재블록의 예측값을 생성한다. 여기서, 참조화소로 이용되는 인접화소들은 이미 예측과 복원으로 재구성된 화소들이다. CTU(Coding Tree Unit)내 CU(Coding Unit)들은 Z-스캔 순서로 처리되기 때문에, 화면 내 예측 과정에서 현재블록의 우측 열과 하측 행의 인접화소들은 참조화소로서 사용될 수 없다.In the intra prediction process, a prediction value of the current block is generated by using adjacent pixels located in the upper row and the left column of the current block as reference pixels. Here, neighboring pixels used as reference pixels are pixels reconstructed by prediction and reconstruction. Since coding units (CUs) in a coding tree unit (CTU) are processed in a Z-scan order, adjacent pixels in the right column and the lower row of the current block cannot be used as reference pixels in the intra prediction process.
도 1을 참조하면, 종래의 평면 예측 모드에서는 우선 현재블록(100)의 우측 열에 위치한 모든 인접화소들의 값을 우측 상단 모서리에 위치한 인접화소(TR, 112)의 값과 동일하게 설정하고, 현재블록(100)의 하측 행에 위치한 모든 인접화소들의 값을 좌측 하단 모서리에 위치한 인접화소(BL, 116)의 값과 동일하게 설정한다. 이에 따라, 현재블록(100)을 둘러싸고 있는 모든 인접화소들의 값이 결정되면, 현재화소(110)에 대하여 수평 방향 및 수직 방향으로 이중 선형 보간(Bilinear Interpolation)을 수행한다. 구체적으로, 도 1의 (a)와 같이 현재화소(110)와 동일한 행에 위치하는 인접화소들(114, 120)을 이용한 선형 보간으로 현재화소(100)의 제1 예측값을 생성하고, 도 1의 (b)와 같이 현재화소(110)와 동일한 열에 위치하는 인접화소들(118, 122)을 이용한 선형 보간으로 현재화소(100)의 제2 예측값을 생성한다. 현재화소(100)의 최종 예측값은 제1 예측값과 제2 예측값의 평균값으로 결정된다.Referring to FIG. 1, in the planar prediction mode of the related art, first, the values of all adjacent pixels located in the right column of the current block 100 are set to be the same as the values of the adjacent pixels TR located in the upper right corner. The value of all adjacent pixels located in the lower row of (100) is set equal to the values of the adjacent pixels BL and 116 located in the lower left corner. Accordingly, when the values of all adjacent pixels surrounding the current block 100 are determined, bilinear interpolation is performed in the horizontal direction and the vertical direction with respect to the current pixel 110. Specifically, as shown in FIG. 1A, a first prediction value of the current pixel 100 is generated by linear interpolation using adjacent pixels 114 and 120 positioned in the same row as the current pixel 110, and FIG. 1. As shown in (b) of FIG. 2, a second prediction value of the current pixel 100 is generated by linear interpolation using adjacent pixels 118 and 122 positioned in the same column as the current pixel 110. The final prediction value of the current pixel 100 is determined as an average value of the first prediction value and the second prediction value.
이러한 종래의 평면 예측 모드는 현재블록(100)의 우측 열에 위치하는 모든 인접화소들의 값과 하측 행에 위치하는 모든 인접화소들의 값을 각각 동일한 값으로 설정하기 때문에, 현재블록(100)의 우측 하단 방향으로 갈수록 화소 예측값의 정확도가 떨어지는 경향을 보인다.Since the conventional planar prediction mode sets the values of all adjacent pixels located in the right column of the current block 100 and the values of all adjacent pixels located in the lower row, respectively, to the same value, the lower right corner of the current block 100. As the direction increases, the accuracy of the pixel prediction value tends to decrease.
본 발명은 화면 내 예측의 정밀도를 높이기 위한 영상 부호화 또는 복호화 기술을 제공한다.The present invention provides an image encoding or decoding technique for increasing the precision of intra prediction.
본 발명의 일 측면에 따르면, 비트스트림으로부터 현재블록의 우측 하단 모서리에 위치한 제1 인접화소를 지시하는 정보를 복호화하여 상기 제1 인접화소의 값을 설정하는 단계; 상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 획득하는 단계; 및 상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 획득하는 단계를 포함하는 영상 복호화 방법을 제공한다.According to an aspect of the present invention, the method comprising: setting a value of the first neighboring pixel by decoding information indicating a first neighboring pixel located at a lower right corner of the current block from a bitstream; The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; Obtaining a value of a lower adjacent pixel and a value of a right adjacent pixel of the pixel; And obtaining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel. It provides a video decoding method comprising the step.
본 발명의 다른 측면에 따르면, 현재블록의 참조화소들을 기반으로 상기 현재블록의 우측 하단 모서리에 위치한 제1 인접화소의 값을 설정하는 단계; 상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 획득하는 단계; 및 상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 획득하는 단계를 포함하는 영상 복호화 방법을 제공한다.According to another aspect of the invention, the step of setting the value of the first adjacent pixel located in the lower right corner of the current block based on the reference pixels of the current block; The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; Obtaining a value of a lower adjacent pixel and a value of a right adjacent pixel of the pixel; And obtaining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel. It provides a video decoding method comprising the step.
본 발명의 또 다른 측면에 따르면, 비트스트림으로부터 현재블록의 우측 하단 모서리에 위치한 제1 인접화소를 지시하는 정보를 복호화하는 복호화부; 상기 제1 인접화소를 지시하는 정보를 이용하여 상기 제1 인접화소의 값을 설정하는 제1 인접화소 설정부; 상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 결정하는 인접화소 예측부; 및 상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 결정하는 현재화소 예측부를 포함하는 영상 복호화 장치를 제공한다.According to another aspect of the invention, the decoding unit for decoding the information indicating the first adjacent pixel located in the lower right corner of the current block from the bitstream; A first neighboring pixel setting unit configured to set a value of the first neighboring pixel by using information indicating the first neighboring pixel; The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; A neighboring pixel predictor configured to determine a value of a lower neighboring pixel and a value of a right neighboring pixel of the pixel; And determining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel. An image decoding apparatus including a current pixel prediction unit is provided.
본 발명의 또 다른 측면에 따르면, 현재블록의 참조화소들을 기반으로 상기 현재블록의 우측 하단 모서리에 위치한 제1 인접화소의 값을 설정하는 제1 인접화소 설정부; 상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 결정하는 인접화소 예측부; 및 상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 결정하는 현재화소 예측부를 포함하는 영상 복호화 장치를 제공한다.According to another aspect of the invention, the first adjacent pixel setting unit for setting the value of the first adjacent pixel located in the lower right corner of the current block based on the reference pixels of the current block; The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; A neighboring pixel predictor configured to determine a value of a lower neighboring pixel and a value of a right neighboring pixel of the pixel; And determining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel. An image decoding apparatus including a current pixel prediction unit is provided.
도 1은 종래기술에 따른 평면 예측 모드를 설명하기 위한 예시도이다.1 is an exemplary diagram for describing a planar prediction mode according to the prior art.
도 2는 본 발명의 실시예에 따른 영상 부호화 장치에 대한 블록도이다.2 is a block diagram of an image encoding apparatus according to an embodiment of the present invention.
도 3은 본 발명의 실시예에 따른 영상 복호화 장치에 대한 블록도이다.3 is a block diagram of an image decoding apparatus according to an embodiment of the present invention.
도 4는 본 발명의 실시예에 따른 평면 예측 모드로 화면 내 예측을 수행하기 위한 장치의 블록도이다.4 is a block diagram of an apparatus for performing intra prediction in a planar prediction mode according to an embodiment of the present invention.
도 5는 본 발명의 실시예에 따른 평면 예측 모드를 설명하기 위한 예시도이다.5 is an exemplary diagram for describing a planar prediction mode according to an embodiment of the present invention.
도 6 내지 도 8은 본 발명의 실시예에 따라서 비트스트림으로부터 복호화한 정보를 이용하는 평면 예측 모드의 예시도이다.6 through 8 are exemplary diagrams of a planar prediction mode using information decoded from a bitstream according to an embodiment of the present invention.
도 9 내지 도 11은 본 발명의 다른 실시예에 따라서 현재블록의 인접화소들을 기초로 계산된 값을 이용하는 평면 예측 모드의 예시도이다.9 to 11 are exemplary views of planar prediction mode using a value calculated based on neighboring pixels of the current block according to another embodiment of the present invention.
도 12 내지 도 15는 본 발명의 다른 실시예에 따른 360 영상에서의 평면 예측 모드의 예시도이다.12 to 15 are exemplary diagrams of a planar prediction mode in a 360 image according to another embodiment of the present invention.
도 16은 본 발명의 다른 실시예에 따라서 현재블록의 인접화소들의 값이 설정되는 방법을 나타내는 예시도이다.16 is an exemplary diagram illustrating a method of setting values of neighboring pixels of a current block according to another embodiment of the present invention.
도 17은 본 발명의 실시예에 따른 화면 내 예측 복호화 방법을 나타내는 흐름도이다.17 is a flowchart illustrating an intra prediction prediction method according to an embodiment of the present invention.
도 18은 본 발명의 다른 실시예에 따른 화면 내 예측 복호화 방법을 나타내는 흐름도이다.18 is a flowchart illustrating an intra prediction prediction method according to another embodiment of the present invention.
본 발명의 구체적인 설명에 앞서, 본 명세서에서 사용되는 용어들에 대해 해석 가능한 의미를 제시한다. 그러나 이하 제시하는 의미는 예시에 불과하며 용어들의 의미가 이에 한정되는 것이 아님을 주의해야 한다.Prior to the detailed description of the present invention, the meanings interpreted for the terms used herein are presented. However, it should be noted that the meanings set forth below are merely examples and the meanings of the terms are not limited thereto.
현재블록은 부호화 또는 복호화하고자 하는 CU(Coding Unit) 또는 PU(Prediction Unit) 등을 지칭할 수 있다. 현재화소는 화면 내 예측 모드에 따라 예측값을 생성하고자 하는 현재블록 내의 화소를 지칭할 수 있다. 인접화소는 현재블록에 바로 인접하는 화소를 지칭할 수 있다. 인접화소는 복원된 화소값일 수도 있고, 다른 화소값을 이용하여 예측된 값일 수도 있다. 주변화소는 현재블록에 인접하지는 않고 주변에 위치한 복원된 화소를 지칭할 수 있다. 참조화소는 현재화소의 예측값을 생성하는데 참조되는 복원된 화소를 지칭할 수 있으며, 현재블록에 인접하는 화소(인접화소) 뿐만 아니라 인접하지 않은 화소(주변화소)도 포함될 수 있다. 한편, '화소'라는 용어는 픽셀, 샘플 등의 용어와 상호 교환 가능하다.The current block may refer to a coding unit (CU) or a prediction unit (PU) to be encoded or decoded. The current pixel may refer to a pixel in a current block for which a prediction value is to be generated according to an intra prediction mode. The adjacent pixel may refer to a pixel immediately adjacent to the current block. The adjacent pixel may be a reconstructed pixel value or may be a value predicted using another pixel value. The peripheral pixel may refer to the reconstructed pixel that is not adjacent to the current block but located in the peripheral. The reference pixel may refer to a reconstructed pixel that is referred to to generate a prediction value of the current pixel, and may include not only pixels (adjacent pixels) adjacent to the current block but also non-adjacent pixels (main pixel). Meanwhile, the term 'pixel' is interchangeable with terms such as pixel and sample.
이하, 본 발명의 일부 실시예들을 예시적인 도면을 통해 상세하게 설명한다. 각 도면의 구성 요소들에 식별 부호를 부가함에 있어서, 동일한 구성요소들에 대해서는 비록 다른 도면상에 표시되더라도 가능한 한 동일한 부호를 가지도록 하고 있음에 유의해야 한다. 또한, 본 발명을 설명함에 있어, 관련된 공지 구성 또는 기능에 대한 구체적인 설명이 본 발명의 요지를 흐릴 수 있다고 판단되는 경우에는 그 상세한 설명은 생략한다.Hereinafter, some embodiments of the present invention will be described in detail through exemplary drawings. In adding an identification code to the components of each drawing, it should be noted that the same components as possible, even if shown on different drawings have the same reference numerals. In addition, in describing the present invention, when it is determined that the detailed description of the related well-known configuration or function may obscure the gist of the present invention, the detailed description thereof will be omitted.
도 2는 본 발명의 실시예에 따른 영상 부호화 장치(200)에 대한 블록도이다. 2 is a block diagram of an image encoding apparatus 200 according to an embodiment of the present invention.
영상 부호화 장치(200)는 블록 분할부(210), 예측부(220), 감산기(230), 변환부(240), 양자화부(245), 부호화부(250), 역양자화부(260), 역변환부(265), 가산기(270), 필터부(280) 및 메모리(290)를 포함한다. 부호화 장치(200)의 각 구성요소는 하드웨어 칩으로 구현될 수 있으며, 또는 소프트웨어로 구현되고 마이크로프로세서가 각 구성요소에 대응하는 소프트웨어의 기능을 실행하도록 구현될 수도 있다.The image encoding apparatus 200 may include a block divider 210, a predictor 220, a subtractor 230, a transformer 240, a quantizer 245, an encoder 250, an inverse quantizer 260, An inverse transform unit 265, an adder 270, a filter unit 280, and a memory 290 are included. Each component of the encoding apparatus 200 may be implemented by a hardware chip, or may be implemented by software and a microprocessor to execute a function of software corresponding to each component.
블록 분할부(210)는 영상을 구성하는 각 픽처(picture)를 복수의 CTU(Coding Tree Unit)로 분할한 이후에, 트리 구조(tree structure)를 이용하여 CTU를 반복적으로(recursively) 분할한다. 트리 구조에서의 리프 노드(leaf node)가 부호화의 기본 단위인 CU(Coding Unit)가 된다. 트리 구조로는 상위 노드가 네 개의 하위 노드로 분할하는 쿼드트리(QuadTree, QT), 또는 QT 구조 및 상위 노드가 두 개의 하위 노드로 분할하는 바이너리트리(BinaryTree, BT) 구조를 혼용한 QTBT (QuadTree plus BinaryTree) 구조가 사용될 수 있다.After dividing each picture constituting the image into a plurality of coding tree units (CTUs), the block dividing unit 210 recursively divides the CTUs using a tree structure. A leaf node in the tree structure becomes a coding unit (CU) which is a basic unit of coding. As a tree structure, QuadTree (QT), which the parent node splits into four child nodes, or QTBT (QuadTree), which uses a QT structure and a binaryTree (BT) structure that the parent node splits into two child nodes. plus BinaryTree) structure can be used.
예측부(220)는 현재블록을 예측하여 예측블록을 생성한다. 예측부(220)는 인트라 예측부(222)와 인터 예측부(224)를 포함한다. 여기서, 현재블록은 상기의 트리 구조에서 리프 노드에 해당하는 부호화의 기본 단위로서, 현재 부호화하고자 하는 CU를 의미한다. 대안적으로, 현재블록은 CU로부터 분할된 복수의 서브블록 중 하나의 서브블록일 수도 있다.The prediction unit 220 generates a prediction block by predicting the current block. The predictor 220 includes an intra predictor 222 and an inter predictor 224. Here, the current block is a basic unit of encoding corresponding to a leaf node in the tree structure, and means a CU to be currently encoded. Alternatively, the current block may be one subblock of the plurality of subblocks divided from the CU.
인트라 예측부(222)는 현재블록이 포함된 현재 픽처 내에서 현재블록의 주변에 위치한 화소(참조화소)들을 이용하여 현재블록 내의 화소들을 예측한다. 예측 방향에 따라 복수의 인트라 예측모드가 존재하며, 각 예측모드에 따라 사용할 참조화소와 연산식이 다르게 정의된다. 특히, 인트라 예측부(222)는 현재 블록을 부호화하는데 사용할 인트라 예측 모드를 결정할 수 있다. 일부 예들에서, 인트라 예측부(222)는 여러 인트라 예측 모드들을 사용하여 현재 블록을 인코딩하고, 테스트된 모드들로부터 사용할 적절한 인트라 예측 모드를 선택할 수도 있다. 예를 들어, 인트라 예측부(222)는 여러 테스트된 인트라 예측 모드들에 대한 레이트 왜곡(rate-distortion) 분석을 사용하여 레이트 왜곡 값들을 계산하고, 테스트된 모드들 중 최선의 레이트 왜곡 특징들을 갖는 인트라 예측 모드를 선택할 수도 있다.The intra predictor 222 predicts pixels in the current block by using pixels (reference pixels) positioned in the vicinity of the current block in the current picture including the current block. There are a plurality of intra prediction modes according to the prediction direction, and reference pixels and calculation expressions to be used are defined differently according to each prediction mode. In particular, the intra predictor 222 may determine an intra prediction mode to be used to encode the current block. In some examples, intra prediction unit 222 may encode the current block using several intra prediction modes and select an appropriate intra prediction mode to use from the tested modes. For example, intra predictor 222 calculates rate distortion values using rate-distortion analysis for several tested intra prediction modes, and has the best rate distortion characteristics among the tested modes. Intra prediction mode may be selected.
본 발명의 실시예에 따른 복수의 인트라 예측 모드는 2개의 비방향성 모드(평면 모드 및 DC 모드)와 65개의 방향성 모드를 포함할 수 있다. 본 발명의 실시예들에 따른 평면(planar) 모드에 대한 구체적인 설명은 다른 도면들을 참조하여 후술한다. 인트라 예측부(222)는 복수의 인트라 예측 모드 중에서 하나의 인트라 예측 모드를 선택하고, 선택된 인트라 예측 모드에 따라 결정되는 참조화소와 연산식을 사용하여 현재 블록을 예측한다. 선택된 인트라 예측 모드에 대한 정보는 부호화부(250)에 의해 부호화되어 영상 복호화 장치로 전달된다.The plurality of intra prediction modes according to the embodiment of the present invention may include two non-directional modes (plane mode and DC mode) and 65 directional modes. A detailed description of the planar mode according to embodiments of the present invention will be described later with reference to other drawings. The intra prediction unit 222 selects one intra prediction mode from among the plurality of intra prediction modes, and predicts the current block by using a reference pixel and an operation formula determined according to the selected intra prediction mode. Information on the selected intra prediction mode is encoded by the encoder 250 and transmitted to the image decoding apparatus.
인터 예측부(224)는 현재 픽처보다 먼저 부호화 및 복호화된 참조 픽처 내에서 현재블록과 가장 유사한 블록을 탐색하고, 그 탐색된 블록을 이용하여 현재블록에 대한 예측블록을 생성한다. 그리고, 현재 픽처 내의 현재블록과 참조 픽처 내의 예측블록 간의 변위(displacement)에 해당하는 움직임벡터(Motion Vector: MV)를 생성한다. 현재블록을 예측하기 위해 사용된 참조 픽처에 대한 정보 및 움직임벡터에 대한 정보를 포함하는 움직임 정보는 부호화부(250)에 의해 부호화되어 영상 복호화 장치로 전달된다.The inter prediction unit 224 searches for the block most similar to the current block in the coded and decoded reference picture before the current picture, and generates a prediction block for the current block using the searched block. A motion vector (MV) corresponding to a displacement between the current block in the current picture and the prediction block in the reference picture is generated. Motion information including information about a reference picture and information about a motion vector used to predict the current block is encoded by the encoder 250 and transmitted to the image decoding apparatus.
감산기(230)는 현재블록으로부터 인트라 예측부(222) 또는 인터 예측부(224)에 의해 생성된 예측블록을 감산하여 잔차 블록을 생성한다.The subtractor 230 subtracts the prediction block generated by the intra predictor 222 or the inter predictor 224 from the current block to generate a residual block.
변환부(240)는 공간 영역의 화소 값들을 가지는 잔차 블록 내의 잔차 신호를 주파수 도메인의 변환 계수로 변환한다. 변환부(240)는 잔차 블록 내의 잔차 신호들을 현재블록의 크기를 변환 단위로 사용하여 변환할 수 있으며, 또는 잔차 블록을 더 작은 복수의 서브블록을 분할하고 서브블록 크기의 변환 단위로 잔차 신호들을 변환할 수도 있다. 잔차 블록을 더 작은 서브블록으로 분할하는 방법은 다양하게 존재할 수 있다. 예컨대, 기 정의된 동일한 크기의 서브블록으로 분할할 수도 있으며, 또는 잔차 블록을 루트 노드로 하는 QT(quadtree) 방식의 분할을 사용할 수도 있다.The converter 240 converts the residual signal in the residual block having pixel values of the spatial domain into a transform coefficient of the frequency domain. The transform unit 240 may convert the residual signals in the residual block using the size of the current block as a conversion unit, or divide the residual block into a plurality of smaller subblocks and convert the residual signals in a subblock-sized transform unit. You can also convert. There may be various ways of dividing the residual block into smaller subblocks. For example, the subblock may be divided into sub-blocks having the same size, or may be divided by a quadtree (QT) method using the residual block as a root node.
양자화부(245)는 변환부(240)로부터 출력되는 변환 계수들을 양자화하고, 양자화된 변환 계수들을 부호화부(250)로 출력한다.The quantization unit 245 quantizes the transform coefficients output from the transform unit 240, and outputs the quantized transform coefficients to the encoder 250.
부호화부(250)는 양자화된 변환 계수들을 CABAC 등의 부호화 방식을 사용하여 부호화하여 비트스트림을 생성한다. 또한, 부호화부(250)는 트리 구조의 최상위 레이어에 위치하는 CTU의 크기에 대한 정보 및 CTU로부터 트리 구조로 블록을 분할하기 위한 분할 정보를 부호화하여, 복호화 장치가 부호화 장치와 동일하게 블록을 분할할 수 있도록 한다. 예컨대, QT 분할의 경우에는 상위 레이어의 블록이 하위 레이어의 네 개의 블록으로 분할되는지 여부를 지시하는 QT 분할 정보를 부호화한다. BT 분할의 경우에는 QT의 리프 노드에 해당하는 블록으로부터 시작하여 각 블록이 두 개의 블록으로 분할되는지 여부 및 분할되는 타입을 지시하는 BT 분할 정보를 부호화한다.The encoder 250 generates a bitstream by encoding the quantized transform coefficients by using an encoding method such as CABAC. In addition, the encoder 250 encodes information about the size of the CTU located in the highest layer of the tree structure and split information for dividing the block into the tree structure from the CTU, so that the decoding apparatus divides the block in the same way as the encoding apparatus. Do it. For example, in the case of QT splitting, QT splitting information indicating whether a block of an upper layer is divided into four blocks of a lower layer is encoded. In the case of BT partitioning, BT partitioning information indicating whether each block is divided into two blocks and the type of partitioning is encoded, starting from the block corresponding to the leaf node of the QT.
부호화부(250)는 현재블록이 인트라 예측에 의해 부호화되었는지 아니면 인터 예측에 의해 부호화되었는지 여부를 지시하는 예측 타입에 대한 정보를 부호화하고, 예측 타입에 따라 인트라 예측정보 또는 인터 예측정보를 부호화한다.The encoder 250 encodes information about a prediction type indicating whether the current block is encoded by intra prediction or inter prediction, and encodes intra prediction information or inter prediction information according to the prediction type.
역양자화부(260)는 양자화부(245)로부터 출력되는 양자화된 변환 계수들을 역양자화하여 변환 계수들을 생성한다. 역변환부(265)는 역양자화부(260)로부터 출력되는 변환 계수들을 주파수 도메인으로부터 공간 도메인으로 변환하여 잔차블록을 복원한다.The inverse quantizer 260 inversely quantizes the quantized transform coefficients output from the quantizer 245 to generate transform coefficients. The inverse transformer 265 restores the residual block by converting the transform coefficients output from the inverse quantizer 260 from the frequency domain to the spatial domain.
가산기(270)는 복원된 잔차블록과 예측부(220)에 의해 생성된 예측블록을 가산하여 현재블록을 복원한다. 복원된 현재블록 내의 화소들은 다음 순서의 블록을 인트라 예측할 때 참조화소로서 사용된다.The adder 270 reconstructs the current block by adding the reconstructed residual block and the prediction block generated by the prediction unit 220. The pixels in the reconstructed current block are used as reference pixels when intra prediction of blocks of the next order.
필터부(280)는 블록 단위의 부호화/복호화로 인해 발생하는 블록킹 현상(blocking artifact)을 제거하기 위해 복원된 블록 간의 경계를 디블록킹 필터링하고 메모리(290)에 저장한다. 한 픽처 내의 모든 블록들이 복원되면, 복원된 픽처는 이후에 부호화하고자 하는 픽처 내의 블록을 인터 예측하기 위한 참조 픽처로 사용된다.The filter unit 280 deblocks and filters the boundary between the reconstructed blocks in order to remove blocking artifacts that occur due to encoding / decoding of blocks. When all the blocks in a picture are reconstructed, the reconstructed picture is used as a reference picture for inter prediction of a block in a picture to be encoded later.
이하에서는 영상 복호화 장치에 대해 설명한다.Hereinafter, an image decoding apparatus will be described.
도 3은 본 발명의 실시예에 따른 영상 복호화 장치(300)에 대한 블록도이다3 is a block diagram of an image decoding apparatus 300 according to an embodiment of the present invention.
영상 복호화 장치(300)는 복호화부(310), 역양자화부(320), 역변환부(330), 예측부(340), 가산기(350), 필터부(360) 및 메모리(370)를 포함한다. 도 3에 도시된 구성요소들은 하드웨어 칩으로 구현될 수 있으며, 또는 소프트웨어로 구현되고 마이크로프로세서가 각 구성요소에 대응하는 소프트웨어의 기능을 실행하도록 구현될 수도 있다.The image decoding apparatus 300 includes a decoder 310, an inverse quantizer 320, an inverse transformer 330, a predictor 340, an adder 350, a filter 360, and a memory 370. . The components shown in FIG. 3 may be implemented in a hardware chip or may be implemented in software and implemented so that the microprocessor executes a function of software corresponding to each component.
복호화부(310)는 영상 부호화 장치로부터 수신한 비트스트림을 복호화하여 블록분할과 관련된 정보를 추출하여 복호화하고자 하는 현재블록을 결정하고, 현재블록을 복원하기 위해 필요한 예측정보와 잔차신호에 대한 정보 등을 추출한다.The decoder 310 decodes the bitstream received from the image encoding apparatus, extracts information related to block division, determines a current block to be decoded, and includes prediction information and residual signal information necessary for reconstructing the current block. Extract
복호화부(310)는 SPS (Sequence Parameter Set) 또는 PPS (Picture Parameter Set)로부터 CTU size에 대한 정보를 추출하여 CTU의 크기를 결정하고, 픽처를 결정된 크기의 CTU로 분할한다. 그리고 CTU를 트리 구조의 최상위 레이어, 즉, 루트 노드로 결정하고, CTU에 대한 분할 정보를 추출함으로써 CTU를 트리 구조를 이용하여 분할한다. 예컨대, QTBT 구조를 사용하여 CTU를 분할하는 경우, 먼저 QT의 분할과 관련된 제1 플래그(QT_split_flag)를 추출하여 각 노드를 하위 레이어의 네 개의 노드로 분할한다. 그리고, QT의 리프 노드에 해당하는 노드에 대해서는 BT의 분할과 관련된 제2 플래그(BT_split_flag) 및 분할 타입 정보를 추출하여 해당 리프 노드를 BT 구조로 분할한다.The decoder 310 extracts information on the CTU size from a Sequence Parameter Set (SPS) or Picture Parameter Set (PPS) to determine the size of the CTU, and divides the picture into a CTU of the determined size. The CTU is determined as the highest layer of the tree structure, that is, the root node, and the CTU is partitioned using the tree structure by extracting partition information about the CTU. For example, when splitting a CTU using a QTBT structure, first, a first flag (QT_split_flag) related to splitting of QT is extracted, and each node is divided into four nodes of a lower layer. For the node corresponding to the leaf node of the QT, the second flag BT_split_flag and the split type information related to the splitting of the BT are extracted to split the corresponding leaf node into the BT structure.
복호화부(310)는 트리 구조의 분할을 통해 복호화하고자 하는 현재블록(현재블록)을 결정하게 되면, 현재블록이 인트라 예측되었는지 아니면 인터 예측되었는지를 지시하는 예측 타입에 대한 정보를 추출한다.When the decoder 310 determines the current block (current block) to be decoded by splitting the tree structure, the decoder 310 extracts information about a prediction type indicating whether the current block is intra predicted or inter predicted.
예측 타입 정보가 인트라 예측을 지시하는 경우, 복호화부(310)는 현재 블록의 인트라 예측정보(예: 인트라 예측 모드, 참조화소에 대한 정보 등)에 대한 신택스 요소(syntax element)를 추출한다. 예측 타입 정보가 인터 예측을 지시하는 경우, 복호화부(310)는 인터 예측정보(예: 인터 예측 모드 등)에 대한 신택스 요소를 추출한다. When the prediction type information indicates intra prediction, the decoder 310 extracts a syntax element for intra prediction information (eg, intra prediction mode, information on a reference pixel, etc.) of the current block. When the prediction type information indicates inter prediction, the decoder 310 extracts a syntax element for inter prediction information (eg, inter prediction mode).
한편, 복호화부(310)는 잔차신호에 대한 정보로서 현재블록의 양자화된 변환계수들에 대한 정보를 추출한다.Meanwhile, the decoder 310 extracts information on the quantized transform coefficients of the current block as information on the residual signal.
역양자화부(320)는 양자화된 변환계수들을 역양자화하고, 역변환부(330)는 역양자화된 변환계수들을 주파수 도메인으로부터 공간 도메인으로 역변환하여 잔차신호들을 복원함으로써 현재블록에 대한 잔차블록을 생성한다.The inverse quantizer 320 inversely quantizes the quantized transform coefficients, and the inverse transformer 330 inversely transforms the inverse quantized transform coefficients from the frequency domain to the spatial domain to generate a residual block for the current block. .
예측부(340)는 인트라 예측부(342) 및 인터 예측부(344)를 포함한다. 인트라 예측부(342)는 현재블록의 예측 타입인 인트라 예측일 때 활성화되고, 인터 예측부(344)는 현재블록의 예측 타입인 인트라 예측일 때 활성화된다.The predictor 340 includes an intra predictor 342 and an inter predictor 344. The intra predictor 342 is activated when the intra prediction is the prediction type of the current block, and the inter predictor 344 is activated when the intra prediction is the prediction type of the current block.
인트라 예측부(342)는 복호화부(310)로부터 추출된 인트라 예측 모드에 대한 신택스 요소로부터 복수의 인트라 예측 모드 중 현재블록의 인트라 예측 모드를 결정하고, 인트라 예측 모드에 따라 현재블록 주변의 참조화소들을 이용하여 현재블록을 예측한다. 또한, 인트라 예측부(342)는 복호화부(310)로부터 추출된 참조화소에 대한 신택스 요소로부터 인트라 예측에 이용할 참조화소의 값 등을 설정할 수 있다.The intra prediction unit 342 determines the intra prediction mode of the current block among the plurality of intra prediction modes from the syntax elements for the intra prediction mode extracted from the decoder 310, and references the reference pixels around the current block according to the intra prediction mode. Predict the current block using In addition, the intra predictor 342 may set a value of a reference pixel to be used for intra prediction from a syntax element for the reference pixel extracted from the decoder 310.
인터 예측부(344)는 복호화부(310)로부터 추출된 인터 예측 모드에 대한 신택스 요소를 이용하여 현재블록의 움직임정보를 결정하고, 결정된 움직임정보를 이용하여 현재블록을 예측한다.The inter prediction unit 344 determines motion information of the current block using a syntax element of the inter prediction mode extracted from the decoder 310, and predicts the current block using the determined motion information.
가산기(350)는 역변환부(330)로부터 출력되는 잔차블록과 인터 예측부(344) 또는 인트라 예측부(342)로부터 출력되는 예측블록을 가산하여 현재 블록을 복원한다. 복원된 현재 블록 내의 화소들은 이후에 복호화할 블록을 인트라 예측할 때의 참조화소로서 활용된다.The adder 350 reconstructs the current block by adding the residual block output from the inverse transformer 330 and the prediction block output from the inter predictor 344 or the intra predictor 342. The pixels in the reconstructed current block are utilized as reference pixels in intra prediction of a block to be subsequently decoded.
필터부(360)는 블록 단위의 복호화로 인해 발생하는 블록킹 현상(blocking artifact)를 제거하기 위해 복원된 블록 간의 경계를 디블록킹 필터링하고 메모리(370)에 저장한다. 한 픽처 내의 모든 블록들이 복원되면, 복원된 픽처는 이후에 복호화하고자 하는 픽처 내의 블록을 인터 예측하기 위한 참조 픽처로 사용된다.The filter unit 360 deblocks and filters the boundary between the reconstructed blocks in order to remove blocking artifacts caused by block-by-block decoding, and stores them in the memory 370. When all the blocks in a picture are reconstructed, the reconstructed picture is used as a reference picture for inter prediction of a block in a picture to be decoded later.
이하, 도 4 및 도 5를 참조하여 본 발명의 실시예에 따른 평면 예측 모드에 대하여 구체적으로 설명한다.Hereinafter, a planar prediction mode according to an embodiment of the present invention will be described in detail with reference to FIGS. 4 and 5.
도 4는 본 발명의 실시예에 따른 평면 예측 모드로 화면 내 예측을 수행하기 위한 장치(400)의 블록도이다. 본 장치(400)는 제1 인접화소 설정부(410), 인접화소 예측부(420) 및 현재화소 예측부(430)를 포함하며, 도 3의 인터 예측부(344)에 구현될 수 있다. 도 4에 도시된 구성요소들은 하드웨어 칩으로 구현될 수 있으며, 또는 소프트웨어로 구현되고 마이크로프로세서가 각 구성요소에 대응하는 소프트웨어의 기능을 실행하도록 구현될 수도 있다.4 is a block diagram of an apparatus 400 for performing intra prediction in a planar prediction mode according to an exemplary embodiment of the present invention. The apparatus 400 may include a first neighboring pixel setting unit 410, a neighboring pixel prediction unit 420, and a current pixel prediction unit 430, and may be implemented in the inter prediction unit 344 of FIG. 3. The components shown in FIG. 4 may be implemented as a hardware chip, or may be implemented in software and such that the microprocessor executes the functions of the software corresponding to each component.
우선 도 4의 장치(400)에 대해 구체적으로 설명하기 전에 도 5를 참조하여 본 발명의 실시예에 따른 평면 예측 모드에 대해 간략히 설명한다. 도 5는 본 발명의 실시예에 따른 평면 예측 모드를 설명하기 위한 예시도이다.Before describing the apparatus 400 of FIG. 4 in detail, the planar prediction mode according to the embodiment of the present invention will be briefly described with reference to FIG. 5. 5 is an exemplary diagram for describing a planar prediction mode according to an embodiment of the present invention.
본 발명에서 제안하는 평면 예측 모드는 도 5의 (a) 및 (b)와 같이 현재화소(510)의 수직방향 및 수평방향으로 이중 보간을 수행하여 현재화소(510)를 예측한다. 다만, 종래기술과 달리 본 발명은 현재블록(500)의 우측 하단 모서리에 위치하는 인접화소(BR, 이하 '제1 인접화소'라 칭함)의 값을 먼저 설정하고, 이 값을 기반으로 현재블록(500)의 하측 인접화소들 및 우측 인접화소들을 예측한다. 본 실시예에서는 제1 인접화소값은 우측 상단 모서리에 위치한 인접화소(TR, 이하 '제2 인접화소'라 칭함)의 값과 좌측 하단 모서리에 위치한 인접화소(BL, 이하 '제3 인접화소'라 칭함)의 값을 이용하여 설정될 수 있다.The planar prediction mode proposed in the present invention predicts the current pixel 510 by performing double interpolation in the vertical and horizontal directions of the current pixel 510 as shown in FIGS. 5A and 5B. However, unlike the prior art, the present invention first sets the value of the adjacent pixel (BR, hereinafter referred to as 'first adjacent pixel') located at the lower right corner of the current block 500, and based on this value, The lower neighboring pixels and the right neighboring pixels of 500 are predicted. In the present exemplary embodiment, the first adjacent pixel value is the value of the adjacent pixel (TR, hereinafter referred to as 'second adjacent pixel') located at the upper right corner, and the adjacent pixel (BL, hereinafter referred to as 'third adjacent pixel') located at the lower left corner. It can be set using a value of).
즉, 도 5의 (a)를 참조하면, 종래에는 현재블록(500)의 우측 인접화소(512)의 값이 우측 상단 모서리에 위치한 제2 인접화소의 값과 동일하게 설정되나, 본 실시예에서는 제2 인접화소(TR)의 값과 제1 인접화소(BR)의 값을 이용한 보간에 의해 획득된 값으로 설정된다.도 5의 (b)를 참조하면, 종래에는 현재블록(500)의 하측 인접화소(516)의 값이 좌측 하단 모서리에 위치한 제3 인접화소의 값과 동일하게 설정되나, 본 실시예에서는 제3 인접화소(BL)의 값과 제1 인접화소(BR)의 값을 이용한 보간에 의해 획득된 값으로 설정된다. 일례로 선형 보간을 사용할 경우, 현재블록(500)의 우측 인접화소(512)의 값과 하측 인접화소(516)의 값을 구하는 과정을 수학식 1과 같이 나타낼 수 있다.That is, referring to FIG. 5A, in the related art, the value of the right adjacent pixel 512 of the current block 500 is set equal to the value of the second adjacent pixel located at the upper right corner. It is set to a value obtained by interpolation using the value of the second neighboring pixel TR and the value of the first neighboring pixel BR. Referring to FIG. 5B, the lower side of the current block 500 is conventionally referred to. The value of the adjacent pixel 516 is set to be the same as the value of the third adjacent pixel located at the lower left corner, but in this embodiment, the value of the third adjacent pixel BL and the value of the first adjacent pixel BR are used. It is set to the value obtained by interpolation. For example, when linear interpolation is used, a process of obtaining the values of the right adjacent pixel 512 and the lower neighboring pixel 516 of the current block 500 may be represented by Equation 1.
Figure PCTKR2018005596-appb-M000001
Figure PCTKR2018005596-appb-M000001
여기서, H는 현재블록의 높이를 의미하며, W는 현재블록의 너비를 의미한다. P(W, y)는 현재블록의 우측 경계에 접하는 인접화소(512)의 값을 의미하고, P(x, H)는 현재블록의 하측 경계에 접하는 인접화소(516)의 값을 의미한다. P(W, H)는 제1 인접화소(BR)의 값을 의미한다. R로 표기된 화소들은 현재블록의 참조화소들이며, R(W, -1)은 제2 인접화소(TR)의 값을 의미하며, R(-1, H)는 제3 인접화소(BL)의 값을 의미한다.Here, H means the height of the current block, W means the width of the current block. P (W, y) means the value of the adjacent pixel 512 adjacent to the right boundary of the current block, and P (x, H) means the value of the adjacent pixel 516 adjacent to the lower boundary of the current block. P (W, H) means the value of the first adjacent pixel BR. Pixels denoted as R are reference pixels of the current block, R (W, -1) denotes the value of the second adjacent pixel TR, and R (-1, H) denotes the value of the third adjacent pixel BL. Means.
현재화소(510)의 제1 예측값은 현재블록(500)의 좌측 인접화소(514)의 값과 우측 인접화소(512)의 값을 이용하여 수평방향으로 보간한 값에 의해 생성된다. 현재화소(510)의 제2 예측값은 현재블록(500)의 하측 인접화소(516)의 값과 상측 인접화소(518)의 값을 이용하여 수직방향으로 보간한 값에 의해 생성된다. 일례로 현재화소(510)의 제1 예측값과 제2 예측값은 선형 보간을 이용하여 구할 경우 수학식 2로 표현할 수 있다.The first prediction value of the current pixel 510 is generated by a value interpolated in the horizontal direction using the value of the left adjacent pixel 514 and the right adjacent pixel 512 of the current block 500. The second prediction value of the current pixel 510 is generated by a value interpolated in the vertical direction by using the value of the lower neighboring pixel 516 and the upper neighboring pixel 518 of the current block 500. For example, the first prediction value and the second prediction value of the current pixel 510 may be expressed by Equation 2 when obtained by using linear interpolation.
Figure PCTKR2018005596-appb-M000002
Figure PCTKR2018005596-appb-M000002
여기서, P1(x, y)는 수평방향으로 보간된 (x, y) 위치의 제1 예측값을 의미하고, P2(x, y)는 수직방향으로 보간된 (x, y) 위치의 제2 예측값을 의미한다. H는 현재블록의 높이를 의미하며, W는 현재블록의 너비를 의미한다. R(x, -1)은 상측 인접화소의 값을 의미하며, R(-1, y)는 좌측 인접화소의 값을 의미한다.Here, P 1 (x, y) means the first predicted value of the position (x, y) interpolated in the horizontal direction, and P 2 (x, y) means the first predicted value of the position (x, y) interpolated in the vertical direction. 2 means a predicted value. H means the height of the current block, W means the width of the current block. R (x, -1) means the value of the upper neighboring pixel, and R (-1, y) means the value of the left neighboring pixel.
본 실시예에 따른 현재화소(510)의 최종 예측값은 제1 예측값과 제2 예측값을 이용하여 결정된다. 일례로 제1 예측값과 제2 예측값의 평균을 이용할 경우, 이를 수식으로 나타내면 수학식 3과 같다.The final prediction value of the current pixel 510 according to the present embodiment is determined using the first prediction value and the second prediction value. For example, in the case of using the average of the first prediction value and the second prediction value, this is expressed by Equation 3 below.
Figure PCTKR2018005596-appb-M000003
Figure PCTKR2018005596-appb-M000003
여기서, P1(x, y)는 수평방향으로 보간된 (x, y) 위치의 제1 예측값을 의미하고, P2(x, y)는 수직방향으로 보간된 (x, y) 위치의 제2 예측값을 의미한다. Pf(x, y)는 본 실시예에 따른 (x, y) 좌표의 최종 예측값을 의미한다. 수학식 3은 하나의 예시이며, 제1 예측값과 제2 예측값을 이용하여 최종 예측값을 계산하는 방법은 이에 한정되지 않는다.Here, P 1 (x, y) means the first predicted value of the position (x, y) interpolated in the horizontal direction, and P 2 (x, y) means the first predicted value of the position (x, y) interpolated in the vertical direction. 2 means a predicted value. P f (x, y) means the final predicted value of the (x, y) coordinates according to the present embodiment. Equation 3 is an example, and a method of calculating the final prediction value using the first prediction value and the second prediction value is not limited thereto.
다른 일례로 제1 예측값과 제2 예측값의 가중 평균을 이용할 수 있다. 일반적으로, 다른 화소를 이용하여 예측된 인접화소 P(즉, 현재블록의 우측 및 하측 인접화소)보다 참조화소 R(즉, 현재블록의 좌측 및 상측 인접화소)이 현재화소(510)의 값과 더 유사할 가능성이 높다. 이러한 경향을 반영하여 수평으로 예측된 제1 예측값과 수직으로 예측된 제2 예측값의 위치를 기반으로 가중치를 달리하는 방법도 가능하다. 이를 수식으로 나타내면 수학식 4와 같다.As another example, a weighted average of the first prediction value and the second prediction value may be used. In general, the reference pixel R (that is, the left and upper neighboring pixels of the current block) is larger than the neighboring pixel P (that is, the right and lower neighboring pixels of the current block) predicted using another pixel. It is likely to be more similar. Reflecting this tendency, a method of varying weight based on the position of the first predicted value predicted horizontally and the second predicted value predicted vertically is also possible. This is expressed as an equation (4).
Figure PCTKR2018005596-appb-M000004
Figure PCTKR2018005596-appb-M000004
여기서, P1(x, y)는 수평방향으로 보간된 (x, y) 위치의 제1 예측값을 의미하고, P2(x, y)는 수직방향으로 보간된 (x, y) 위치의 제2 예측값을 의미한다. Pf(x, y)는 본 실시예에 따른 (x, y) 좌표의 최종 예측값을 의미한다. 수직방향의 거리 (y+1)과 수평방향의 거리 (x+1)이 가중치로 사용된다. 수직방향으로 y의 값이 커질수록 그리고 수평방향으로 x의 값이 커질수록 각 방향에서 최종 예측값이 참조화소 R로부터 멀어지고, 인접화소 P에 가까워지게 되므로, 두 가중치가 제1 예측값과 제2 예측값에 역으로 곱해진다. 즉, 제1 예측값에 적용되는 제1 가중치는 현재화소와 동일한 열에 위치한 상측 참조화소로부터 현재화소까지의 거리에 의해 결정되고, 제2 예측값에 적용되는 제2 가중치는 현재화소와 동일한 행에 위치한 좌측 참조화소로부터 현재화소까지의 거리에 의해 결정된다. 수학식 4에서 x와 y의 범위는 각각 (0 ~ W-1)과 (0 ~ H-1)이므로 (x, y)가 (0, 0)일 때 수학식 3과 동일한 계산을 수행하며, 그 외의 경우에는 제1 예측값과 제2 예측값 중에서 더 참조화소 R에 가까운 예측값에 큰 가중치를 부여하여 최종 예측값을 획득한다.Here, P 1 (x, y) means the first predicted value of the position (x, y) interpolated in the horizontal direction, and P 2 (x, y) means the first predicted value of the position (x, y) interpolated in the vertical direction. 2 means a predicted value. P f (x, y) means the final predicted value of the (x, y) coordinates according to the present embodiment. The vertical distance (y + 1) and the horizontal distance (x + 1) are used as weights. As the value of y in the vertical direction increases and the value of x in the horizontal direction, the final prediction value in each direction moves away from the reference pixel R and approaches the adjacent pixel P, so that the two weights are equal to the first prediction value and the second prediction value. Is multiplied inversely. That is, the first weight applied to the first prediction value is determined by the distance from the upper reference pixel located in the same column as the current pixel to the current pixel, and the second weight applied to the second prediction value is located on the left in the same row as the current pixel. It is determined by the distance from the reference pixel to the current pixel. In Equation 4, the ranges of x and y are (0 to W-1) and (0 to H-1), respectively, so that the same calculation as in Equation 3 is performed when (x, y) is (0, 0). In other cases, the final predicted value is obtained by giving a greater weight to the predicted value closer to the reference pixel R among the first predicted value and the second predicted value.
도 4에는 본 발명의 평면 예측 모드로 화면 내 예측을 수행하기 위한 장치(400)의 블록도가 도시되어 있다. 제1 인접화소 설정부(410)는 복호화부(310)에 의해 비트스트림으로부터 복호화된 현재블록의 제1 인접화소(BR)를 지시하는 정보를 이용하여 제1 인접화소(BR)의 값을 설정할 수 있다. 대안적으로, 제1 인접화소 설정부(410)는 현재블록의 참조화소들을 기반으로 제1 인접화소(BR)의 값을 설정할 수 있다. 인접화소 예측부(420)는 제1 인접화소(BR)의 값, 제2 인접화소(TR)의 값 및 제3 인접화소(BL)의 값을 이용한 보간에 의해 현재블록의 하측 인접화소의 값과 우측 인접화소의 값을 결정한다. 구체적으로, 인접화소 예측부(420)는 제1 인접화소(BR)의 값 및 제2 인접화소(TR)의 값을 이용한 보간에 의해 우측 인접화소의 값을 결정하고, 제1 인접화소(BR)의 값 및 제3 인접화소(BL)의 값을 이용한 보간에 의해 하측 인접화소의 값을 결정한다.4 is a block diagram of an apparatus 400 for performing intra prediction in a planar prediction mode of the present invention. The first neighboring pixel setting unit 410 sets the value of the first neighboring pixel BR by using the information indicating the first neighboring pixel BR of the current block decoded from the bitstream by the decoder 310. Can be. Alternatively, the first neighboring pixel setting unit 410 may set the value of the first neighboring pixel BR based on the reference pixels of the current block. The neighboring pixel prediction unit 420 may determine the value of the lower neighboring pixel of the current block by interpolation using the value of the first neighboring pixel BR, the value of the second neighboring pixel TR, and the value of the third neighboring pixel BL. And the value of the right adjacent pixel. In detail, the neighboring pixel predictor 420 determines the value of the right neighboring pixel by interpolation using the value of the first neighboring pixel BR and the value of the second neighboring pixel TR, and determines the first neighboring pixel BR. The value of the lower adjacent pixel is determined by interpolation using the value of) and the value of the third adjacent pixel BL.
현재화소 예측부(430)는 현재블록의 좌측 인접화소의 값, 상측 인접화소의 값, 하측 인접화소의 값 및 우측 인접화소의 값을 이용한 보간에 의해 현재블록 내 현재화소의 예측값을 결정한다. 구체적으로, 현재화소 예측부(430)는 좌측 인접화소의 값과 우측 인접화소의 값을 이용한 보간에 의해 현재화소의 제1 예측값을 결정하고, 상측 인접화소의 값과 하측 인접화소의 값을 이용한 보간에 의해 현재화소의 제2 예측값을 결정한다. 그리고 현재화소 예측부(430)는 현재화소의 제1 예측값과 제2 예측값을 이용하여 최종적인 예측값을 결정한다. 한편, 좌측 인접화소와 상측 인접화소는 현재블록의 참조화소로서 이미 예측과 복원에 의해 그 값이 재구성된 화소들이다.The current pixel prediction unit 430 determines the prediction value of the current pixel in the current block by interpolation using the value of the left neighboring pixel, the upper neighboring pixel, the lower neighboring pixel, and the right neighboring pixel of the current block. Specifically, the current pixel prediction unit 430 determines the first prediction value of the current pixel by interpolation using the value of the left neighboring pixel and the value of the right neighboring pixel, and uses the value of the upper neighboring pixel and the value of the lower neighboring pixel. The second prediction value of the current pixel is determined by interpolation. The current pixel prediction unit 430 determines a final prediction value by using the first prediction value and the second prediction value of the current pixel. On the other hand, the left neighboring pixels and the upper neighboring pixels are pixels that have already been reconstructed by prediction and reconstruction as reference pixels of the current block.
본 실시예의 평면 예측 모드의 방법에서 중요한 이슈는 현재블록의 우측 하단 모서리에 위치하는 제1 인접화소(BR)의 값을 어떻게 설정하느냐 하는 것이다. 제1 인접화소(BR)의 값은 복호화 장치에서 다른 인접화소 및/또는 주변화소들을 이용하여 부호화 장치와 동일한 연산식에 의해 계산되거나, 부호화 장치에서 결정된 제1 인접화소(BR)의 값에 대한 정보가 복호화 장치로 전송될 수 있다. 다만, 본 실시예는 전송되는 제1 인접화소(BR)에 대한 데이터보다 본 실시예의 평면 예측에 따라 감소된 잔차(residual) 데이터가 더 작을 경우에는 개선된 효과를 얻을 수 없다. 따라서, 제1 인접화소(BR)의 값을 효율적으로 전송하는 것이 매우 중요하다.An important issue in the method of planar prediction mode of this embodiment is how to set the value of the first adjacent pixel BR located at the lower right corner of the current block. The value of the first neighboring pixel BR is calculated by the same operation as the encoding apparatus using other neighboring pixels and / or neighboring pixels in the decoding apparatus, or for the value of the first neighboring pixel BR determined by the encoding apparatus. Information may be transmitted to the decoding device. However, the present embodiment may not obtain an improved effect when the residual data reduced according to the planar prediction of the present embodiment is smaller than the data of the first adjacent pixel BR transmitted. Therefore, it is very important to efficiently transmit the value of the first adjacent pixel BR.
이하, 다른 도면들을 참조하여 제1 인접화소(BR)의 값을 효율적으로 전송하거나 보간하기 위한 다양한 실시예들에 대해 설명한다. 우선, 도 6 내지 도 8을 참조하여 제1 인접화소(BR)의 값이 효율적으로 전송되는 경우에 대하여 설명한다. 도 6 내지 도 8은 본 발명의 실시예에 따라서 비트스트림으로부터 복호화된 제1 인접화소(BR)를 지시하는 정보를 이용하는 평면 예측 모드의 예시도이다. 여기서, 제1 인접화소(BR)를 지시하는 정보는 제1 인접화소(BR)의 값 및 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 화소의 위치를 나타내는 정보를 포함한다.Hereinafter, various embodiments for efficiently transmitting or interpolating a value of the first adjacent pixel BR will be described with reference to other drawings. First, a case in which the value of the first adjacent pixel BR is efficiently transmitted will be described with reference to FIGS. 6 to 8. 6 to 8 are exemplary diagrams of a planar prediction mode using information indicating a first neighboring pixel BR decoded from a bitstream according to an embodiment of the present invention. The information indicating the first adjacent pixel BR includes information indicating a position of a pixel having a value most similar to the original value of the first adjacent pixel BR and the original value of the first adjacent pixel BR.
도 6은 영상 부호화 장치가 제1 인접화소(BR)의 값에 대한 정보를 영상 복호화 장치로 직접 전송하는 일 실시예를 도시한다. 영상 복호화 장치는 평면 예측 모드가 선택된 경우, 제1 인접화소(BR)의 값을 지시하는 신택스 요소(예: br_value)를 비트스트림으로부터 복호화하여 해당 신택스 요소가 지시하는 값을 제1 인접화소(BR)의 값으로 설정할 수 있다.FIG. 6 illustrates an embodiment in which the image encoding apparatus directly transmits information about a value of the first adjacent pixel BR to the image decoding apparatus. When the planar prediction mode is selected, the image decoding apparatus decodes a syntax element (eg, br_value ) indicating a value of the first neighboring pixel BR from the bitstream and decodes a value indicated by the corresponding syntax element in the first neighboring pixel BR. It can be set to the value of).
도 7은 영상 부호화 장치가 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 화소의 위치에 대한 정보를 영상 복호화 장치로 전송하는 일 실시예를 도시한다. 여기서, 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 참조화소의 위치에 대한 정보는 1차원 인덱스로 표현될 수 있다. 영상 복호화 장치는 평면 예측 모드가 선택된 경우, 참조화소의 위치에 대한 정보를 지시하는 신택스 요소(예: br_idx)를 비트스트림으로부터 복호화하여 해당 신택스 요소가 지시하는 위치의 참조화소의 값을 제1 인접화소(BR)의 값으로 설정할 수 있다.FIG. 7 illustrates an embodiment in which the image encoding apparatus transmits information about a position of a pixel having a value most similar to the original value of the first adjacent pixel BR to the image decoding apparatus. Here, the information about the position of the reference pixel having the value most similar to the original value of the first adjacent pixel BR may be represented by a one-dimensional index. When the planar prediction mode is selected, the image decoding apparatus decodes a syntax element (for example, br_idx ) indicating information about a position of a reference pixel from the bitstream to determine a value of a reference pixel at a position indicated by the corresponding syntax element. It may be set to the value of the pixel BR.
br_idx 신택스 요소는 현재블록의 참조화소들 중 어느 하나의 화소의 위치를 (0, 0)으로 하여 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 화소의 위치를 1차원 오프셋으로 지시할 수 있다. 예를 들어, 현재블록의 좌측 상단 모서리에 위치하는 화소(TL), 제2 인접화소(TR) 및 제3 인접화소(BL) 중 어느 하나의 위치를 (0, 0)으로 하여 1차원 오프셋이 결정될 수 있다. 도 7은 화소 TL의 위치가 (0, 0)으로 설정된 경우를 나타낸다. 추가적으로, 영상 부호화 장치는 참조 방향을 지시하는 신택스 요소(예: br_direction)를 전송하여 영상 복호화 장치로 하여금 해당 신택스 요소를 복호화하여 참조 방향을 수평 또는 수직으로 결정하게 할 수 있다. The br_idx syntax element indicates the position of a pixel having a value most similar to the original value of the first neighboring pixel BR as a one-dimensional offset by setting the position of one pixel among the reference pixels of the current block as (0, 0). can do. For example, the one-dimensional offset is set by setting any one of the pixel TL, the second adjacent pixel TR, and the third adjacent pixel BL positioned at the upper left corner of the current block as (0, 0). Can be determined. 7 illustrates a case where the position of the pixel TL is set to (0, 0). In addition, the apparatus for encoding an image may transmit a syntax element (eg, br_direction ) indicating a reference direction so that the apparatus for decoding an image may determine the reference direction horizontally or vertically by decoding the syntax element.
도 8은 영상 부호화 장치가 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 화소의 위치에 대한 정보를 영상 복호화 장치로 전송하는 다른 실시예를 도시한다. 여기서, 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 주변화소 및/또는 인접화소의 위치에 대한 정보는 2차원 벡터로 표현될 수 있다. 영상 복호화 장치는 평면 예측 모드가 선택된 경우, 유사한 값을 갖는 화소의 위치에 대한 정보를 지시하는 신택스 요소(예: u_br, v_br)를 비트스트림으로부터 복호화하여 해당 신택스 요소가 지시하는 위치의 주변화소의 값을 제1 인접화소(BR)의 값으로 설정할 수 있다.FIG. 8 illustrates another embodiment in which the image encoding apparatus transmits information about a position of a pixel having a value most similar to the original value of the first adjacent pixel BR to the image decoding apparatus. Here, the information about the position of the neighboring pixel and / or the adjacent pixel having the value most similar to the original value of the first neighboring pixel BR may be expressed as a two-dimensional vector. The image decoding device further comprises: a flat prediction mode is selected if, syntax elements indicating the information on the location of the pixels having similar values: decoding (for example, u_br, v_br) from the bit stream of the peripheral pixels of the location indicated by the corresponding syntax element The value may be set to the value of the first adjacent pixel BR.
u_br 및 v_br 신택스 요소는 현재블록의 참조화소들 중 어느 하나의 화소의 위치를 (0, 0)으로 하여 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 화소의 위치를 2차원 오프셋으로 지시할 수 있다. 예를 들어, 현재블록의 좌측 상단 모서리에 위치하는 화소(TL), 제2 인접화소(TR) 및 제3 인접화소(BL) 중 어느 하나의 위치를 (0, 0)으로 하여 2차원 오프셋이 결정될 수 있다. 도 8은 화소 TL의 위치가 (0, 0)으로 설정된 경우를 나타낸다. The u_br and v_br syntax elements are two-dimensional offsets of the position of the pixel having the value most similar to the original value of the first adjacent pixel BR, with the position of one pixel among the reference pixels of the current block as (0, 0). Can be indicated. For example, a two-dimensional offset is set by setting any one of the pixel TL, the second adjacent pixel TR, and the third adjacent pixel BL positioned at the upper left corner of the current block as (0, 0). Can be determined. 8 illustrates a case where the position of the pixel TL is set to (0, 0).
이하, 도 9 내지 도 11을 참조하여 제1 인접화소(BR)의 값이 다른 주변화소 및/또는 인접화소들을 이용한 연산식에 의해 계산되는 경우에 대하여 설명한다. 9 내지 도 11은 본 발명의 다른 실시예에 따라서 현재블록의 주변화소 및/또는 인접화소들을 기초로 계산된 값을 이용하는 평면 예측 모드의 예시도이다.Hereinafter, a case in which the value of the first neighboring pixel BR is calculated by a calculation formula using other neighboring pixels and / or neighboring pixels will be described with reference to FIGS. 9 through 11. 9 to 11 are exemplary views of planar prediction mode using a value calculated based on neighboring pixels and / or neighboring pixels of a current block according to another embodiment of the present invention.
도 9는 영상 복호화 장치가 제1 인접화소(BR)의 값에 대한 정보를 영상 부호화 장치로부터 별도로 수신하지 않고 복원된 인접화소 및 주변화소들의 값의 특정방향으로의 변화량을 기반으로 제1 인접화소(BR)의 값을 보간하는 방법의 일 실시예를 나타낸다. 영상 복호화 장치는 제1 인접화소(BR)로부터 연장된 적어도 하나의 특정 방향의 직선 상에 위치하는 현재블록의 참조화소(예: TR, TL, BL) 또는 인접화소의 값에 이와 동일한 직선 상에 위치하는 이미 복원된 주변화소의 값과의 변화량을 합산하여 해당 방향에서의 제1 인접화소(BR) 예측값을 계산한다. 그리고 각 방향으로부터 예측된 제1 인접화소(BR) 예측값들의 평균으로 제1 인접화소(BR)의 값을 설정한다. 여기서, 특정 방향은 제1 인접화소(BR)로부터 수평 방향, 수직 방향 및 대각선 방향 중 적어도 하나일 수 있고, 참조화소는 현재블록의 좌측 상단 모서리에 위치한 인접화소(TL), 제2 인접화소(TR) 및 제3 인접화소(BL) 중 적어도 하나일 수 있다.FIG. 9 illustrates a first neighboring pixel based on an amount of change of values of neighboring pixels and neighboring pixels reconstructed without separately receiving information on a value of a first neighboring pixel from the image encoding apparatus. An embodiment of a method of interpolating a value of (BR) is shown. The apparatus for decoding an image may be arranged on the same straight line as a reference pixel (eg, TR, TL, BL) or a value of a neighboring pixel of the current block located on a straight line in at least one specific direction extending from the first adjacent pixel BR. The first neighboring pixel (BR) prediction value in the corresponding direction is calculated by summing change amounts with the values of neighboring pixels which are already restored. A value of the first adjacent pixel BR is set as an average of the first neighboring pixel BR predicted values from each direction. Here, the specific direction may be at least one of a horizontal direction, a vertical direction, and a diagonal direction from the first adjacent pixel BR, and the reference pixel may include the adjacent pixel TL and the second adjacent pixel located at the upper left corner of the current block. TR) and at least one of the third adjacent pixels BL.
도 9는 현재블록의 모서리에 위치한 3개의 참조화소(TR, TL, BL)의 각 방향으로의 변화량을 기반으로 제1 인접화소(BR)의 값을 보간하는 경우를 예시한다. 제1 인접화소(BR)의 수직 방향 직선에 위치하는 제2 인접화소(TR)의 값과 이미 복원된 주변화소(910)의 값 간의 변화량을 제2 인접화소(TR)의 값에 합산하여 제1 인접화소(BR)의 수직방향 예측값을 계산한다. 제1 인접화소(BR)의 대각선 방향 직선에 위치하는 인접화소(TL)의 값과 이미 복원된 주변화소(920)의 값 간의 변화량을 주변화소(TL)의 값에 합산하여 제1 인접화소(BR)의 대각선방향 예측값을 계산한다. 또한, 제1 인접화소(BR)의 수평 방향 직선에 위치하는 제3 인접화소(BL)의 값과 이미 복원된 주변화소(930)의 값 간의 변화량을 제3 인접화소(BL)의 값에 합산하여 제1 인접화소(BR)의 수평방향 예측값을 계산한다. 각 인접화소들(TR, TL, BL)로부터 계산된 제2 인접화소(BR)의 예측값들의 평균으로 제1 인접화소(BR)의 값을 생성한다.FIG. 9 illustrates a case of interpolating values of the first neighboring pixel BR based on the amount of change in each direction of three reference pixels TR, TL, and BL located at the edge of the current block. The amount of change between the value of the second adjacent pixel TR located on a straight line in the vertical direction of the first adjacent pixel BR and the value of the neighboring pixel 910 which has already been restored is added to the value of the second adjacent pixel TR to add the first change. 1 The vertical prediction value of the adjacent pixel BR is calculated. The amount of change between the value of the adjacent pixel TL positioned on the diagonal line of the first adjacent pixel BR and the value of the neighboring pixel 920 already restored is added to the value of the peripheral pixel TL to add the first adjacent pixel ( Calculate the diagonal prediction value of BR). In addition, the amount of change between the value of the third adjacent pixel BL positioned on the horizontal straight line of the first adjacent pixel BR and the value of the neighboring pixel 930 that is already restored is added to the value of the third adjacent pixel BL. The horizontal prediction value of the first adjacent pixel BR is calculated. The value of the first adjacent pixel BR is generated as an average of the predicted values of the second adjacent pixel BR calculated from the adjacent pixels TR, TL, and BL.
본 실시예에서 제안하는 제1 인접화소(BR)의 보간 방법에 따른 연산식은 수학식 5와 같다. Equation 5 according to the interpolation method of the first adjacent pixel BR proposed in the present embodiment is shown in Equation 5 below.
Figure PCTKR2018005596-appb-M000005
Figure PCTKR2018005596-appb-M000005
여기서, H는 현재블록의 높이를 의미하며, W는 현재블록의 너비를 의미한다. P(W, H)는 제1 인접화소(BR)의 값을 의미하고, R(-1, -1)은 현재블록의 좌측 상단 모서리에 위치한 인접화소(TL)의 값을 의미하며, R(W, -1)은 제2 인접화소(TR)의 값을 의미하고, R(-1, H)는 제3 인접화소(BL)의 값을 의미한다. R(-1-W, -1-H), R(W, -1-H) 및 R(-1-W, H)는 각 방향으로 현재블록의 크기만큼 떨어진 이미 복원된 주변화소의 값을 의미한다.Here, H means the height of the current block, W means the width of the current block. P (W, H) means the value of the first adjacent pixel BR, R (-1, -1) means the value of the adjacent pixel TL located at the upper left corner of the current block, and R ( W and -1 mean a value of the second adjacent pixel TR, and R (-1, H) means a value of the third adjacent pixel BL. R (-1-W, -1-H), R (W, -1-H), and R (-1-W, H) are the values of the neighboring pixels that have already been restored, separated by the size of the current block in each direction. it means.
도 10은 도 9의 평면 예측 모드에 따라 제1 인접화소(BR)의 보간 값을 계산한 일례를 나타낸다. 제2 인접화소(TR)의 값은 50, 제2 인접화소(TR)로부터 수직방향으로 떨어진 위치의 복원 주변화소(1010)의 값은 60, 좌측 상단 모서리 화소(TL)의 값은 71, 좌측 상단 모서리 화소(TL)로부터 대각선 방향으로 떨어진 위치의 복원 주변화소(1020)의 값은 85, 제3 인접화소(BL)의 값은 50, 제3 인접화소(BL)로부터 수평방향으로 떨어진 위치의 복원 주변화소(1030)의 값은 60이라고 가정한다. 각 방향으로의 화소값의 변화량이 반영된 제1 인접화소(BR)의 보간값은 46으로 계산된다. 본 실시예에서는 블록 간의 간격을 기반으로 변화량을 계산했지만 변화량을 계산하는 방법은 이에 한정되지 않으며, 화소 간의 간격 등 다양한 방법이 사용될 수 있다.FIG. 10 illustrates an example of calculating an interpolation value of the first adjacent pixel BR according to the planar prediction mode of FIG. 9. The value of the second adjacent pixel TR is 50, the value of the reconstructed peripheral pixel 1010 at a position vertically away from the second adjacent pixel TR is 60, the value of the upper left corner pixel TL is 71, the left side. The value of the reconstructed peripheral pixel 1020 at a position diagonally away from the upper edge pixel TL is 85, the value of the third adjacent pixel BL is 50, and the value of the position is horizontally separated from the third adjacent pixel BL. It is assumed that the value of the restored peripheral pixel 1030 is 60. The interpolation value of the first adjacent pixel BR reflecting the amount of change in the pixel value in each direction is calculated as 46. In the present embodiment, the amount of change is calculated based on the distance between blocks, but the method of calculating the amount of change is not limited thereto, and various methods such as the distance between pixels may be used.
다른 실시예로, 보간 과정에서 사용되는 각 화소 값에 서로 다른 가중치를 부가할 수도 있다. 예컨대, 각 변화량이 보정된 참조화소들(TL, TR, BL)의 값에 가중치를 부가하는 경우의 연산식은 수학식 6 같다.In another embodiment, different weights may be added to each pixel value used in the interpolation process. For example, the equation in the case where the weight is added to the values of the reference pixels TL, TR, and BL in which each change amount is corrected is expressed by Equation 6.
Figure PCTKR2018005596-appb-M000006
Figure PCTKR2018005596-appb-M000006
여기서, cd, cw, ch는 가중치를 의미하며 C= cd + cw + ch이다. 각 가중치는 영상의 특성에 따라 영상 부호화 장치 및 복호화 장치가 동일한 방법으로 결정하며, 특정 방법으로 한정되지 않는다.Here, c d , c w , and c h denote weights and C = c d + c w + c h . Each weight is determined in the same manner by the image encoding apparatus and the decoding apparatus according to the characteristics of the image, and is not limited to a specific method.
도 11은 영상 복호화 장치가 제1 인접화소(BR)의 값에 대한 정보를 영상 부호화 장치로부터 별도로 수신하지 않고 현재블록의 참조화소들의 평균값 또는 가중 평균값 기반으로 제1 인접화소(BR)의 값을 설정하는 방법의 일 실시예를 나타낸다. 본 실시예에 따라 계산되는 제1 인접화소(BR) 값의 예시들은 다음과 같다.11 shows a value of a first neighboring pixel BR based on an average value or a weighted average value of reference pixels of a current block without separately receiving information about the value of the first neighboring pixel BR from the image encoding apparatus. An embodiment of a setting method is shown. Examples of the first adjacent pixel value BR calculated according to the present embodiment are as follows.
1) TR 및 BL 값들의 평균값1) average of TR and BL values
2) TR 및 BL 값들의 가중 평균값2) Weighted average of TR and BL values
3) TR, BL 및 TL 값들의 가중 평균값3) Weighted average of TR, BL and TL values
4) TR, TR+1, BL, BL+1 값들의 평균값4) Average of TR, TR + 1, BL, BL + 1 values
5) TR, TR-1, TR+1, BL, BL-1, BL+1 값들의 평균값5) Average of TR, TR-1, TR + 1, BL, BL-1, BL + 1 values
6) TR, TR+1, TR+2, BL, BL+1, BL+2 값들의 평균값6) Average of TR, TR + 1, TR + 2, BL, BL + 1, BL + 2 values
위의 예시들에서, TR은 현재블록의 우측 상단 모서리에 위치하는 인접화소(즉, 제2 인접화소), BL은 좌측 하단 모서리에 위치하는 인접화소(즉, 제3 인접화소), TL은 좌측 상단 모서리에 위치하는 인접화소를 의미한다. TR±x는 TR을 기준으로 좌우 방향에 위치하는 화소를 의미하고, BL±y는 BL을 기준으로 상하 방향에 위치하는 화소를 의미한다. 예컨대, TR+1은 TR의 우측에 접하는 화소를 의미하고 TR+2는 TR+1의 우측에 접하는 화소를 의미한다.In the above examples, TR is the adjacent pixel located at the upper right corner of the current block (ie, the second adjacent pixel), BL is the adjacent pixel located at the lower left corner (ie, the third adjacent pixel), and TL is left It means the adjacent pixel located in the upper corner. TR ± x means pixels located in the left and right directions with respect to TR, and BL ± y means pixels located in the up and down direction with respect to BL. For example, TR + 1 means a pixel in contact with the right side of TR and TR + 2 means a pixel in contact with the right side of TR + 1.
제1 인접화소(BR)의 값을 계산하는 방법들은 위의 5개의 예시들에 한정되지 않고 현재블록의 참조화소들을 이용하여 계산하는 여러 가지 방법들을 포함할 수 있다. 제1 인접화소(BR)의 값은 영상 부호화 장치 및 복호화 장치에서 동일한 방법으로 계산된다.Methods of calculating a value of the first neighboring pixel BR are not limited to the above five examples, and may include various methods of calculating using reference pixels of the current block. The value of the first adjacent pixel BR is calculated by the same method in the image encoding apparatus and the decoding apparatus.
일례로, 영상 부호화 장치 및 복호화 장치는 수학식 7과 같은 연산식을 이용하여 제1 인접화소(BR)의 값을 동일하게 설정할 수 있다.For example, the image encoding apparatus and the decoding apparatus may set the same value of the first neighboring pixel BR by using an equation such as Equation (7).
Figure PCTKR2018005596-appb-M000007
Figure PCTKR2018005596-appb-M000007
여기서, H는 현재블록의 높이를 의미하며, W는 현재블록의 너비를 의미한다. P(W, H)는 제1 인접화소(BR)의 값을 의미하고, R(W, -1)은 제2 인접화소(TR)의 값을 의미하며, R(-1, H)는 제3 인접화소(BL)의 값을 의미한다. R(W-1, -1), R(W-2, -1), R(W+1, -1) 및 R(W+2, -1)은 제2 인접화소(TR)의 인접화소들을 의미하며, R(-1, H-1), R(-1, H-2), R(-1, H+1) 및 R(-1, H+2)는 제3 인접화소(BL)의 인접화소들을 의미한다.Here, H means the height of the current block, W means the width of the current block. P (W, H) means the value of the first adjacent pixel BR, R (W, -1) means the value of the second adjacent pixel TR, and R (-1, H) is the zero It means the value of 3 adjacent pixels BL. R (W-1, -1), R (W-2, -1), R (W + 1, -1) and R (W + 2, -1) are adjacent pixels of the second adjacent pixel TR R (-1, H-1), R (-1, H-2), R (-1, H + 1) and R (-1, H + 2) are the third adjacent pixels BL ) Means adjacent pixels.
또 다른 일례로, 영상 부호화 장치 및 복호화 장치는 수학식 8과 같은 연산식을 이용하여 제1 인접화소(BR)의 값을 동일하게 설정할 수 있다.As another example, the image encoding apparatus and the decoding apparatus may set the same value of the first neighboring pixel BR by using an equation such as Equation (8).
Figure PCTKR2018005596-appb-M000008
Figure PCTKR2018005596-appb-M000008
여기서, H는 현재블록의 높이를 의미하며, W는 현재블록의 너비를 의미한다. P(W, H)는 제1 인접화소(BR)의 값을 의미하고, R(W, -1)은 제2 인접화소(TR)의 값을 의미하며, R(-1, H)는 제3 인접화소(BL)의 값을 의미한다. cw, ch는 가중치를 의미하며 C= cw + ch이다. 각 가중치는 영상의 특성에 따라 영상 부호화 장치 및 복호화 장치가 동일한 방법으로 결정된다. 예를 들어, cw 는 현재블록의 너비를, ch 는 현재블록의 높이를 가중치로 사용할 수 있다.Here, H means the height of the current block, W means the width of the current block. P (W, H) means the value of the first adjacent pixel BR, R (W, -1) means the value of the second adjacent pixel TR, and R (-1, H) is the zero It means the value of 3 adjacent pixels BL. c w , c h means the weight and C = c w + c h . Each weight is determined by the same method as the image encoding apparatus and the decoding apparatus according to the characteristics of the image. For example, c w may use the width of the current block and c h may use the height of the current block as a weight.
다른 실시예로, 영상 부호화 장치는 미리 정의된 복수의 제1 인접화소(BR)의 후보값들 중에서 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 후보값을 선택하여 선택된 후보값을 지시하는 정보를 영상 복호화 장치로 전송할 수 있다. 영상 복호화 장치는 비트스트림으로부터 복수의 제1 인접화소(BR)의 후보값들 중 하나를 지시하는 정보를 복호화하여 복호화된 정보가 지시하는 제1 인접화소(BR)의 후보값을 제1 인접화소(BR)의 값으로 설정할 수 있다. 여기서, 제1 인접화소(BR)의 후보값을 지시하는 정보로 br_idx 신택스 요소가 사용할 수 있다. 표 1은 br_idx 및 제1 인접화소(BR)의 후보값의 일례를 나타낸다.In another embodiment, the apparatus for encoding an image selects a candidate value having a value most similar to the original value of the first neighboring pixel BR among candidate values of the plurality of predefined first neighboring pixels BR, and is selected. Information indicating a may be transmitted to the image decoding apparatus. The image decoding apparatus decodes information indicating one of the candidate values of the plurality of first adjacent pixels BR from the bitstream, and determines the candidate value of the first adjacent pixel BR indicated by the decoded information. Can be set to a value of (BR). Here, the br_idx syntax element may be used as information indicating a candidate value of the first adjacent pixel BR. Table 1 shows an example of candidate values of br_idx and the first adjacent pixel BR.
br_idxbr_idx Associated candidateAssociated candidate
00 TR 및 BL의 평균값Average value of TR and BL
1One TR, TR+1, BL, BL+1 값들의 평균값Average of TR, TR + 1, BL, BL + 1 values
22 TR, TR-1, TR+1, BL, BL-1, BL+1 값들의 평균값Average of TR, TR-1, TR + 1, BL, BL-1, BL + 1 values
33 TR, TR+1, TR+2, BL, BL+1, BL+2 값들의 평균값Average value of TR, TR + 1, TR + 2, BL, BL + 1, BL + 2 values
영상 부호화 장치가 선택된 하나의 br_idx 값을 영상 복호화 장치로 전송하면, 영상 복호화 장치는 전송된 br_idx에 해당하는 후보값을 구하여 제1 인접화소(BR)의 값으로 설정한다. 예를 들어, 영상 복호화 장치가 비트스트림으로부터 복호화한 br_idx의 값이 1인 경우, TR, TR+1, BL 및 BL+1의 값들의 평균값을 구하여 제1 인접화소(BR)의 값으로 설정할 수 있다.When the image encoding apparatus transmits the selected br_idx value to the image decoding apparatus, the image decoding apparatus obtains a candidate value corresponding to the transmitted br_idx and sets it to the value of the first neighboring pixel BR. For example, when the value of br_idx decoded from the bitstream by the image decoding apparatus is 1, the average value of the values of TR, TR + 1, BL, and BL + 1 may be obtained and set as the value of the first adjacent pixel BR. have.
또 다른 실시예로, 영상 부호화 장치는 제1 인접화소(BR)의 원본값과 참조화소들(예: TL, TR, BL 등)의 값을 기반으로 보간한 제1 인접화소(BR)의 예측값의 차분값을 전송할 수 있다. 예를 들어, 차분값은 br_residual 신택스 요소를 통해 전송될 수 있다. 영상 복호화 장치는 차분값 외에 차분값에 대한 부호 정보를 추가 전송할 수 있다. 영상 복호화 장치는 비트스트림으로부터 제1 인접화소(BR)의 원본값과 제1 인접화소(BR)의 예측값의 차분값을 나타내는 정보(br_residual 신택스 요소)를 복호화하고, 영상 부호화 장치와 동일한 방법으로 획득한 제1 인접화소(BR)의 예측값에 복호화된 차분값을 보상한 값을 제1 인접화소(BR)의 값으로 설정할 수 있다. 이때, 영상 복호화 장치는 차분값에 대한 부호 정보를 별도로 복호화하고, 복호화된 부호 정보를 보상한 값을 제1 인접화소(BR)의 값으로 설정할 수 있다. 본 실시예에서 제1 인접화소(BR)의 예측값은 도 9 내지 도 11을 참조하여 설명한 방법 등에 의해 획득될 수 있으나, 반드시 이에 한정되는 것은 아니다.In another embodiment, the apparatus for encoding an image may predict the first neighboring pixel BR based on the original value of the first neighboring pixel BR and values of reference pixels (eg, TL, TR, BL, etc.). The difference value of can be transmitted. For example, the difference value can be transmitted via the br_residual syntax element. The image decoding apparatus may additionally transmit code information about the difference value in addition to the difference value. The image decoding apparatus decodes information ( br_residual syntax element) representing a difference value between the original value of the first neighboring pixel BR and the predicted value of the first neighboring pixel BR from the bitstream, and obtains the same method as the image encoding apparatus. A value that compensates for the decoded difference value in the predicted value of one first adjacent pixel BR may be set as the value of the first adjacent pixel BR. In this case, the image decoding apparatus may separately decode code information about the difference value, and set a value compensated for the decoded code information as a value of the first adjacent pixel BR. In the present exemplary embodiment, the predicted value of the first adjacent pixel BR may be obtained by the method described with reference to FIGS. 9 to 11, but is not necessarily limited thereto.
예를 들어, 영상 부호화 장치는 복원된 제2 인접화소(TR)의 값이 55이고, 복원된 제3 인접화소(BL)의 값이 45이며, 제1 인접화소(BR)의 원본값이 60인 경우에, 제2 인접화소(TR) 및 제3 인접화소(BL) 값들의 평균값(50)을 제1 인접화소(BR)의 예측값으로서 생성한 후, 제1 인접화소(BR)의 원본값(60)과 예측값(50)의 차분값(10)을 전송한다. 영상 부호화 장치는 추가적으로 차분값(10)에 대한 부호 정보(+/-)를 전송한다. 영상 복호화 장치는 이와 동일한 방법으로 제1 인접화소(BR)의 예측값을 생성한 후, 전송된 차분값(10)을 더하고, 전송된 부호(+/-)를 보상하여 제1 인접화소(BR)의 값(60)을 생성할 수 있다. 대안적으로, 제1 인접화소(BR)의 예측값은 계산된 값이 아닌 특정 위치의 복원된 화소값으로 설정될 수 있다.For example, in the image encoding apparatus, the value of the reconstructed second neighboring pixel TR is 55, the value of the reconstructed third neighboring pixel BL is 45, and the original value of the first neighboring pixel BR is 60. In the case of, the average value 50 of the values of the second adjacent pixel TR and the third adjacent pixel BL is generated as the predicted value of the first adjacent pixel BR, and then the original value of the first adjacent pixel BR is generated. The difference value 10 between 60 and the predicted value 50 is transmitted. The image encoding apparatus additionally transmits code information (+/−) of the difference value 10. The image decoding apparatus generates the predicted value of the first neighboring pixel BR in the same manner, adds the transmitted difference value 10, compensates the transmitted sign (+/−), and compensates the first neighboring pixel BR. Can produce a value of 60. Alternatively, the predicted value of the first adjacent pixel BR may be set to a reconstructed pixel value of a specific position instead of the calculated value.
도 12 내지 도 15는 본 발명의 다른 실시예에 따른 360 영상에서의 평면 예측 모드의 예시도이다. 이하, 360 영상과 같은 특수한 포맷의 영상을 위한 제1 인접화소(BR)의 값을 설정하는 방법에 대하여 설명한다. 본 실시예에서 현재블록은 2D 이미지로 부호화된 360 영상에서 복호화하고자 하는 제1 면에 위치한다.12 to 15 are exemplary diagrams of a planar prediction mode in a 360 image according to another embodiment of the present invention. Hereinafter, a method of setting a value of the first adjacent pixel BR for an image of a special format such as a 360 image will be described. In the present embodiment, the current block is located on the first surface to be decoded from the 360 image encoded into the 2D image.
도 12 (a)는 360 영상을 정방형 포맷(equirectangular format)으로 투영한 결과이다. 정방형 투영 포맷의 특징 중 하나는 이미지의 좌측 경계와 우측 경계가 서로 맞닿는 것이다. 이러한 특징을 고려하면, 도 12 (b)와 같이 이미지 내 좌측 경계에 접하는 복원 화소들(1210)의 값을 이미지의 우측 경계에 접하는 인접화소들(1220)의 값들로서 활용하여 이미지의 우측 경계에 접하는 블록에 대한 제1 인접화소(BR)의 값을 결정할 수 있다.12 (a) shows the result of projecting a 360 image in an equirectangular format. One of the features of the square projection format is that the left and right boundaries of an image abut each other. In consideration of this feature, as shown in FIG. 12B, the values of the reconstructed pixels 1210 adjacent to the left boundary of the image are used as the values of the adjacent pixels 1220 adjacent to the right boundary of the image. The value of the first adjacent pixel BR for the adjacent block may be determined.
즉, 영상 복호화 장치는 현재블록의 우측 경계가 360 영상에서 복호화하고자 하는 제1 면의 우측 경계와 일치할 때, 360 영상을 기준으로 현재블록의 우측 경계에 접하며 이미 복호화된 블록을 식별한다. 그리고 영상 복호화 장치는 식별된 블록의 좌측 경계에 접하는 블록 내 화소들(1210) 중 360 영상을 기준으로 제1 인접화소(BR)와 접하는 화소의 값을 제1 인접화소(BR)의 값으로 설정한다. 이에 더하여, 영상 복호화 장치는 식별된 블록의 좌측 경계에 접하는 블록 내 화소들(1210)의 적어도 일부를 현재블록의 참조화소들(1220)의 일부로 설정할 수 있다.That is, when the right boundary of the current block coincides with the right boundary of the first surface to be decoded in the 360 image, the image decoding apparatus contacts the right boundary of the current block based on the 360 image and identifies the already decoded block. The image decoding apparatus sets a value of a pixel in contact with the first adjacent pixel BR based on a 360 image among the pixels 1210 in the block adjacent to the left boundary of the identified block to the value of the first adjacent pixel BR. do. In addition, the image decoding apparatus may set at least some of the pixels 1210 in the block adjacent to the left boundary of the identified block as part of the reference pixels 1220 of the current block.
도 13은 현재블록의 우측 경계가 360 영상이 정방형 포맷으로 투영된 에서 2D 이미지의 우측 경계와 일치하는 경우를 예시한다. 현재블록의 제1 인접화소(BR, 1300)의 값은 360 영상을 기준으로 현재블록의 우측 경계에 접하며 이미 복호화된 화소들(1210) 중 제1 인접화소(BR. 1300)과 접하는 화소의 값으로 설정된다. 한편, 현재블록의 우측 인접화소들(1220)의 값 역시 정방형 포맷의 특성을 이용하여 360 영상을 기준으로 현재블록의 우측 경계에 접하며 이미 복호화된 블록의 좌측 경계의 인접화소들(1210)의 값으로 설정될 수 있다. 이 경우, 제1 인접화소(BR)와 제2 인접화소(TR)를 이용하여 보간을 할 필요가 없다.FIG. 13 illustrates a case where a right boundary of the current block coincides with a right boundary of a 2D image in which a 360 image is projected in a square format. The value of the first adjacent pixels BR and 1300 of the current block is in contact with the right boundary of the current block with respect to the 360 image and the value of the pixel in contact with the first adjacent pixels BR. 1300 among the decoded pixels 1210. Is set. On the other hand, the values of the adjacent pixels 1220 on the right side of the current block are also adjacent to the right boundary of the current block on the basis of 360 images using the characteristics of the square format, and the values of the adjacent pixels 1210 on the left boundary of the already decoded block. Can be set. In this case, it is not necessary to interpolate using the first adjacent pixel BR and the second adjacent pixel TR.
이하, 도 14 및 도 15를 참조하여 360 영상이 큐브 맵 포맷(cube map format)으로 투영된 경우에 대하여 설명한다. 도 14를 참조하면, (a)는 레이아웃(layout)의 변경이 없는 non-compact 레이아웃의 2D 이미지를, (b)는 각 면(face)을 재배치하어 공백이 없는 직사각형 형태로 레이아웃을 변경한 compact 레이아웃을 도시한다. Top, Bottom, Front, Back, Right, Left는 360 영상이 큐브에 투영될 때 각 투영 면(face)을 나타낸다. 각 면의 경계들은 360 영상을 기준으로 서로 인접 관계가 존재한다. 도 14에서는 360 영상에서의 각 인접관계를 도형으로 표현하였다. 예를 들어, non-compact 레이아웃에서 Front 면의 좌측 경계는 360 영상에서 Left 면의 우측 경계와 접한다.Hereinafter, a case in which a 360 image is projected in a cube map format will be described with reference to FIGS. 14 and 15. Referring to FIG. 14, (a) shows a 2D image of a non-compact layout without changing the layout, and (b) shows a compact layout in which a layout is changed to a rectangular shape without spaces by rearranging faces. The layout is shown. Top, Bottom, Front, Back, Right, Left represent each projection face when 360 images are projected onto the cube. The boundaries of each plane are adjacent to each other based on the 360 image. In FIG. 14, each adjacent relationship in the 360 image is represented by a figure. For example, in a non-compact layout, the left border of the front plane is in contact with the right border of the left plane in a 360 image.
큐브 맵 포맷의 경우에도 현재블록이 어느 한 면(face)의 경계에 접하는 경우, 이미 복호화된 다른 면의 경계에 위치한 화소들의 값을 참조하여 제1 인접화소(BR), 우측 열의 인접화소들, 및/또는 하측 행의 인접화소들의 값을 설정할 수 있다. 도 15를 참조하면, 현재블록이 제1 면(1510)에 위치하고 현재블록의 우측 경계가 제1 면(1510)의 우측 경계와 일치할 때, 영상 복호화 장치는 360 영상을 기준으로 제1 면(1510)의 우측 경계와 접하며 이미 복호화된 제2 면(1520)을 식별한다. 영상 복호화 장치는 제2 면(1520)의 좌측 경계에 접하는 제2 면(1520) 내 화소들 중 360 영상에서 제1 인접화소(BR)와 접하는 화소의 값을 제1 인접화소(BR)의 값으로 설정한다. 또한, 영상 복호화 장치는 제2 면(1520)의 좌측 경계에 접하는 제 2면(1520) 내 화소들(1522)의 적어도 일부를 현재블록의 우측 인접화소들(1512)의 적어도 일부로 설정할 수 있다.In the case of the cube map format, when the current block is in contact with the boundary of one face, the first adjacent pixel BR, adjacent pixels in the right column, And / or set the values of adjacent pixels in the lower row. Referring to FIG. 15, when the current block is located on the first surface 1510 and the right boundary of the current block coincides with the right boundary of the first surface 1510, the image decoding apparatus may determine the first surface (based on 360 images). A second face 1520 that is in contact with the right boundary of 1510 and has already been decrypted is identified. The image decoding apparatus may determine a value of a pixel in contact with the first adjacent pixel BR in the 360 image among the pixels in the second surface 1520 that is in contact with the left boundary of the second surface 1520. Set to. Also, the image decoding apparatus may set at least some of the pixels 1522 in the second surface 1520 adjacent to the left boundary of the second surface 1520 as at least some of the right adjacent pixels 1512 of the current block.
도 16은 본 발명의 다른 실시예에 따라서 현재블록에 인접한 주변화소들의 값이 설정되는 방법을 나타내는 예시도이다. 제1 인접화소(BR)의 값을 설정하여 제2 인접화소(TR) 및 제3 인접화소(BL)와의 보간을 이용하는 방법 외에도, 도 16과 같이 현재블록에 인접한 우측 인접화소들의 값 P(W, y)들 및 현재블록에 인접한 하측 인접화소들의 값 P(x, H)들을 설정할 수 있다. P(W, y)들은 각각 현재블록의 참조화소들 중 제2 인접화소(TR)을 포함한 우상측에 위치하는 복원된 각각의 화소들의 값으로 설정될 수 있다. P(x, H)들은 각각 현재블록의 참조화소들 중 제3 인접화소(BL)을 포함한 좌하측에 위치하는 복원된 각각의 화소들의 값으로 설정될 수 있다.16 is a diagram illustrating a method of setting values of neighboring pixels adjacent to a current block according to another embodiment of the present invention. In addition to the method of setting the value of the first adjacent pixel BR to use interpolation with the second adjacent pixel TR and the third adjacent pixel BL, as shown in FIG. 16, the value P (W) of the right adjacent pixels adjacent to the current block is shown. , y) and the values P (x, H) of lower neighboring pixels adjacent to the current block can be set. P (W, y) may be set to the value of each of the reconstructed pixels positioned on the upper right side including the second adjacent pixel TR among the reference pixels of the current block. P (x, H) may be set to the value of each of the reconstructed pixels located at the lower left side including the third adjacent pixel BL among the reference pixels of the current block.
이하, 도 17 및 도 18을 참조하여 본 발명의 실시예에 따른 화면 내 예측 복호화 방법에 대하여 설명한다. 도 17은 본 발명의 실시예에 따른 화면 내 예측 복호화 방법을 나타내는 흐름도이다.Hereinafter, an intra prediction prediction method according to an embodiment of the present invention will be described with reference to FIGS. 17 and 18. 17 is a flowchart illustrating an intra prediction prediction method according to an embodiment of the present invention.
우선 영상 복호화 장치는 비트스트림으로부터 현재블록의 우측 하단 모서리에 위치한 제1 인접화소를 지시하는 정보를 복호화하여 제1 인접화소(BR)의 값을 설정한다(S1710). 제1 인접화소(BR)의 값을 설정하는 방법에는 전술한 다양한 실시예들에 따른 방법이 적용될 수 있다.First, the image decoding apparatus decodes information indicating the first adjacent pixel located at the lower right corner of the current block from the bitstream and sets a value of the first adjacent pixel BR (S1710). The method according to the aforementioned various embodiments may be applied to the method of setting the value of the first adjacent pixel BR.
예를 들어, 제1 인접화소(BR)를 지시하는 정보는 제1 인접화소(BR)의 원본값과 가장 유사한 값을 갖는 화소의 위치를 나타내는 정보일 수 있다. 이 경우, 단계 S1710에서는 현재블록의 참조화소들 중 가장 유사한 값을 갖는 화소의 위치를 나타내는 정보가 지시하는 위치의 화소의 값을 제1 인접화소(BR)의 값으로 설정할 수 있다. 화소의 위치를 나타내는 정보는 현재블록의 좌측 상단 모서리에 위치한 인접화소(TL), 우측 상단 모서리에 위치한 인접화소(TR) 및 우측 하단 모서리에 위치한 인접화소(BL) 중 어느 하나의 위치를 기준으로 하여 해당 화소의 위치를 지시할 수 있다. 다른 예로, 영상 복호화 장치는 비트스트림으로부터 현재블록의 참조화소들의 평균값 또는 가중 평균값으로 구성된 복수의 제1 인접화소(BR)의 후보값들 중 하나를 지시하는 정보를 복호화하여, 복호화된 정보가 지시하는 하나의 제1 인접화소(BR)의 후보값을 제1 인접화소(BR)의 값으로 설정할 수 있다. 여기서, 복호화된 정보가 지시하는 하나의 제1 인접화소(BR)의 후보값은, 복수의 제1 인접화소(BR)의 후보값들 중 제1 인접화소(BR)의 원본값과 가장 유사한 값을 의미할 수 있다.For example, the information indicating the first adjacent pixel BR may be information indicating a position of a pixel having a value most similar to the original value of the first adjacent pixel BR. In this case, in operation S1710, the pixel value at the position indicated by the information indicating the position of the pixel having the most similar value among the reference pixels of the current block may be set as the value of the first adjacent pixel BR. The information indicating the position of the pixel is based on the position of any one of adjacent pixels TL located at the upper left corner of the current block, adjacent pixels TR located at the upper right corner, and adjacent pixels BL located at the lower right corner. To indicate the position of the pixel. As another example, the apparatus for decoding an image decodes information indicating one of candidate values of a plurality of first neighboring pixels BR configured as an average value or a weighted average value of the reference pixels of the current block from the bitstream, and the decoded information is indicated. The candidate value of one first adjacent pixel BR may be set to the value of the first adjacent pixel BR. Here, the candidate value of one first adjacent pixel indicated by the decoded information is the value most similar to the original value of the first adjacent pixel BR among the candidate values of the plurality of first adjacent pixels BR. It may mean.
제1 인접화소(BR)의 값이 설정되면, 영상 복호화 장치는 제1 인접화소(BR)의 값, 현재블록의 우측 상단 모서리에 위치한 제2 인접화소(TR)의 값 및 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소(BL)의 값을 이용한 보간에 의해 현재블록의 하측 인접화소의 예측값 및 우측 인접화소의 예측값을 획득한다(S1720). 구체적으로, 우측 인접화소의 예측값은 제1 인접화소(BR)의 값 및 제2 인접화소(TR)의 값을 이용한 보간에 의해 획득되고, 하측 인접화소의 예측값은 제1 인접화소(BR)의 값 및 제3 인접화소(BL)의 값을 이용한 보간에 의해 획득된다. 여기서, 좌측 인접화소와 우측 인접화소는 현재화소와 동일한 행에 위치하고, 상측 인접화소와 하측 인접화소는 현재화소와 동일한 열에 위치할 수 있다.When the value of the first neighboring pixel BR is set, the image decoding apparatus determines the value of the first neighboring pixel BR, the value of the second neighboring pixel TR located at the upper right corner of the current block, and the lower left of the current block. The interpolation using the value of the third neighboring pixel BL located at the corner is obtained to obtain the predicted value of the lower neighboring pixel and the predicted value of the right neighboring pixel in step S1720. Specifically, the predicted value of the right neighboring pixel is obtained by interpolation using the value of the first neighboring pixel BR and the value of the second neighboring pixel TR, and the predicted value of the lower neighboring pixel is obtained from the first neighboring pixel BR. Value is obtained by interpolation using the value and the value of the third adjacent pixel BL. Here, the left adjacent pixel and the right adjacent pixel may be positioned in the same row as the current pixel, and the upper adjacent pixel and the lower adjacent pixel may be positioned in the same column as the current pixel.
마지막으로, 영상 복호화 장치는 현재블록의 좌측 인접화소의 값, 현재블록의 상측 인접화소의 값, 하측 인접화소의 예측값 및 우측 인접화소의 예측값을 이용한 보간에 의해 현재블록 내 현재화소의 예측값을 획득한다(S1730). 구체적으로, 영상 복호화 장치는 좌측 인접화소의 값과 우측 인접화소의 예측값을 이용한 보간에 의해 현재화소의 제1 예측값을 획득하고, 상측 인접화소의 값과 하측 인접화소의 예측값을 이용한 보간에 의해 현재화소의 제2 예측값을 획득한다. 현재화소의 최종 예측값은 제1 예측값과 제2 예측값을 이용하여 획득된다. 예컨대, 최종 예측값은 제1 예측값과 제2 예측값의 평균값이 될 수 있다.Finally, the image decoding apparatus obtains the predicted value of the current pixel in the current block by interpolation using the value of the left neighboring pixel of the current block, the upper neighboring pixel of the current block, the predicted value of the lower neighboring pixel, and the predicted value of the right neighboring pixel. (S1730). Specifically, the image decoding apparatus obtains a first prediction value of the current pixel by interpolation using the value of the left neighboring pixel and the predicted value of the right neighboring pixel, and currently interpolates using the value of the upper neighboring pixel and the prediction value of the lower neighboring pixel. Obtain a second predicted value of the pixel. The final prediction value of the current pixel is obtained using the first prediction value and the second prediction value. For example, the final prediction value may be an average value of the first prediction value and the second prediction value.
도 18은 본 발명의 다른 실시예에 따른 화면 내 예측 복호화 방법을 나타내는 흐름도이다.18 is a flowchart illustrating an intra prediction prediction method according to another embodiment of the present invention.
우선 영상 복호화 장치는 현재블록의 참조화소들을 기반으로 현재블록의 우측 하단 모서리에 위치한 제1 인접화소(BR)의 값을 설정한다(S1810). 본 실시예에 따르면, 제1 인접화소(BR)의 값은 영상 부호화 장치에서 이용된 방법과 동일한 방법으로 영상 복호화 장치에서 직접 계산될 수 있다. 제1 인접화소(BR)의 값을 계산하는 방법은 전술한 다양한 실시예들에 따른 방법들을 포함할 수 있다.First, the image decoding apparatus sets the value of the first adjacent pixel BR located at the lower right corner of the current block based on the reference pixels of the current block (S1810). According to the present embodiment, the value of the first adjacent pixel BR may be directly calculated by the image decoding apparatus in the same manner as that used by the image encoding apparatus. The method of calculating a value of the first adjacent pixel BR may include methods according to the aforementioned various embodiments.
예를 들어, 영상 복호화 장치는 제1 인접화소(BR)로부터 연장된 적어도 하나의 특정 방향의 직선 상에 위치하는 현재블록의 참조화소 또는 인접화소의 값에 이와 동일 직선 상에 위치하는 이미 복원된 화소의 값과의 변화량을 합산하여 해당 방향에서의 제1 인접화소(BR) 예측값을 계산하고, 제1 인접화소(BR) 예측값들의 평균 또는 가중 평균으로 제1 인접화소(BR)의 값을 설정할 수 있다. 여기서, 특정 방향은 제1 인접화소(BR)로부터 직선 방향, 수직 방향 및 대각선 방향 중 적어도 하나일 수 있고, 현재블록의 참조화소는 현재블록의 좌측 상단 모서리에 위치한 주변화소, 제2 인접화소(TR) 및 제3 인접화소(BL) 중 적어도 하나일 수 있다. 다른 예로, 영상 복호화 장치는 현재블록의 참조화소들의 평균값 또는 가중 평균값을 제1 인접화소(BR)의 값으로 설정할 수 있다. 또 다른 예로, 영상 복호화 장치는 현재블록의 참조화소들을 기반으로 제1 인접화소(BR)의 예측값을 결정하고, 비트스트림으로부터 제1 인접화소(BR)의 원본값과 예측값의 차분값을 나타내는 정보를 복호화하여 직접 결정한 예측값에 복호화된 차분값을 더함으로써 제1 인접화소(BR)의 값을 결정할 수 있다.For example, the image decoding apparatus may have already been restored on the same straight line to the value of the reference pixel or the adjacent pixel of the current block located on the straight line in at least one specific direction extending from the first adjacent pixel BR. The first neighboring pixel BR prediction value in the corresponding direction is calculated by summing the amount of change with the value of the pixel, and the value of the first neighboring pixel BR is set as an average or a weighted average of the first neighboring pixel BR prediction values. Can be. Here, the specific direction may be at least one of a linear direction, a vertical direction, and a diagonal direction from the first adjacent pixel BR, and the reference pixel of the current block may include a peripheral pixel located at the upper left corner of the current block, a second adjacent pixel ( TR) and at least one of the third adjacent pixels BL. As another example, the apparatus for decoding an image may set the average value or weighted average value of the reference pixels of the current block to the value of the first adjacent pixel BR. As another example, the apparatus for decoding an image may determine the prediction value of the first neighboring pixel BR based on the reference pixels of the current block, and indicate the difference value between the original value and the prediction value of the first neighboring pixel BR from the bitstream. The value of the first neighboring pixel BR may be determined by decoding the sum and adding the decoded difference value to the directly determined prediction value.
도 17 및 도 18에서는 각 과정을 순차적으로 실행하는 것으로 기재하고 있으나, 반드시 이에 한정되는 것은 아니다. 다시 말해, 도 17 및 도 18에 기재된 과정을 변경하여 실행하거나 하나 이상의 과정을 병렬적으로 실행하는 것으로 적용 가능할 것이므로, 도 17 및 도 18은 시계열적인 순서로 한정되는 것은 아니다.17 and 18 are described as sequentially executing each process, but is not necessarily limited thereto. In other words, since the process described in FIGS. 17 and 18 may be applied by changing the process or executing one or more processes in parallel, FIGS. 17 and 18 are not limited to the time series order.
도 17 및 도 18에 기재된 본 실시예에 따른 영상 부호화 또는 복호화 방법은 컴퓨터 프로그램으로 구현되고 컴퓨터로 읽을 수 있는 기록매체에 기록될 수 있다. 본 실시예에 따른 영상 부호화 또는 복호화 방법을 구현하기 위한 컴퓨터 프로그램이 기록되고 컴퓨터가 읽을 수 있는 기록매체는 컴퓨팅 시스템에 의하여 읽혀질 수 있는 데이터가 저장되는 모든 종류의 기록장치를 포함한다.The image encoding or decoding method according to the present embodiment described in FIGS. 17 and 18 may be implemented in a computer program and recorded on a computer-readable recording medium. A computer program for recording an image encoding or decoding method according to the present embodiment and a computer readable recording medium include all kinds of recording devices that store data that can be read by a computing system.
이상의 설명은 본 실시예의 기술 사상을 예시적으로 설명한 것에 불과한 것으로서, 본 실시예가 속하는 기술 분야에서 통상의 지식을 가진 자라면 본 실시예의 본질적인 특성에서 벗어나지 않는 범위에서 다양한 수정 및 변형이 가능할 것이다. 따라서, 본 실시예들은 본 실시예의 기술 사상을 한정하기 위한 것이 아니라 설명하기 위한 것이고, 이러한 실시예에 의하여 본 실시예의 기술 사상의 범위가 한정되는 것은 아니다. 본 실시예의 보호 범위는 아래의 청구범위에 의하여 해석되어야 하며, 그와 동등한 범위 내에 있는 모든 기술 사상은 본 실시예의 권리범위에 포함되는 것으로 해석되어야 할 것이다.The above description is merely illustrative of the technical idea of the present embodiment, and those skilled in the art to which the present embodiment belongs may make various modifications and changes without departing from the essential characteristics of the present embodiment. Therefore, the present embodiments are not intended to limit the technical idea of the present embodiment but to describe the present invention, and the scope of the technical idea of the present embodiment is not limited by these embodiments. The scope of protection of the present embodiment should be interpreted by the following claims, and all technical ideas within the scope equivalent thereto should be construed as being included in the scope of the present embodiment.
CROSS-REFERENCE TO RELATED APPLICATIONCROSS-REFERENCE TO RELATED APPLICATION
본 특허출원은, 본 명세서에 그 전체가 참고로서 포함되는, 2017년 05월 18일 및 2017년 07월 12일에 각각 한국에 출원한 특허출원번호 제10-2017-0061887호 및 제10-2017-0088703호에 대해 우선권을 주장한다.This patent application is incorporated by reference in Korean Patent Application Nos. 10-2017-0061887 and 10-2017, which were filed in Korea on May 18, 2017 and July 12, 2017, which are hereby incorporated by reference in their entirety. Claim priority on -0088703.

Claims (16)

  1. 비트스트림으로부터 현재블록의 우측 하단 모서리에 위치한 제1 인접화소를 지시하는 정보를 복호화하여 상기 제1 인접화소의 값을 설정하는 단계;Setting a value of the first neighboring pixel by decoding information indicating a first neighboring pixel located at a lower right corner of the current block from a bitstream;
    상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 획득하는 단계; 및The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; Obtaining a value of a lower adjacent pixel and a value of a right adjacent pixel of the pixel; And
    상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 획득하는 단계Obtaining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel
    를 포함하는 영상 복호화 방법.Image decoding method comprising a.
  2. 제1항에 있어서,The method of claim 1,
    상기 제1 인접화소를 지시하는 정보는, 상기 제1 인접화소의 원본값과 가장 유사한 값을 갖는 화소의 위치를 나타내는 정보이고,The information indicating the first adjacent pixel is information indicating a position of a pixel having a value most similar to the original value of the first adjacent pixel.
    상기 제1 인접화소의 값을 설정하는 단계는, 상기 현재블록의 참조화소들 중 상기 화소의 위치를 나타내는 정보가 지시하는 위치의 화소의 값을 상기 제1 인접화소의 값으로 설정하는 단계를 포함하는, 영상 복호화 방법.The setting of the value of the first neighboring pixel includes setting the value of the pixel at the position indicated by the information indicating the position of the pixel among the reference pixels of the current block to the value of the first neighboring pixel. A video decoding method.
  3. 제2항에 있어서,The method of claim 2,
    상기 화소의 위치를 나타내는 정보는, 상기 현재블록의 좌측 상단 모서리에 위치한 주변화소, 상기 제2 인접화소 및 상기 제3 인접화소 중 어느 하나의 위치를 기준으로 하여 상기 제1 인접화소의 값과 가장 유사한 값을 갖는 화소의 위치를 지시하는, 영상 복호화 방법.Information indicating the position of the pixel is based on the position of any one of the neighboring pixel, the second neighboring pixel, and the third neighboring pixel located at the upper left corner of the current block. And a position of a pixel having a similar value.
  4. 현재블록의 참조화소들을 기반으로 상기 현재블록의 우측 하단 모서리에 위치한 제1 인접화소의 값을 설정하는 단계;Setting a value of a first adjacent pixel located at a lower right corner of the current block based on reference pixels of the current block;
    상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 획득하는 단계; 및The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; Obtaining a value of a lower adjacent pixel and a value of a right adjacent pixel of the pixel; And
    상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 획득하는 단계Obtaining a predicted value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel
    를 포함하는 영상 복호화 방법.Image decoding method comprising a.
  5. 제4항에 있어서,The method of claim 4, wherein
    상기 제1 인접화소의 값을 설정하는 단계는,The setting of the value of the first adjacent pixel may include:
    상기 제1 인접화소로부터 연장된 적어도 하나의 특정 방향의 직선 상에 위치하는 상기 현재블록의 참조화소의 값에 이와 동일 직선 상에 위치하는 이미 복원된 주변화소의 값과의 변화량을 합산하여 상기 적어도 하나의 특정 방향 별로 상기 제1 인접화소의 예측값을 계산하는 단계; 및The value of the reference pixel of the current block located on a straight line in at least one specific direction extending from the first adjacent pixel is added to the change amount with the value of the already restored peripheral pixel located on the same straight line to add the at least Calculating a prediction value of the first neighboring pixel in one specific direction; And
    상기 적어도 하나의 특정 방향 별로 계산된 상기 제1 인접화소의 예측값의 평균을 상기 제1 인접화소의 값으로 설정하는 단계를 포함하는, 영상 복호화 방법.And setting an average of predicted values of the first neighboring pixel calculated for each of the at least one specific direction to a value of the first neighboring pixel.
  6. 제5항에 있어서,The method of claim 5,
    상기 특정 방향은, 상기 제1 인접화소로부터 수평 방향, 수직 방향 및 대각선 방향 중 적어도 하나이고,The specific direction is at least one of a horizontal direction, a vertical direction and a diagonal direction from the first adjacent pixel,
    상기 현재블록의 참조화소는, 상기 현재블록의 좌측 상단 모서리에 위치한 인접화소, 상기 제2 인접화소 및 상기 제3 인접화소 중 적어도 하나인, 영상 복호화 방법.The reference pixel of the current block is at least one of a neighboring pixel located at an upper left corner of the current block, the second neighboring pixel, and the third neighboring pixel.
  7. 제4항에 있어서,The method of claim 4, wherein
    상기 제1 인접화소의 값을 설정하는 단계는,The setting of the value of the first adjacent pixel may include:
    상기 현재블록의 참조화소들의 평균값 또는 가중 평균값을 상기 제1 인접화소의 값으로 설정하는 단계를 포함하는, 영상 복호화 방법.And setting an average value or a weighted average value of the reference pixels of the current block to the value of the first adjacent pixel.
  8. 제4항에 있어서,The method of claim 4, wherein
    상기 제1 인접화소의 값을 설정하는 단계는,The setting of the value of the first adjacent pixel may include:
    비트스트림으로부터 상기 현재블록의 참조화소들의 평균값 또는 가중 평균값으로 구성된 복수의 상기 제1 인접화소의 후보값들 중 하나를 지시하는 정보를 복호화하는 단계; 및Decoding information indicating one of candidate values of a plurality of first neighboring pixels, the average value or a weighted average value of reference pixels of the current block, from a bitstream; And
    상기 복호화된 정보가 지시하는 하나의 상기 제1 인접화소의 후보값을 상기 제1 인접화소의 값으로 설정하는 단계를 더 포함하되,Setting a candidate value of one first neighboring pixel indicated by the decoded information to a value of the first neighboring pixel,
    상기 복호화된 정보가 지시하는 하나의 상기 제1 인접화소의 후보값은, 상기 복수의 제1 인접화소의 후보값들 중 상기 제1 인접화소의 원본값과 가장 유사한 값인, 영상 복호화 방법.And a candidate value of one first neighboring pixel indicated by the decoded information is a value most similar to an original value of the first neighboring pixel among candidate values of the plurality of first neighboring pixels.
  9. 제4항에 있어서,The method of claim 4, wherein
    상기 제1 인접화소의 값을 설정하는 단계는,The setting of the value of the first adjacent pixel may include:
    상기 현재블록의 참조화소들을 기반으로 상기 제1 인접화소의 예측값을 결정하는 단계; 및Determining a prediction value of the first neighboring pixel based on reference pixels of the current block; And
    비트스트림으로부터 상기 제1 인접화소의 원본값과 상기 예측값의 차분값을 나타내는 정보를 복호화하여 상기 예측값에 상기 차분값을 더함으로써 결정된 값을 상기 제1 인접화소의 값으로 설정하는 단계를 포함하는, 영상 복호화 방법.Decoding the information representing the difference between the original value of the first neighboring pixel and the prediction value from the bitstream and setting the value determined by adding the difference value to the prediction value to the value of the first neighboring pixel; Image Decoding Method.
  10. 제4항에 있어서,The method of claim 4, wherein
    상기 현재블록은, 2D 이미지로 부호화된 360 영상에서 복호화하고자 하는 제1 면에 위치하고,The current block is located on a first surface to be decoded from a 360 image encoded as a 2D image,
    상기 제1 인접화소의 값을 설정하는 단계는,The setting of the value of the first adjacent pixel may include:
    상기 현재블록의 우측 경계가 상기 제1 면의 우측 경계와 일치할 때, 상기 360 영상을 기준으로 상기 현재블록의 우측 경계에 접하며 이미 복호화된 블록을 식별하는 단계;When the right boundary of the current block coincides with the right boundary of the first surface, identifying a block that is in contact with the right boundary of the current block and is already decoded based on the 360 image;
    상기 식별된 블록의 좌측 경계에 접하는 상기 식별된 블록 내 화소들 중 상기 360 영상을 기준으로 상기 제1 인접화소와 접하는 화소의 값을 상기 제1 인접화소의 값으로 설정하는 단계를 포함하는, 영상 복호화 방법.And setting a value of a pixel in contact with the first neighboring pixel based on the 360 image among the pixels in the identified block in contact with a left boundary of the identified block to the value of the first neighboring pixel. Decryption method.
  11. 제1항 또는 제4항에 있어서,The method according to claim 1 or 4,
    상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 획득하는 단계는,Acquiring the value of the lower adjacent pixel and the value of the right adjacent pixel,
    상기 제1 인접화소의 값 및 상기 제2 인접화소의 값을 이용한 보간에 의해 상기 우측 인접화소의 값을 획득하는 단계; 및Obtaining a value of the right adjacent pixel by interpolation using the value of the first adjacent pixel and the value of the second adjacent pixel; And
    상기 제1 인접화소의 값 및 상기 제3 인접화소의 값을 이용한 보간에 의해 상기 하측 인접화소의 값을 획득하는 단계를 포함하는, 영상 복호화 방법.And obtaining a value of the lower neighboring pixel by interpolation using the value of the first neighboring pixel and the value of the third neighboring pixel.
  12. 제1항 또는 제4항에 있어서,The method according to claim 1 or 4,
    상기 현재화소의 예측값을 획득하는 단계는,Acquiring a prediction value of the current pixel,
    상기 좌측 인접화소의 값과 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재화소의 제1 예측값을 획득하는 단계;Obtaining a first predicted value of the current pixel by interpolation using the value of the left neighboring pixel and the value of the right neighboring pixel;
    상기 상측 인접화소의 값과 상기 하측 인접화소의 값을 이용한 보간에 의해 상기 현재화소의 제2 예측값을 획득하는 단계; 및Obtaining a second predicted value of the current pixel by interpolation using the value of the upper neighboring pixel and the value of the lower neighboring pixel; And
    상기 제1 예측값과 상기 제2 예측값을 이용하여 상기 현재화소의 예측값을 획득하는 단계를 포함하되,Obtaining a prediction value of the current pixel by using the first prediction value and the second prediction value,
    상기 좌측 인접화소와 상기 우측 인접화소는 상기 현재화소와 동일한 행에 위치하고, 상기 상측 인접화소와 상기 하측 인접화소는 상기 현재화소와 동일한 열에 위치하는, 영상 복호화 방법.And the left neighboring pixel and the right neighboring pixel are located in the same row as the current pixel, and the upper neighboring pixel and the lower neighboring pixel are located in the same column as the current pixel.
  13. 제12항에 있어서,The method of claim 12,
    상기 현재화소의 예측값은,The prediction value of the current pixel,
    상기 제1 예측값과 상기 제2 예측값을 가중평균하여 생성하는, 영상 복호화 방법.And generating a weighted average of the first prediction value and the second prediction value.
  14. 제13항에 있어서,The method of claim 13,
    상기 제1 예측값에 적용되는 제1 가중치는 상기 현재화소와 동일한 열에 위치한 상측 참조화소로부터 상기 현재화소까지의 거리에 의해 결정되고,The first weight applied to the first prediction value is determined by a distance from an upper reference pixel located in the same column as the current pixel to the current pixel,
    상기 제2 예측값에 적용되는 제2 가중치는 상기 현재화소와 동일한 행에 위치한 좌측 참조화소로부터 상기 현재화소까지의 거리에 의해 결정되는, 영상 복호화 방법.And a second weight applied to the second prediction value is determined by a distance from a left reference pixel located in the same row as the current pixel to the current pixel.
  15. 비트스트림으로부터 현재블록의 우측 하단 모서리에 위치한 제1 인접화소를 지시하는 정보를 복호화하는 복호화부;A decoder which decodes information indicating a first adjacent pixel located at a lower right corner of the current block from the bitstream;
    상기 제1 인접화소를 지시하는 정보를 이용하여 상기 제1 인접화소의 값을 설정하는 제1 인접화소 설정부;A first neighboring pixel setting unit configured to set a value of the first neighboring pixel by using information indicating the first neighboring pixel;
    상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 결정하는 인접화소 예측부; 및The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; A neighboring pixel predictor configured to determine a value of a lower neighboring pixel and a value of a right neighboring pixel of the pixel; And
    상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 결정하는 현재화소 예측부A current value for determining a prediction value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel Pixel prediction unit
    를 포함하는 영상 복호화 장치.Image decoding apparatus comprising a.
  16. 현재블록의 참조화소들을 기반으로 상기 현재블록의 우측 하단 모서리에 위치한 제1 인접화소의 값을 설정하는 제1 인접화소 설정부;A first neighboring pixel setting unit configured to set a value of a first neighboring pixel located at a lower right corner of the current block based on reference pixels of the current block;
    상기 제1 인접화소의 값, 상기 현재블록의 우측 상단 모서리에 위치한 제2 인접화소의 값 및 상기 현재블록의 좌측 하단 모서리에 위치한 제3 인접화소의 값을 이용한 보간(interpolation)에 의해 상기 현재블록의 하측 인접화소의 값 및 우측 인접화소의 값을 결정하는 인접화소 예측부; 및The current block by interpolation using the value of the first neighboring pixel, the value of the second neighboring pixel located at the upper right corner of the current block, and the value of the third neighboring pixel located at the lower left corner of the current block; A neighboring pixel predictor configured to determine a value of a lower neighboring pixel and a value of a right neighboring pixel of the pixel; And
    상기 현재블록의 좌측 인접화소의 값, 상기 현재블록의 상측 인접화소의 값, 상기 하측 인접화소의 값 및 상기 우측 인접화소의 값을 이용한 보간에 의해 상기 현재블록 내 현재화소의 예측값을 결정하는 현재화소 예측부A current value for determining a prediction value of the current pixel in the current block by interpolation using a value of a left neighboring pixel of the current block, a value of an upper neighboring pixel of the current block, a value of the lower neighboring pixel, and a value of the right neighboring pixel Pixel prediction unit
    를 포함하는 영상 복호화 장치.Image decoding apparatus comprising a.
PCT/KR2018/005596 2017-05-18 2018-05-16 Intra prediction encoding or decoding method and device WO2018212582A1 (en)

Applications Claiming Priority (4)

Application Number Priority Date Filing Date Title
KR20170061887 2017-05-18
KR10-2017-0061887 2017-05-18
KR10-2017-0088703 2017-07-12
KR1020170088703A KR102356317B1 (en) 2017-05-18 2017-07-12 Method and Apparatus for Intra Prediction Encoding and Decoding

Publications (1)

Publication Number Publication Date
WO2018212582A1 true WO2018212582A1 (en) 2018-11-22

Family

ID=64274198

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/KR2018/005596 WO2018212582A1 (en) 2017-05-18 2018-05-16 Intra prediction encoding or decoding method and device

Country Status (1)

Country Link
WO (1) WO2018212582A1 (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10645381B2 (en) * 2018-04-30 2020-05-05 Google Llc Intra-prediction for smooth blocks in image/video
US10834410B2 (en) 2017-10-23 2020-11-10 Google Llc Context modeling for intra-prediction modes

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110249741A1 (en) * 2010-04-09 2011-10-13 Jie Zhao Methods and Systems for Intra Prediction
KR20140008503A (en) * 2012-07-10 2014-01-21 한국전자통신연구원 Method and apparatus for image encoding/decoding
KR20160143586A (en) * 2015-06-05 2016-12-14 인텔렉추얼디스커버리 주식회사 Method and apparatus of image encoding/decoding using reference pixel composition in intra prediction
US20170085917A1 (en) * 2015-09-23 2017-03-23 Nokia Technologies Oy Method, an apparatus and a computer program product for coding a 360-degree panoramic video
KR20170034799A (en) * 2017-02-22 2017-03-29 엠앤케이홀딩스 주식회사 Apparatus for decoding an image

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20110249741A1 (en) * 2010-04-09 2011-10-13 Jie Zhao Methods and Systems for Intra Prediction
KR20140008503A (en) * 2012-07-10 2014-01-21 한국전자통신연구원 Method and apparatus for image encoding/decoding
KR20160143586A (en) * 2015-06-05 2016-12-14 인텔렉추얼디스커버리 주식회사 Method and apparatus of image encoding/decoding using reference pixel composition in intra prediction
US20170085917A1 (en) * 2015-09-23 2017-03-23 Nokia Technologies Oy Method, an apparatus and a computer program product for coding a 360-degree panoramic video
KR20170034799A (en) * 2017-02-22 2017-03-29 엠앤케이홀딩스 주식회사 Apparatus for decoding an image

Cited By (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US10834410B2 (en) 2017-10-23 2020-11-10 Google Llc Context modeling for intra-prediction modes
US10893280B2 (en) 2017-10-23 2021-01-12 Google Llc Refined entropy coding for level maps
US11178409B2 (en) 2017-10-23 2021-11-16 Google Llc Probability mapping for entropy coding
US11477462B2 (en) 2017-10-23 2022-10-18 Google Llc Entropy coding using transform class
US10645381B2 (en) * 2018-04-30 2020-05-05 Google Llc Intra-prediction for smooth blocks in image/video
US11039131B2 (en) 2018-04-30 2021-06-15 Google Llc Intra-prediction for smooth blocks in image/video

Similar Documents

Publication Publication Date Title
WO2017052000A1 (en) Method and apparatus for motion vector refinement-based inter prediction in image coding system
WO2015142054A1 (en) Method and apparatus for processing multiview video signals
WO2013069932A1 (en) Method and apparatus for encoding image, and method and apparatus for decoding image
WO2016200043A1 (en) Method and apparatus for inter prediction on basis of virtual reference picture in video coding system
WO2011068331A2 (en) Video encoding device and encoding method thereof, video decoding device and decoding method thereof, and directional intra-prediction method to be used thereto
WO2013169031A1 (en) Method and apparatus for processing video signals
WO2019198997A1 (en) Intra-prediction-based image coding method and apparatus thereof
WO2020009390A1 (en) Image processing method and device by means of inter-prediction in image coding system
WO2016056822A1 (en) 3d video coding method and device
WO2019194500A1 (en) Intra-prediction-based image coding method and device therefor
WO2019240425A1 (en) Inter-prediction method and image decoding device
WO2017052272A1 (en) Method and apparatus for intra prediction in video coding system
WO2018062699A1 (en) Image decoding method and apparatus in image coding system
WO2014010918A1 (en) Method and device for processing video signal
WO2019135628A1 (en) Method and device for encoding or decoding image
WO2022114742A1 (en) Apparatus and method for video encoding and decoding
WO2020141885A1 (en) Image coding method and device using deblocking filtering
WO2018212582A1 (en) Intra prediction encoding or decoding method and device
WO2014073877A1 (en) Method and apparatus for processing multiview video signal
WO2013133587A1 (en) Method and apparatus for processing video signals
WO2019135636A1 (en) Image coding/decoding method and apparatus using correlation in ycbcr
WO2020004931A1 (en) Method and apparatus for processing image according to inter-prediction in image coding system
WO2018131838A1 (en) Method and device for image decoding according to intra-prediction in image coding system
WO2018128511A1 (en) Device and method for encoding or decoding image
WO2020009375A1 (en) Intra prediction method and device in image coding system

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 18802183

Country of ref document: EP

Kind code of ref document: A1

NENP Non-entry into the national phase

Ref country code: DE

122 Ep: pct application non-entry in european phase

Ref document number: 18802183

Country of ref document: EP

Kind code of ref document: A1