JP2018056685A - Image encoder, image encoding method and image encoding program, and image decoder, image decoding method and image decoding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image encoder which enables the decrease in code amount by utilizing the dependence between video signal components in CCLM prediction to adjust quantization values of a reference source and a reference destination.SOLUTION: In an image encoder 100, an intra prediction part 14 uses one of first and second prediction modes to produce predicted values of color components in respective block in a picture, provided that in the first prediction mode, decoded pixel values of brightness components are used to produce predicted values of color components according to a linear model, and in the second prediction mode predicted values of color components are produced without using decoded pixel values of brightness signals; a subtracter 2 subtracts a predicted value from each color component of an original image to produce a predictive residue; an orthogonal conversion part 3 performs an orthogonal transformation of the predictive residue to produce an orthogonal transform coefficient; and a quantization part 4 sets quantization values of reference destination color components to be higher than those of reference source color components to quantize orthogonal transform coefficients when a predicted value-generation part generates predicted values of color components with the first prediction mode is selected.SELECTED DRAWING: Figure 2

Description

  The present disclosure relates to an image encoding device, an image encoding method, an image encoding program, an image decoding device, an image decoding method, and an image decoding program.

  H.265 / MPEG-H HEVC (High Efficiency Video Coding) (hereinafter referred to as HEVC), which can encode image information more efficiently than H.264 / MPEG-4 AVC (Advanced Video Coding). Is abbreviated). In HEVC, intra prediction that generates a prediction value by performing spatial prediction within a frame and inter prediction that generates a prediction value by performing motion compensation prediction between frames are used.

“Algorithm Description of Joint Exploration Test Model 1”, JVET-B0021, February 2016 “CE6.a.4: Chroma intra prediction by reconstructed luma samples”, JCTVC-E266, March2011

  In Non-Patent Documents 1 and 2, as an improvement in intra prediction by developing HEVC, a cross-component liner model prediction that generates a prediction value of a color difference signal using a decoded pixel value of a luminance signal is proposed. ) Has been proposed. Non-Patent Document 1 describes that the prediction value of one color difference signal is corrected using the prediction residual of one color difference signal. Cross component linear model prediction is abbreviated as CCLM prediction.

  Unlike other intra predictions, CCLM prediction has a dependency relationship between components of a video signal (luminance signal and color difference signal). In Non-Patent Documents 1 and 2, there is no means for adjusting the quantization value of the reference source and the reference destination by using the dependency relationship and reducing the code amount.

  Embodiments provide an image encoding apparatus and an image encoding method capable of adjusting a quantization value of a reference source and a reference destination and reducing a code amount by using a dependency relationship between components of a video signal in CCLM prediction And an image encoding program, an image decoding device, an image decoding method, and an image decoding program.

  According to the first aspect of the embodiment, the first prediction mode for generating the prediction value of the color component by the linear model using the decoded pixel value of the luminance component, and the color component without using the decoded pixel value of the luminance signal A prediction value generation unit that generates a prediction value of a color component in each block in the picture using any one of the second prediction modes for generating a prediction value of the image, and the prediction from the color component of the original image A subtractor that subtracts a value to generate a prediction residual, an orthogonal transform unit that orthogonally transforms the prediction residual to generate an orthogonal transform coefficient, and the predicted value generation unit selects the first prediction mode. A quantizing unit configured to quantize the orthogonal transform coefficient by setting a quantized value of the reference destination color component higher than that of the reference source color component. Is provided.

  According to the second aspect of the embodiment, the first prediction mode for generating the prediction value of the color component by the linear model using the decoded pixel value of the luminance component, and the color component without using the decoded pixel value of the luminance signal A prediction value of a color component is generated using any one of the second prediction modes for generating a prediction value of the image, a prediction residual is generated by subtracting the prediction value from the color component of the original image, When the prediction residual is orthogonally transformed to generate an orthogonal transform coefficient, and when the first prediction mode is selected to generate a color component prediction value, the reference destination color component is compared with the reference source color component An image encoding method is provided in which the orthogonal transform coefficient is quantized by setting a high quantization value of.

  According to the third aspect of the embodiment, the computer uses the first prediction mode for generating the prediction value of the color component by the linear model using the decoded pixel value of the luminance component, and the decoded pixel value of the luminance signal. The second prediction mode for generating the color component prediction value without generating the color component prediction value, and subtracting the prediction value from the color component of the original image to generate the prediction residual. A step of generating a difference, a step of orthogonally transforming the prediction residual to generate an orthogonal transformation coefficient, and a color component of a reference source when the first prediction mode is selected to generate a color component prediction value And a step of quantizing the orthogonal transform coefficient by setting a quantized value of a reference color component higher than that of the image encoding program.

  According to the fourth aspect of the embodiment, the first prediction mode for generating the prediction value of the color component by the linear model using the decoded pixel value of the luminance component, and the color component without using the decoded pixel value of the luminance signal Entropy decoding for entropy decoding a bitstream in which a prediction value of a color component in each block in a picture is generated and encoded using any one of the second prediction modes for generating a prediction value of And the orthogonal transform coefficient entropy-decoded by the entropy decoding unit in the block in which the prediction value of the color component is generated and encoded using the first prediction mode is compared with the color component of the reference source Thus, an image decoding apparatus is provided that includes an inverse quantization unit configured to inversely quantize by setting a quantization value of a reference color component high.

  According to the fifth aspect of the embodiment, the first prediction mode for generating the prediction value of the color component by the linear model using the decoded pixel value of the luminance component, and the color component without using the decoded pixel value of the luminance signal Entropy decoding the bitstream in which the prediction value of the color component in each block in the picture is generated and encoded using any one of the second prediction modes for generating the prediction value of In the block in which the predicted value of the color component is generated and encoded using the first prediction mode, the entropy-decoded orthogonal transform coefficient is compared with the color component of the reference source and the quantum of the color component of the reference destination An image decoding method is provided that performs inverse quantization by setting a high quantization value.

  According to the sixth aspect of the embodiment, the computer uses the first prediction mode for generating the prediction value of the color component by the linear model using the decoded pixel value of the luminance component, and the decoded pixel value of the luminance signal. Entropy decoding of the bitstream in which the prediction value of the color component in each block in the picture is generated and encoded using one of the second prediction modes for generating the prediction value of the color component without any And comparing the entropy-decoded orthogonal transform coefficient with a reference source color component in a block in which a prediction value of a color component is generated and encoded using the first prediction mode. An image decoding program is provided that executes a step of performing inverse quantization by setting a quantization value of a color component of the image to a high value.

  According to the image encoding device, the image encoding method, and the image encoding program of the embodiment, and the image decoding device, the image decoding method, and the image decoding program, the dependency relationship between the components of the video signal in CCLM prediction is used. Thus, the amount of code can be reduced by adjusting the quantization values of the reference source and the reference destination.

FIG. 1 is a block diagram illustrating an overall configuration example of an image encoding / decoding system including an image encoding device and an image decoding device. FIG. 2 is a block diagram illustrating an image encoding device according to an embodiment. FIG. 3 is a diagram illustrating the structure of the coding tree unit. FIG. 4 is a diagram illustrating the structure of the prediction unit. FIG. 5 is a diagram illustrating the structure of the conversion unit. FIG. 6 is a diagram illustrating a configuration example of a frame divided into a plurality of slices. FIG. 7 is a diagram illustrating an example of a prediction mode of intra prediction. FIG. 8 is a diagram showing the relationship between the block Y signal block of the Cb and Cr signals in the case of a 4: 2: 0 format Y, Cb, and Cr video signal. FIG. 9 is a block diagram illustrating a specific configuration example of the quantization unit 4 and the inverse quantization unit 8 in FIG. FIG. 10 is a diagram illustrating an example of syntax transmitted by the image encoding device according to the embodiment. FIG. 11 is a flowchart illustrating the operation of the image encoding device according to the embodiment, the processing by the image encoding method according to the embodiment, and the processing executed by the image encoding program according to the embodiment. FIG. 12A is a block diagram illustrating a schematic configuration of a computer including a storage unit that stores an image encoding program according to an embodiment. FIG. 12B is a block diagram illustrating a schematic configuration of a computer including a storage unit that stores an image decoding program according to an embodiment. FIG. 13 is a block diagram illustrating an image decoding apparatus according to an embodiment. FIG. 14 is a block diagram illustrating a specific configuration example of the inverse quantization unit 23 in FIG. 13. FIG. 15 is a flowchart illustrating an operation of the image decoding apparatus according to the embodiment, processing by the image decoding method according to the embodiment, and processing executed by the image decoding program according to the embodiment.

  Hereinafter, an image encoding device, an image encoding method, an image encoding program, an image decoding device, an image decoding method, and an image decoding program according to an embodiment will be described with reference to the accompanying drawings.

  First, an overall configuration example of an image encoding / decoding system including an image encoding device and an image decoding device will be described with reference to FIG. In FIG. 1, the preprocessing device 50 sets various conditions for encoding image information by the image encoding device 100 according to an embodiment in accordance with a user operation. The image encoding device 100 encodes the image information according to the setting information from the preprocessing device 50 and outputs a bit stream. The bit stream includes encoded image information and syntax.

  The transmission device 110 transmits a bit stream to a predetermined transmission path 120. The transmission path 120 is wired or wireless, and may be a communication line such as the Internet, a telephone line, or a radio wave for terrestrial broadcasting or satellite broadcasting that transmits a television signal. The receiving device 150 receives the bit stream transmitted through the transmission path 120. The image decoding apparatus 200 according to an embodiment decodes a bit stream and outputs decoded image information.

<Image encoding device>
FIG. 2 shows a specific configuration example of the image encoding device 100. In FIG. 2, pixels of image information of a digital signal are sequentially input to a reorder buffer 1. If the image information is an analog signal, it may be converted into a digital signal by the A / D converter in the previous stage of the rearrangement buffer 1. The image information is, for example, a luminance signal Y (hereinafter referred to as Y signal) and color difference signals Cb and Cr (hereinafter referred to as Cb and Cr signals), and Y, Cb, Cr video signals in 4: 2: 0 format are taken as an example.

  The rearrangement buffer 1 accumulates input pixels for a plurality of frames, and rearranges and reads frames (pictures) as necessary. Each frame of image information is encoded using an I picture that is encoded using pixels in the frame, a P picture that is encoded predictively using pixels in a past frame, and a predictive encoding using pixels in past and future frames Is set to one of the B pictures. The I picture, P picture, and B picture may be set by the preprocessing device 50, or may be selected by the image encoding device 100 according to a predetermined rule.

  The rearrangement buffer 1 rearranges the order of frames for encoding when a plurality of picture groups (GOPs), which are constituent units of a sequence constituting a bit stream to be described later, are included.

  The frame output from the rearrangement buffer 1 is processed by being divided into units composed of a plurality of pixels as follows. As shown in FIG. 3, the pixels constituting the frame are divided into, for example, a coding tree unit (CTU: Coding Tree Unit) having 64 horizontal pixels and 64 vertical pixels. CTU is CU hierarchy 0. Each CTU may be divided into variable size coding units (CUs) based on recursive quadtree block partitioning.

  A CTU may be divided into CUs of CU hierarchy 1 having 32 horizontal pixels and 32 vertical pixels. The CU of CU layer 1 may be divided into CUs of CU layer 2 of 16 pixels horizontally and 16 pixels vertically. The CU of CU layer 2 may be divided into CUs of CU layer 3 of 8 horizontal pixels and 8 vertical pixels. When the CTU is not divided, the CTU becomes the CU as it is.

  The maximum size CU is referred to as a maximum coding unit (LCU), and the minimum size CU is referred to as a minimum coding unit (SCU).

  By the way, in the 4: 2: 0 format Y, Cb, Cr video signal, the resolution of the Cb signal and the Cr signal is ½ both in the horizontal and vertical directions compared to the resolution of the Y signal. Therefore, the CU includes a Y signal coding block (CB) having the same size as that of the CU, and a Cb signal and a CB of the Cr signal that are ¼ the size of the CU.

  The CU is divided into prediction units (PUs) for prediction processing in the intra prediction unit 14 and the inter prediction unit 15 in FIG. As shown in FIG. 4, 2N × 2N or N × N PUs are selected in a CU (intra CU) encoded based on intra-frame prediction by the intra prediction unit 14. However, the N × N PU is used only by the SCU.

  As shown in FIG. 4, in the CU (inter CU) encoded based on the inter-frame prediction by the inter prediction unit 15, 2N × 2N, N × N, 2N × N, N × 2N, 2N × nU, PUs of 2N × nD, nL × 2N, and nR × 2N are selected. However, N × N can be used only by SCUs where N is 16 or more, and 2N × nU, 2N × nD, nL × 2N, and nR × 2N can be used only when asymmetric motion division is enabled. . 2N × nU, 2N × nD, nL × 2N, and nR × 2N are PUs that divide the CU by 1: 3 vertically, 3: 1 vertically, 1: 3 horizontally, and 3: 1 horizontally.

  Similarly, the PU includes a prediction block (PB) of a Y signal having the same size as the PU, and a Pb of a Cb signal and a Cr signal having a size of 1/4 of the PU.

  Further, the CU is a variable size transform unit (TU: Transform Unit) based on recursive quadtree block division for the orthogonal transform process in the orthogonal transform unit 3 and the quantization process in the quantization unit 4 in FIG. ).

  As shown in FIG. 5, the maximum size of a TU is 32 pixels horizontally and 32 pixels vertically. If the TU of 32 horizontal pixels and 32 vertical pixels is TU layer 0, TU layer 0 may be divided into TUs of TU layer 1 of 16 horizontal pixels and 16 vertical pixels. The TU of TU layer 1 may be divided into TUs of TU layer 2 of 8 horizontal pixels and 8 vertical pixels. The TU of the TU hierarchy 2 may be divided into TUs of the TU hierarchy 3 of 4 horizontal pixels and 4 vertical pixels.

  However, the parent node of the quadtree in the intra CU is PU, and the parent node of the quadtree in the inter CU is CU.

  Similarly, the TU includes a Y signal conversion block (TB) having the same size as the TU, and a Cb signal and a Cr signal TB that are ¼ the size of the TU.

  The variable size CU, PU, and TU described above are selected so that the cost function value described later is minimized. Since the CU, PU, and TU that minimize the cost function value are different according to the pattern of the image information, the pixels of each frame are divided in a state where CU, PU, and TU having different sizes are mixed.

  3 to 5 exemplify the CU, PU, and TU structures employed in HEVC, the CU, PU, and TU structures are not particularly limited. Instead of CU, PU, and TU, a macroblock structure similar to that employed in MPEG2 may be used. In this embodiment, the case where the structure of CU, PU, and TU is adopted is taken as an example.

  As shown in FIG. 6, a frame made up of a plurality of CTUs includes at least one slice. In FIG. 6, a thick solid line indicates a boundary between slices, and the frame is divided into three slices SL1 to SL3. How the frame is divided into a plurality of slices is set by the preprocessing device 50. Each of the slices SL1 to SL3 includes at least one continuous CTU.

  The image encoding device 100 encodes image information in units of slices, and the image decoding device 200 decodes image information in units of slices. In FIG. 6, arrows indicated by solid lines indicate the order of encoding and decoding.

  Returning to FIG. 2, the subtracter 2 uses the prediction value (predicted image) generated by the intra prediction unit 14 or the inter prediction unit 15 described later from the CU of the image information that is the original image output from the rearrangement buffer 1. Subtract to produce a prediction residual. The subtracter 2 supplies the prediction residual to the orthogonal transformation unit 3.

  The orthogonal transform unit 3 performs orthogonal transform on the prediction residual in units of TUs to convert the prediction residual into a frequency domain signal. The orthogonal transform unit 3 uses the discrete sine transform (DST) only when the intra prediction is selected and the TU is 4 pixels horizontal and 4 pixels vertical, and the discrete cosine transform (DCT) is used in other cases. The orthogonal transformation is performed on the prediction residual. The orthogonal transform unit 3 supplies the orthogonal transform coefficient to the quantization unit 4.

  The quantization unit 4 quantizes the orthogonal transform coefficient and supplies it to the entropy coding unit 5 and the inverse quantization unit 8. A specific configuration and operation of the quantization unit 4 will be described later. The entropy encoding unit 5 assigns codes having different lengths to the quantized orthogonal transform coefficients based on the occurrence probabilities, and entropy codes the orthogonal transform coefficients.

  The entropy encoding unit 5 also entropy encodes the syntax for encoding image information. The syntax includes various syntax elements such as information for specifying a prediction mode, a motion vector, and a reference pixel selected by the intra prediction unit 14 or the inter prediction unit 15.

  As an example, the entropy encoding unit 5 can entropy encode orthogonal transform coefficients and syntax using context adaptive arithmetic code (CABAC: Context-based Adaptive Binary Arithmetic Coding).

  The rate control unit 7 controls the rate of the quantization operation in the quantization unit 4 so that the encoded data output from the entropy encoding unit 5 does not overflow or underflow.

  An HRD (Hypothetical Reference Decoder) buffer 6 temporarily accumulates and outputs a bit stream composed of encoded data output from the entropy encoding unit 5.

  The inverse quantization unit 8 inversely quantizes the quantized orthogonal transform coefficient in units of TU and supplies the quantized orthogonal transform coefficient to the inverse orthogonal transform unit 9. The inverse quantization operation in the inverse quantization unit 8 is an operation opposite to the quantization operation in the quantization unit 4.

  The inverse orthogonal transform unit 9 performs inverse orthogonal transform on the input orthogonal transform coefficient in units of TU, and supplies the prediction residual to the adder 10. The adder 10 adds the input prediction residual and the prediction value generated by the intra prediction unit 14 or the inter prediction unit 15 selected by the prediction value selection unit 16 to generate a decoded signal. The decoded signal is supplied to the loop filter 11 and the frame memory 12.

  The loop filter 11 reduces coding noise of the decoded signal. The loop filter 11 includes a deblocking filter that reduces distortion generated at a block boundary and a pixel adaptive offset that reduces ringing distortion. The decoded signal filtered by the loop filter 11 is supplied to the frame memory 12.

  The frame memory 12 stores the decoded signal output from the adder 10 that has not been subjected to the filter processing by the loop filter 11 and the decoded signal that has been subjected to the filter processing by the loop filter 11. The switch 13 supplies the decoded signal stored in the frame memory 12 that has not been subjected to the filter processing to the intra prediction unit 14 and supplies the decoded signal stored in the frame memory 12 to which the filter processing has been performed to the inter prediction unit 15. To do.

  For the Y signal, the intra prediction unit 14 performs, for example, a total of 35 types of prediction in 33 units of directional prediction modes in 33 types of directions, DC (direct current) prediction mode, and planar (Planar) prediction mode described later. Generate predicted values in mode. However, the prediction mode is selected in units of PUs.

  For the Cb and Cr signals, the intra prediction unit 14 uses the vertical prediction modes shown in FIGS. 7A to 7E when the same prediction mode as the Y signal prediction mode is not used in TU units. The prediction value is generated in the horizontal prediction mode, the DC prediction mode, the planar prediction mode, and the CCLM prediction mode. However, the prediction mode is selected in units of PUs.

  In FIGS. 7A to 7D, H is a reference pixel located above a TU that is a target for generating a predicted value, and V is a reference pixel located to the left of the TU. The vertical prediction mode illustrated in FIG. 7A is a mode in which the prediction pixel in the TU is generated in the vertical direction using the reference pixel H. The horizontal prediction mode illustrated in FIG. 7B is a mode in which the prediction pixel in the TU is generated in the horizontal direction using the reference pixel V.

  The DC prediction mode shown in FIG. 7C is a mode for generating a prediction pixel in the TU using an average value of the reference pixels H and V. The planar prediction mode illustrated in FIG. 7D is a mode in which prediction pixels in the TU are generated by interpolation prediction using four reference pixels of the reference pixels H and V. The CCLM prediction mode shown in FIG. 7E is a mode for generating prediction pixels in the TU as follows.

  In the case of a 4: 2: 0 format Y, Cb, Cr video signal, the TB of the Cb and Cr signals and the TB of the Y signal have the relationship shown in FIGS. 8A and 8B. Here, an example in which N is 8 is shown. The peripheral pixels P1 to P16 located above and to the left of the TB of the Cb and Cr signals shown in FIG. 8A are the peripheral pixels located above and to the left of the TB of the Y signal shown in FIG. It corresponds to the peripheral interpolation pixels Pi1 to Pi16 at the illustrated positions interpolated.

The predicted value pred _C (i, j) of the chrominance signal is expressed by equation (1) using a pixel value rec _L (i, j) obtained by shifting the decoded pixel value of the Y signal to the phase of the C signal by a linear model. expressed. (i, j) indicates a pixel position. The intra prediction unit 14 generates a prediction value pred _C (i, j) of the Cb and Cr signals using Expression (1). Α and β in Expression (1) are obtained by, for example, a linear least square method using the peripheral pixels P1 to P16 and the peripheral interpolation pixels Pi1 to Pi16. Therefore, it is not necessary to transmit α and β as a bit stream.

pred _C (i, j) = αrec _L (i, j) + β (1)

Further, the intra prediction unit 14 corrects the prediction value pred _Cr (i, j) of the _Cr signal using the prediction residual resi _Cb ′ (i, j) of the Cb signal based on the equation (2), and corrects the correction. The predicted value pred * _Cr (i, j) of the Cr signal is generated.

pred * _Cr (i, j) = pred _Cr (i, j) + α · resi _Cb ′ (i, j) (2)

  As described above, the intra prediction unit 14 generates a prediction value in a plurality of prediction modes including the CCLM prediction mode. The prediction modes that do not use the decoded pixel value of the Y signal other than the CCLM prediction mode are not limited to the prediction modes illustrated in FIGS. Even when the intra prediction unit 14 generates the prediction values of the Cb and Cr signals using the same prediction mode as the Y signal prediction mode, the prediction mode is not particularly limited.

  The intra prediction unit 14 calculates the cost function value in one determination mode of the High Complexity Mode and the Low Complexity Mode, selects the size of the CU, PU, and TU that minimizes the cost function value, and The prediction value of the prediction mode that minimizes the cost function value is selected.

  High Complexity Mode is a determination mode for obtaining a cost function value Cost_Func based on Expression (3). In Expression (3), D is a difference between the original image and the decoded image, λ is a Lagrange multiplier, and R is a generated code amount including orthogonal transform coefficients. In the High Complexity Mode, it is necessary to calculate the generated code amount by encoding once in all the prediction modes as candidates.

  Cost_Func = D + λR (3)

Low Complexity Mode is a determination mode for obtaining the cost function value Cost_Func based on Expression (4). In Equation (4), SA (T) D is a value obtained by multiplying the difference between the original image and the decoded image by a Hadamard matrix, QP ₀ (QP) is a function of a quantization parameter (QP), and Header_Bit includes an orthogonal transform coefficient There is no code amount. In Low Complexity Mode, prediction images need only be generated in all candidate prediction modes, and there is no need to generate decoded images.

Cost_Func = SA (T) D + QP ₀ (QP) x Header_Bit (4)

  Returning to FIG. 2, the inter prediction unit 15 detects the motion of the image, and sets the CU size as an upper limit to a minimum of horizontal 8 pixels, vertical 4 pixels or horizontal 4 pixels, and vertical 8 pixels from a PU of 64 horizontal pixels. The prediction value is generated by performing inter-frame motion compensation prediction with a PU of 64 vertical pixels. The inter prediction unit 15 generates a predicted value (predicted image) with reference to a past frame, a future frame, or a past and future frame. Each of the past and future frames may be a plurality of frames.

  Similarly, the inter prediction unit 15 calculates the cost function value in one determination mode of the High Complexity Mode and the Low Complexity Mode, and selects the size of the CU, PU, and TU that minimizes the cost function value. In addition, the prediction value of the prediction mode that minimizes the cost function value is selected.

  The prediction value selection unit 16 selects the smaller one of the prediction value selected by the intra prediction unit 14 and the prediction value selected by the inter prediction unit 15 as the final prediction value, and adds the subtractor 2 and the addition. Supply to the vessel 10.

  Note that only the prediction value by the intra prediction unit 14 is used when encoding an I picture. When the P picture and the B picture are encoded, the smaller one of the prediction value selected by the intra prediction unit 14 and the prediction value selected by the inter prediction unit 15 is selected.

  Information indicating the prediction mode selected by the intra prediction unit 14 is supplied to the CCLM prediction mode detection unit 17. When the CCLM prediction mode detection unit 17 detects that the input information indicating the prediction mode indicates the CCLM prediction mode and the intra prediction unit 14 selects the CCLM prediction mode, the quantization unit detects the detection information indicating that fact. 4 and the inverse quantization unit 8. The quantization unit 4 and the inverse quantization unit 8 that have received the detection information are configured to correct the quantization parameter.

  A specific configuration example of the quantization unit 4 and the inverse quantization unit 8 and a method of correcting the quantization parameter will be described with reference to FIG. The quantization unit 4 includes a default quantization parameter holding unit 41, a quantization parameter correction unit 42, and a quantization processing unit 43. The inverse quantization unit 8 includes a default quantization parameter holding unit 81, a quantization parameter correction unit 82, and an inverse quantization processing unit 83.

Quantization parameters for the Y signal and the Cb and Cr signals obtained by the normal decoding process are QP _Y , QP _Cb and QP _Cr , respectively. The default quantization parameter holding unit 41 holds the quantization parameters QP _Cb and QP _Cr as default quantization parameters. The quantization parameter correction unit 42 holds the first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset according to the setting by the preprocessing device 50.

The first offset value cclm_1st_QP_offset indicates an offset value of the quantization parameters QP _Cb and QP _Cr with respect to the quantization parameter QP _Y. The second offset value cclm_2nd_QP_offset indicates the offset value of the quantization parameter QP _{Cr with} respect to the quantization parameter QP _Cb .

When the CCLM prediction mode detection unit 17 detects that the CCLM prediction mode is selected in the intra prediction unit 14, the quantization parameter correction unit 42 obtains Cb and Cr in the CCLM prediction mode as shown in equations (5) and (6). Correct the quantization parameter for the signal. The correction quantization parameters for the Cb and Cr signals in the CCLM prediction mode are QP _{Cb_cclm} and QP _{Cr_cclm} .

QP _{Cb_cclm} = Clip (QP _Cb + cclm_1st_QP_offset, 0, Max_QP_value) (5)
QP _{Cr_cclm} = Clip (QP _Cr + cclm_1st_QP_offset + cclm_2nd_QP_offset, 0, Max_QP_value) (6)

Expression (5) adds the first offset value cclm_1st_QP_offset to the quantization parameter QP _Cb, and limits the added value between the minimum value (here, 0) and the maximum value (Max_QP_value) in the quantization parameter standard. It is shown that the obtained value is set as a quantization parameter QP _{Cb_cclm} .

The expression (6) adds the first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset to the quantization parameter QP _Cr and restricts the added value between the minimum value and the maximum value of the quantization parameter standard. This indicates that the value is the quantization parameter QP _{Cr_cclm} .

Rather than individually setting the offset value of the quantization parameter QP _Cb for the quantization parameter QP _Y and the offset value of the quantization parameter QP _Cr for the quantization parameter QP _Y , the first offset value cclm — 1st_QP_offset and Setting the second offset value cclm_2nd_QP_offset has the following advantages.

Since the quantization parameter QP _Cb and the quantization parameter QP _Cr are approximated, the second offset value cclm_2nd_QP_offset may be zero. Therefore, if the first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset are set as described above, the generated code amount can be reduced.

The quantization parameter correction unit 42 supplies correction quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} to the quantization processing unit 43 when the CCLM prediction mode is selected, and default quantization when a prediction mode other than the CCLM prediction mode is selected. The parameters QP _Cb and QP _Cr are supplied to the quantization processing unit 43.

The quantization processing unit 43 quantizes the input orthogonal transform coefficient with the predetermined quantization parameters QP _Cb and QP _Cr or the corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} and outputs a quantized orthogonal transform coefficient.

The operations of the default quantization parameter holding unit 81 and the quantization parameter correction unit 82 in the inverse quantization unit 8 are the same as the operations of the default quantization parameter holding unit 41 and the quantization parameter correction unit 42, respectively. The inverse quantization processing unit 83 inversely quantizes the quantized orthogonal transform coefficient with the predetermined quantization parameters QP _Cb and QP _Cr or the corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} , and outputs an inverse quantized orthogonal transform coefficient.

  The first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset may be transmitted with a syntax such as a slice header of a bitstream or a picture parameter set (Picture Parameter Set). The first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset need only be transmitted when the CCLM prediction mode is enabled in each slice.

  FIG. 10 shows an example of the syntax generated and transmitted by the entropy encoding unit 5. As shown in FIG. 10, if the CCLM prediction mode enable flag is 1, syntax elements indicating the first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset are included in the syntax and transmitted. Cclm_1st_qp_offset and cclm_2nd_qp_offset in FIG. 10 are syntax elements indicating the first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset, respectively.

  The image encoding apparatus 100 has the configuration shown in FIG. 9, so that when the CCLM prediction mode is selected, the reference Cb and Cr signals are compared with the reference source Cb and Cr signals (referenced Cb and Cr signals). The quantization step of the Cr signal is increased and the quantization value is set high. Therefore, the image coding apparatus 100 can improve the compression rate by reducing the amount of code necessary to compress image information.

<Image coding method>
The process by the image coding method of this embodiment performed with the image coding apparatus 100 is demonstrated using the flowchart shown in FIG. FIG. 11 shows processing executed by the quantization unit 4 and the inverse quantization unit 8.

  In FIG. 11, in step S11, the quantization unit 4 receives the orthogonal transform coefficient of the block in the slice of the frame to be encoded. In step S11, the inverse quantization unit 8 receives the quantized orthogonal transform coefficient of the block in the slice of the frame to be encoded. In step S12, the quantization unit 4 and the inverse quantization unit 8 determine whether the CCLM prediction mode is enabled in the slice of the frame.

  If the CCLM prediction mode is enabled (YES), the quantization unit 4 and the inverse quantization unit 8 determine whether the detection information is supplied from the CCLM prediction mode detection unit 17 in step S13. It is determined whether the CCLM prediction mode is selected in the block.

If the CCLM prediction mode is selected (YES), the quantization unit 4 and the inverse quantization unit 8 add the first offset value cclm — 1st_QP_offset and the second value to the predetermined quantization parameters QP _Cb and QP _Cr in step S14. The offset values cclm_2nd_QP_offset are added to generate corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} .

Subsequently, in step S15, the quantization unit 4 quantizes the orthogonal transform coefficient using the corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} , and shifts the processing to step S17. In step S15, the inverse quantization unit 8 inversely quantizes the quantized orthogonal transform coefficient using the corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} , and shifts the processing to step S17.

  On the other hand, if the CCLM prediction mode is not enabled in step S12 (NO), the quantization unit 4 and the inverse quantization unit 8 shift the processing to step S16. Also, if the CCLM prediction mode is not selected in step S13 (NO), the quantization unit 4 and the inverse quantization unit 8 shift the processing to step S16.

In step S16, the quantization unit 4 quantizes the orthogonal transform coefficient using the predetermined quantization parameters QP _Cb and QP _Cr , and shifts the processing to step S17. In step S16, the inverse quantization unit 8 inversely quantizes the quantized orthogonal transform coefficient using the predetermined quantization parameters QP _Cb and QP _Cr , and shifts the processing to step S17.

  In step S17, the quantization unit 4 and the inverse quantization unit 8 determine whether or not the quantization process and the inverse quantization process have been completed for all the blocks of the slice. If the quantization process and the inverse quantization process have not been completed for all the blocks (NO), the quantization unit 4 and the inverse quantization unit 8 return the process to step S13, and repeat the same process in the next block.

  If the quantization process and the inverse quantization process have been completed for all the blocks (YES), the quantization unit 4 and the inverse quantization unit 8 respectively perform the quantization process for all the slices of the frame in step S18. And it is determined whether or not the inverse quantization process is completed. If the quantization process and the inverse quantization process for all slices are not completed (NO), the quantization unit 4 and the inverse quantization unit 8 return the process to step S12, and repeat the same process in the next slice.

  If the quantization process and the inverse quantization process for all the slices are completed (YES), the quantization unit 4 and the inverse quantization unit 8 respectively perform the quantization process and the inverse quantum for each frame in step S19. It is determined whether or not the digitization process has been completed. If the quantization process and the inverse quantization process for all the frames are not completed (NO), the quantization unit 4 and the inverse quantization unit 8 return the process to step S11 and repeat the same process in the next frame.

  If the quantization process and the inverse quantization process for all the frames have been completed (YES), the quantization unit 4 and the inverse quantization unit 8 end the process.

<Image coding program>
The operation in the image encoding device 100 shown in FIG. 2 can be executed by a computer by a computer program (image encoding program). Each process shown in FIG. 11 can be executed by a computer using an image encoding program.

  In FIG. 12A, the computer 300 includes a central processing unit (CPU) 301 and a storage unit 302. An operation unit 310 is connected to the computer 300. The storage unit 302 stores an image encoding program. The operation unit 310 can function as the pre-processing device 50 shown in FIG. By causing the CPU 301 to execute the image encoding program, the computer 300 can function as the image encoding device 100 that encodes the input image information.

  The storage unit 302 is an arbitrary non-temporary storage medium such as a semiconductor memory, a hard disk drive, or an optical disk. The image encoding program may be provided to the computer 300 via a communication line such as the Internet.

<Image decoding device>
FIG. 13 shows a specific configuration example of the image decoding device 200. In FIG. 13, the HRD buffer 21 temporarily accumulates the bit stream and supplies it to the entropy decoding unit 22. The entropy decoding unit 22 entropy decodes the orthogonal transform coefficients and syntax included in the bitstream.

  The orthogonal transform coefficient is supplied to the inverse quantization unit 23. Information indicating which of intra syntax and inter prediction is adopted among the syntax elements is supplied to the switch 32. Of the syntax elements, information related to intra prediction is supplied to the intra prediction unit 30 and the CCLM prediction mode detection unit 33. Information related to inter prediction among the syntax elements is supplied to the inter prediction unit 31.

  The inverse quantization unit 23 inversely quantizes the orthogonal transform coefficient in units of TU and supplies the inverse transform coefficient to the inverse orthogonal transform unit 24. The inverse orthogonal transform unit 24 performs inverse orthogonal transform on the inversely quantized orthogonal transform coefficient in units of TU, and supplies the prediction residual to the adder 25. The adder 25 adds the input prediction residual and the prediction value generated by the intra prediction unit 30 or the inter prediction unit 31 supplied from the switch 32 to generate a decoded signal. The decoded signal is supplied to the loop filter 26 and the frame memory 28.

  The loop filter 26 has the same configuration as the loop filter 11 and reduces the encoding noise of the decoded signal. The re-order buffer 27 accumulates pixels supplied from the loop filter 26 for a plurality of frames. If the frame order is rearranged, the rearrangement buffer 27 rearranges the frames in the order of the image information of the original image and outputs it as decoded image information. The decoded image information is converted into an analog signal by a D / A converter as necessary.

  The frame memory 28 stores the decoded signal output from the adder 25 that has not been subjected to the filter processing by the loop filter 26 and the decoded signal that has been subjected to the filter processing by the loop filter 26. The switch 29 supplies the decoded signal that has not been subjected to the filter processing accumulated in the frame memory 28 to the intra prediction unit 30, and supplies the decoded signal that has been subjected to the filter processing accumulated in the frame memory 28 to the inter prediction unit 31. To do.

  The intra prediction unit 30 performs intra-frame prediction according to information indicating a prediction mode of intra prediction, and generates respective prediction values of the Y signal, the Cb signal, and the Cr signal. The inter prediction unit 31 performs inter-frame prediction in accordance with information related to inter prediction, and generates respective prediction values for the Y signal and the Cb and Cr signals. The switch 32 supplies the adder 25 with the prediction value generated by the intra prediction unit 30 or the inter prediction unit 31 according to information indicating which of intra prediction and inter prediction is adopted.

  When the CCLM prediction mode detection unit 33 detects that the input information indicating the prediction mode indicates the CCLM prediction mode and the intra prediction unit 14 selects the CCLM prediction mode, the detection information indicating that is dequantized. To the unit 23. The inverse quantization unit 23 that has received the detection information is configured to correct the quantization parameter.

  A specific configuration example of the inverse quantization unit 23 and a method of correcting the quantization parameter will be described with reference to FIG. The inverse quantization unit 23 includes a default quantization parameter holding unit 231, a quantization parameter correction unit 232, and an inverse quantization processing unit 233. The operation of the inverse quantization unit 23 is the same as the operation of the inverse quantization unit 8 in FIG.

The default quantization parameter holding unit 231 holds default quantization parameters QP _Cb and QP _Cr . When the CCLM prediction mode detection unit 33 detects that encoding is performed in the CCLM prediction mode, the quantization parameter correction unit 232 generates corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} according to equations (5) and (6). To do.

When it is detected that the encoding is performed in the CCLM prediction mode, the quantization parameter correction unit 232 supplies the corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} to the inverse quantization processing unit 233, so that the prediction modes other than the CCLM prediction mode When it is detected that the data is encoded by the above, the predetermined quantization parameters QP _Cb and QP _Cr are supplied to the inverse quantization processing unit 233.

The inverse quantization processing unit 233 inversely quantizes the input orthogonal transform coefficient with the predetermined quantization parameters QP _Cb and QP _Cr or the corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} , and outputs an inverse quantization orthogonal transform coefficient.

  The image decoding apparatus 200 has the configuration shown in FIG. 14, whereby an image is encoded in the CCLM prediction mode, and the quantization value of the reference destination Cb and Cr signal is set higher than that of the reference source Cb and Cr signal. By doing so, it is possible to decode a bit stream with an improved compression rate.

<Image decoding method>
The process by the image decoding method of this embodiment performed with the image decoding apparatus 200 is demonstrated using the flowchart shown in FIG. FIG. 15 shows processing executed by the inverse quantization unit 23.

  In FIG. 15, the inverse quantization unit 23 receives the orthogonal transform coefficient of the block in the slice of the frame to be decoded in step S21. In step S22, the inverse quantization unit 23 determines whether the CCLM prediction mode is used in the slice based on whether the enable flag of the CCLM prediction mode is 1.

  If the CCLM prediction mode is used (YES), the inverse quantization unit 23 receives the first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset decoded by the entropy decoding unit 22 in step S23. Based on whether detection information is supplied from the CCLM prediction mode detection unit 33, the inverse quantization unit 23 determines whether the block is encoded in the CCLM prediction mode.

If the block is encoded in the CCLM prediction mode (YES), the inverse quantization unit 23 adds the first offset value cclm_1st_QP_offset and the second offset to the predetermined quantization parameters QP _Cb and QP _Cr in step S25. The value cclm_2nd_QP_offset is added to generate corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} .

Subsequently, in step S26, the inverse quantization unit 23 inversely quantizes the orthogonal transform coefficient using the corrected quantization parameters QP _{Cb_cclm} and QP _{Cr_cclm} , and shifts the processing to step S28.

On the other hand, if the CCLM prediction mode is not used in step S22 (NO), the inverse quantization unit 23 shifts the process to step S27. Moreover, the inverse quantization part 23 will transfer a process to step S27, if the said block is not encoded by CCLM prediction mode in step S24 (NO). In step S27, the inverse quantization unit 23 inversely quantizes the orthogonal transform coefficient using the predetermined quantization parameters QP _Cb and QP _Cr , and shifts the processing to step S28.

  In step S28, the inverse quantization unit 23 determines whether or not the inverse quantization processing of all the blocks of the slice has been completed. If the inverse quantization process has not been completed for all the blocks (NO), the inverse quantization unit 23 returns the process to step S24 and repeats the same process in the next block.

  If the inverse quantization process for all the blocks has been completed (YES), the inverse quantization unit 23 determines whether or not the inverse quantization process for all slices of the frame has been completed in step S29. If the inverse quantization process has not been completed for all slices (NO), the inverse quantization unit 23 returns the process to step S22 and repeats the same process in the next slice.

  If the inverse quantization process for all slices has been completed (YES), the inverse quantization unit 23 determines whether the inverse quantization process for all frames has been completed in step S30. If the inverse quantization process has not been completed for all the frames (NO), the inverse quantization unit 23 returns the process to step S21 and repeats the same process in the next frame. If the inverse quantization process for all the frames has been completed (YES), the inverse quantization unit 23 ends the process.

<Image decoding program>
The operation of the image decoding apparatus 200 shown in FIG. 13 can be executed by a computer by a computer program (image decoding program). Each process shown in FIG. 15 can be executed by a computer using an image decoding program.

  As shown in FIG. 12B, the computer has the same configuration as that in FIG. 12A, and the description of the common parts with FIG. 12A is omitted. In FIG. 12B, the storage unit 302 stores an image decoding program instead of the image encoding program. By causing the CPU 301 to execute the image decoding program, the computer 300 can function as the image decoding device 200 that decodes the encoded bit stream.

  It is also possible to store the image encoding program and the image decoding program in the storage unit 302 so that the computer 300 functions as the image encoding device 100 and the image decoding device 200.

<Modification>
In the image encoding device, the image encoding method, and the image encoding program, and the image decoding device, the image decoding method, and the image decoding program of the present embodiment described above, the following modifications can be made. .

  The first offset value cclm_1st_QP_offset and the second offset value cclm_2nd_QP_offset may be set as a predetermined fixed value without being transmitted in the bit stream.

Different values of the first offset value cclm — 1st_QP_offset and the second offset value cclm — 2nd_QP_offset may be used depending on the value of the quantization parameter for the sub-LCU (subLCU) to which the block to be quantized belongs. A sub-LCU is a maximum coding unit having a size smaller than that of a CTU. For example, when the bit rate is high (the quantization parameter is low), the offset value is not added to the default quantization parameters QP _Cb and QP _Cr , and when the bit rate is low (the quantization parameter is high), the offset value is added. May be.

A threshold of a quantization parameter for selecting whether or not to add an offset value to the predetermined quantization parameters QP _Cb and QP _Cr may be transmitted as a bit stream. The threshold value of the quantization parameter may be transmitted using a slice header or a picture parameter set.

  In the present embodiment described above, Y, Cb, and Cr video signals are taken as examples of image information. However, the color space is not limited to Y, Cb, and Cr, and may be Y, Co, and Cg video signals. The image information may be a video signal in an arbitrary color space including a luminance component and a color component composed of two color difference signals. In this embodiment, the Y, Cb, Cr video signal in the 4: 2: 0 format is taken as an example, but the Y, Cb, Cr video signal is in the 4: 2: 2 format or the 4: 4: 4 format. Also good.

  In the present embodiment described above, the CCLM prediction mode is used for intra prediction, but the CCLM prediction mode may be used for inter prediction. That is, the image coding apparatus 100 predicts a color component without using a first prediction mode in which a prediction value of a color component is generated by a linear model using the decoded pixel value of the luminance component, and a decoded pixel value of the luminance signal. What is necessary is just to provide the predicted value generation part which generates the predicted value of the color component in each block in a picture using either of the 2nd prediction modes which generate a value. The same applies to the image decoding apparatus 200.

  In FIG. 2, the intra prediction unit 14 may be a prediction value generation unit, the inter prediction unit 15 may be a prediction value generation unit, and the intra prediction unit 14 and the inter prediction unit 15 may be prediction value generation units. . In FIG. 13, the intra prediction unit 30 may be a prediction value generation unit, the inter prediction unit 31 may be a prediction value generation unit, and the intra prediction unit 30 and the inter prediction unit 31 may be prediction value generation units. .

  The present invention is not limited to the embodiment described above, and various modifications can be made without departing from the scope of the present invention. In the present embodiment, the CCLM prediction mode for generating the predicted value of the color difference signal using the decoded pixel value of the luminance signal is taken as an example, but the CCLM for generating the predicted value of the Cr signal using the decoded pixel value of the Cb signal. It is also applicable to the prediction mode. That is, the configuration of the present embodiment can be applied if there is a dependency relationship between components of the video signal (luminance signal and color difference signal).

  The image encoding device 100 and the image decoding device 200 may be configured by hardware such as an integrated circuit, may be configured by software, or both may be mixed. The image encoding method and the image decoding method may be executed by any hardware resource such as an integrated circuit or a computer.

2 Subtractor 3 Orthogonal Transformer 4 Quantizer 5 Entropy Encoder 8, 23 Inverse Quantizer 9, 24 Inverse Orthogonal Transformer 10, 25 Adder 14, 30 Intra Predictor (Predicted Value Generator)
15, 31 Inter prediction unit (prediction value generation unit)
22 Entropy Decoding Unit 42, 82, 232 Quantization Parameter Correction Unit 100 Image Encoding Device 200 Image Decoding Device 300 Computer 302 Storage Unit

Claims (10)

A first prediction mode for generating a prediction value of a color component by a linear model using a decoded pixel value of a luminance component, and a second prediction mode for generating a prediction value of a color component without using a decoded pixel value of a luminance signal A prediction value generation unit that generates a prediction value of a color component in each block in the picture using any one of
A subtractor that generates a prediction residual by subtracting the prediction value from a color component of an original image;
An orthogonal transform unit that orthogonally transforms the prediction residual to generate an orthogonal transform coefficient;
When the prediction value generation unit selects the first prediction mode to generate a color component prediction value, the quantization value of the reference color component is set higher than the reference color component, A quantization unit for quantizing the orthogonal transform coefficient;
An image encoding device comprising: The quantization unit includes a quantization parameter correction unit that generates a correction quantization parameter by adding an offset value to a predetermined quantization parameter;
When the prediction value generation unit selects the first prediction mode to generate a color component prediction value, the quantization value of the reference color component is set higher than the reference color component. The image encoding apparatus according to claim 1, wherein the orthogonal transform coefficient is quantized using the corrected quantization parameter.   The image coding apparatus according to claim 2, wherein a syntax including a syntax element indicating the offset value is transmitted in a bitstream. A first prediction mode for generating a prediction value of a color component by a linear model using a decoded pixel value of a luminance component, and a second prediction mode for generating a prediction value of a color component without using a decoded pixel value of a luminance signal Is used to generate a color component prediction value,
Subtract the prediction value from the color component of the original image to generate a prediction residual,
Orthogonal transform the prediction residual to generate orthogonal transform coefficients;
When the first prediction mode is selected to generate a color component prediction value, the quantization value of the reference color component is set higher than the reference color component, and the orthogonal transform coefficient is quantized. An image encoding method characterized by comprising: On the computer,
A first prediction mode for generating a prediction value of a color component by a linear model using a decoded pixel value of a luminance component, and a second prediction mode for generating a prediction value of a color component without using a decoded pixel value of a luminance signal Generating a predicted value of a color component using any of
Subtracting the prediction value from the color component of the original image to generate a prediction residual;
Orthogonally transforming the prediction residual to generate orthogonal transform coefficients;
When the first prediction mode is selected to generate a color component prediction value, the quantization value of the reference color component is set higher than the reference color component, and the orthogonal transform coefficient is quantized. Steps to
An image encoding program characterized in that A first prediction mode for generating a prediction value of a color component by a linear model using a decoded pixel value of a luminance component, and a second prediction mode for generating a prediction value of a color component without using a decoded pixel value of a luminance signal An entropy decoding unit that entropy-decodes a bitstream in which a predicted value of a color component in each block in the picture is generated and encoded, and
In the block in which the prediction value of the color component is generated and encoded using the first prediction mode, the orthogonal transform coefficient entropy-decoded by the entropy decoding unit is compared with the reference color component for reference. An inverse quantization unit configured to inversely quantize by setting the quantization value of the previous color component high;
An image decoding apparatus comprising: The inverse quantization unit includes a quantization parameter correction unit that generates a correction quantization parameter by adding an offset value to a predetermined quantization parameter;
To set a quantization value of a reference color component higher than that of a reference color component in a block in which a prediction value of a color component is generated and encoded using the first prediction mode The image decoding apparatus according to claim 6, wherein the orthogonal transform coefficient is inversely quantized using the corrected quantization parameter. The entropy decoding unit receives a syntax including a syntax element indicating the offset value transmitted in the bitstream;
The image decoding apparatus according to claim 7, wherein the inverse quantization unit generates the corrected quantization parameter by adding the offset value received by the entropy decoding unit to the predetermined quantization parameter. A first prediction mode for generating a prediction value of a color component by a linear model using a decoded pixel value of a luminance component, and a second prediction mode for generating a prediction value of a color component without using a decoded pixel value of a luminance signal And entropy decoding the bitstream in which the predicted value of the color component in each block in the picture is generated and encoded,
In the block in which the prediction value of the color component is generated and encoded using the first prediction mode, the entropy-decoded orthogonal transform coefficient is compared with the color component of the reference source and the color component of the reference destination is compared. An image decoding method, characterized in that a quantization value is set high and inverse quantization is performed. On the computer,
A first prediction mode for generating a prediction value of a color component by a linear model using a decoded pixel value of a luminance component, and a second prediction mode for generating a prediction value of a color component without using a decoded pixel value of a luminance signal And entropy decoding a bitstream in which a predicted value of a color component in each block in the picture is generated and encoded using any of
In the block in which the prediction value of the color component is generated and encoded using the first prediction mode, the entropy-decoded orthogonal transform coefficient is compared with the color component of the reference source and the color component of the reference destination is compared. A step of dequantizing by setting a high quantization value;
An image decoding program characterized in that

Images