JP2008271371A - Moving image encoding apparatus, moving image decoding apparatus, moving image encoding method, moving image decoding method, and program - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To obtain excellent encoding efficiency by efficiently performing encoding while making small a predictive residual between the predictive value of each pixel and the value of a corresponding pixel in an input image. <P>SOLUTION: A moving image encoding apparatus which encodes a moving image while dividing each of images constituting the moving image into a plurality of blocks is characterized in comprising: an intra-prediction unit which calculates predictive values of each pixel constituting a block on the basis of the value of a pixel adjacent to the block and the value of a pixel spaced apart from the block by one or more pixels longitudinally or laterally; a predictive residual calculation unit for calculating a difference between the predictive value of each pixel and the value of a corresponding pixel in an input image; and an encoding unit which encodes the difference to generate an encoded moving image. <P>COPYRIGHT: (C)2009,JPO&INPIT

Description

  The present invention relates to a moving image encoding device, a moving image decoding device, a moving image encoding method, a moving image decoding method, and a program, and in particular, divides an image constituting a moving image into blocks, and sets pixel values of the block to surroundings The present invention relates to a moving image encoding device, a moving image decoding device, a moving image encoding method, a moving image decoding method, and a program that perform prediction based on pixel values.

A conventional moving image encoding method, for example, H.264. In H.264 or MPEG-4 AVC (Advanced Video Coding), an image constituting a moving image is divided into a plurality of blocks each composed of 4 × 4 pixels, 8 × 8 pixels, or 16 × 16 pixels. Intra prediction for predicting the pixel value of the same image from the pixel value of the same image encoded and adjacent to the block, or inter prediction for predicting from the pixel value of the encoded image, or this intra prediction, or The difference between the prediction result of the inter prediction and the actual image is transformed using orthogonal transform obtained by partially modifying the discrete cosine transform, and the result of quantization and entropy coding is used as encoded image data (for example, non-patent) Reference 1 and Non-Patent Document 2).
ITU-T (International Telecommunication Union Telecommunication Standardization Sector) Recommendation H.264 MPEG-4 part 10 Advanced Video Coding, ISO / IEC 14496-10

  However, in the conventional encoding method, when the number of pixels increases as in HDTV (High Definition Television), the amount of information before encoding increases, so the amount of information of encoded image data also increases. There's a problem.

  The present invention has been made in view of such circumstances, and an object of the present invention is to provide a moving picture coding apparatus, a moving picture decoding apparatus, a moving picture coding method, and a moving picture for a moving picture code exhibiting excellent coding efficiency. It is to provide a decoding method and a program.

  The present invention has been made to solve the above-described problems, and a moving image encoding apparatus according to the present invention is a moving image encoding apparatus that divides and encodes an image constituting a moving image into a plurality of blocks. An intra prediction unit that calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block; A prediction residual calculating unit that calculates a difference between a predicted value of each pixel and a value of a corresponding pixel of the input image; and an encoding unit that encodes the difference to generate an encoded moving image. It is characterized by.

  The moving image coding apparatus of the present invention refers not only to the value of the pixel adjacent to the block to be processed, but also to the value of a pixel that is at least one pixel from the block, and based on the value of these pixels, Since the pixel value is predicted, the pixel value can be predicted with high accuracy. For this reason, since the prediction residual, which is the difference between the predicted image and the input image, becomes a small value, these can be efficiently encoded, and excellent encoding efficiency is achieved.

  Also, the moving image encoding device of the present invention is the above-described moving image encoding device, wherein the intra prediction unit is configured to calculate an external value between the adjacent pixel value and the pixel value between the one or more pixels. A predicted value is calculated by insertion.

  Moreover, the moving image encoding device of the present invention is the above-described moving image encoding device, wherein the intra prediction unit calculates an average of the value of the adjacent pixel and the value of the pixel between the one or more pixels. The value is a predicted value.

  The moving picture encoding apparatus of the present invention is any one of the moving picture encoding apparatuses described above, wherein the intra prediction unit includes a value of a pixel adjacent to the block and a vertical or horizontal direction from the block. A predicted value of each pixel constituting the block is calculated on the basis of the value of the pixel between one pixel.

  The moving picture encoding apparatus according to the present invention is any one of the moving picture encoding apparatuses described above, wherein the intra prediction unit selects a pixel from among pixels at least one pixel away from the block. The prediction value of each pixel constituting the block is calculated based on the value of the selected pixel and the value of the pixel adjacent to the block, and the encoding unit adds the difference to the intra prediction unit. The encoded image includes information indicating the position of the pixel between the one or more pixels used in the encoded image.

  The moving picture encoding apparatus of the present invention is any one of the moving picture encoding apparatuses described above, wherein the intra prediction unit determines a predetermined rule among pixels at least one pixel from the block. The prediction value of each pixel constituting the block is calculated based on the value of the pixel selected based on and the value of the pixel adjacent to the block.

  Also, the moving image decoding apparatus of the present invention decodes the encoded moving image in the moving image decoding apparatus that decodes the encoded moving image by dividing the image constituting the moving image into a plurality of blocks. A decoding unit for calculating a prediction residual of each pixel constituting the block, a value of a pixel adjacent to the block, and a value of a pixel at least one pixel in the vertical or horizontal direction from the block An intra prediction unit that calculates a prediction value of each pixel that constitutes the block; and an addition unit that calculates the value of each pixel by adding the prediction residual and the prediction value for each pixel. It is characterized by.

  The moving image decoding apparatus according to the present invention refers not only to the value of the pixel adjacent to the block to be processed, but also to the value of a pixel at least one pixel from the block, and based on the value of these pixels, each pixel Therefore, the pixel value can be predicted with high accuracy. For this reason, since the prediction residual, which is the difference between the predicted image and the input image, becomes a small value, these can be encoded efficiently, and excellent encoding efficiency is achieved.

  The moving picture decoding apparatus according to the present invention is the moving picture decoding apparatus described above, wherein the intra prediction unit performs extrapolation between a value of the adjacent pixel and a value of a pixel between the one or more pixels. A predicted value is calculated.

  The moving picture decoding apparatus according to the present invention is the moving picture decoding apparatus described above, wherein the intra prediction unit calculates an average value of the value of the adjacent pixel and the value of the pixel between the one or more pixels. It is characterized by a predicted value.

  The moving picture decoding apparatus according to the present invention is any one of the moving picture decoding apparatuses described above, wherein the intra prediction unit includes a pixel value adjacent to the block and one pixel in the vertical or horizontal direction from the block. A predicted value of each pixel constituting the block is calculated on the basis of the value of the pixel having a gap.

  The moving picture decoding apparatus according to the present invention is any one of the moving picture decoding apparatuses described above, wherein the decoding unit is used when the predicted value is calculated by the intra prediction unit from the encoded moving picture. Information indicating the position of a pixel between one or more pixels is extracted, and the intra prediction unit performs pixel detection from among the pixels between the one or more pixels according to the information indicating the pixel position extracted by the decoding unit. And a predicted value of each pixel constituting the block is calculated based on the value of the selected pixel and the value of the pixel adjacent to the block.

  The moving picture decoding apparatus according to the present invention is any one of the above-described moving picture decoding apparatuses, wherein the intra prediction unit is based on a predetermined rule among pixels at least one pixel from the block. Based on the value of a selected pixel and the value of a pixel adjacent to the block, a predicted value of each pixel constituting the block is calculated.

  Further, the moving image encoding method of the present invention is a moving image encoding method in a moving image encoding device that divides and encodes an image constituting a moving image into a plurality of blocks. A first step of calculating a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block; A second process in which the moving image encoding device calculates a difference between a predicted value of each pixel and a value of a corresponding pixel of the input image; and the moving image encoding device encodes the difference. And a third step of generating an encoded moving image.

The moving picture decoding method of the present invention is a moving picture decoding method in a moving picture decoding apparatus for decoding a moving picture that is encoded by dividing an image constituting a moving picture into a plurality of blocks.
A first step in which the video decoding device decodes the encoded video and calculates a prediction residual of each pixel constituting the block; and the video decoding device is adjacent to the block. A second step of calculating a predicted value of each pixel constituting the block based on a value of the pixel to be processed and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block; The decoding apparatus includes a third process of calculating a value of each pixel by adding the prediction residual and the prediction value for each pixel.

  In addition, the program of the present invention provides a computer that divides and encodes an image constituting a moving image into a plurality of blocks, and includes a pixel value adjacent to the block and one or more pixels in the vertical or horizontal direction from the block. An intra prediction unit that calculates a predicted value of each pixel that constitutes the block based on the value of the pixel that has the pixel, and a prediction that calculates a difference between the predicted value of each pixel and the value of the corresponding pixel of the input image The residual calculation unit functions as an encoding unit that encodes the difference to generate an encoded moving image.

  Further, the program of the present invention configures the block by decoding a computer that decodes an encoded moving image by dividing an image constituting the moving image into a plurality of blocks, and decodes the encoded moving image. A decoding unit that calculates a prediction residual of each pixel to be configured, and configures the block based on a value of a pixel adjacent to the block and a value of a pixel that is at least one pixel in the vertical or horizontal direction from the block An intra prediction unit that calculates a prediction value of each pixel, and an addition unit that calculates the value of each pixel by adding the prediction residual and the prediction value for each pixel.

  According to the present invention, not only the value of the pixel adjacent to the block to be processed, but also the value of a pixel that is at least one pixel from the block is referred to, and the value of each pixel is determined based on the value of these pixels. Since the prediction is performed, the pixel value can be predicted with high accuracy. For this reason, since the prediction residual, which is the difference between the predicted image and the input image, becomes a small value, these can be efficiently encoded, and excellent encoding efficiency is achieved.

[First Embodiment]
Embodiments of the present invention will be described below with reference to the drawings. FIG. 1 is a schematic block diagram showing the configuration of a moving image encoding apparatus 100 according to the first embodiment of the present invention. Reference numeral 101 denotes an addition unit that takes a difference between an input image and a predicted image and outputs a prediction residual that is a result of the difference. 102 uses an orthogonal transform such as discrete cosine transform to transform the prediction residual to obtain a coefficient, quantize the coefficient with a predetermined quantization width, and output a transform coefficient that is a result of the quantization It is a transform / quantization unit. Reference numeral 103 denotes an encoding unit that encodes the transform coefficient and prediction information described later using an entropy code such as a Huffman code or an arithmetic code, and outputs encoded image data that is a result of the encoding.

  104 performs inverse transformation of quantization and orthogonal transformation performed by the transformation / quantization unit 102 on the transformation coefficient outputted by the transformation / quantization unit 102, and outputs a prediction residual as a result of the inverse transformation This is an inverse quantization / inverse transform unit. Reference numeral 105 denotes an adder that adds the prediction residual output from the inverse quantization / inverse transform unit 104 and the predicted image and outputs a decoded image that is a result of the addition. 106 performs intra prediction on each pixel of the target block using the pixel of the decoded image of the already encoded block as a reference pixel among the images to be encoded. It is an intra prediction part which outputs a certain prediction image. Further, the intra prediction unit 106 outputs prediction information including information indicating the prediction mode used for the intra prediction to the encoding unit 103. Here, the block is a block obtained by dividing an image constituting a moving image every four vertical pixels and every four horizontal pixels.

  Reference numeral 107 denotes a frame memory that stores the decoded image output from the adding unit 105. 108 is stored in the frame memory 107, and a prediction in the time direction is performed with reference to a decoded image of an image encoded before the image to be encoded. It is the inter prediction part which outputs a certain prediction image. In addition, the inter prediction unit 108 outputs information for identifying a decoded image referred to in prediction and prediction information including a motion vector used for prediction to the encoding unit 103. 109, if the image to be encoded is an I picture that does not refer to a decoded image at another time, the prediction image output from the intra prediction unit 106 is output to the addition units 101 and 105, and the other time If it is a P picture or a B picture that refers to the decoded image, it is a switch unit that outputs the prediction image output by the inter prediction unit 108 to the addition units 101 and 105. Note that whether each picture is an I picture, a P picture, or a B picture is selected based on a coding cost, a GOP (Group Of Pictures) structure, or a coding setting by a user instruction. The GOP structure is a structural unit of a moving image including at least one I picture and an arbitrary number of P pictures and B pictures.

  FIG. 2 is a flowchart for explaining an outline of the operation of the moving image encoding apparatus 100. The moving image encoding apparatus 100 performs the following steps Sa2 to Sa10 for each input moving image (Sa1). First, processing is divided depending on whether the target image is encoded using intra prediction or inter prediction (Sa2). In the present embodiment, the processing is described as being divided into either intra prediction or inter prediction. However, both intra prediction and inter prediction are performed, and the results of these processes are compared to determine the encoding efficiency. You may make it choose the better one. When encoding using intra prediction, the switch unit 109 is set to output an input from the intra prediction unit 106, and processing is performed for each block obtained by dividing the image into four vertical pixels and four horizontal pixels. (Sa3). The intra prediction unit 106 generates a predicted image of the processing target block based on the decoded image obtained by decoding the encoded block of the image (Sa4). The adding unit 101 calculates the prediction residual of the processing target block by taking the difference between the predicted image of the processing target block generated by the intra prediction unit 106 and the corresponding block of the input image (Sa5). The transform / quantization unit 102 performs orthogonal transform and quantization on the prediction residual to calculate a transform coefficient (Sa6).

  The inverse quantization / inverse transform unit 104 performs inverse quantization and inverse orthogonal transform, which are inverse operations of the quantization and orthogonal transform performed by the transform / quantization unit 102, and calculates a prediction residual of the processing target block ( Sa7). The adding unit 105 adds the prediction image generated in step Sa4 to the prediction residual to generate a decoded image of the processing target block (Sa8). Here, since the processing of step Sa7 and step Sa8 is the same as the processing performed in the decoding device, this decoded image is an image when the encoded moving image encoded by the moving image encoding device 100 is decoded by the decoding device. Matches. Further, this decoded image is used when a predicted image is generated by the intra prediction unit 106 in step Sa4 and a predicted image is generated by the inter prediction unit 108 in step Sa12 described later. After the decoded image is generated in step Sa8, if there is an unprocessed block in the image to be processed, the processing in steps Sa4 to Sa8 is repeated until returning to step Sa4 and processing all the blocks (Sa9). . After performing the processing of steps Sa4 to Sa8 for all the blocks of the image to be processed, the encoding unit 103 calculates the transform coefficient calculated for each block in step Sa6 and the parameters when performing the intra prediction of step Sa4. Are entropy-coded to generate encoded image data (Sa10). If there is a continuation in the input moving image, the process returns to step Sa1 (Sa11), and the processes of steps Sa2 to Sa10 are repeated for the subsequent images.

  On the other hand, when inter prediction is used in the conditional branch of step Sa2, the inter prediction unit 108 generates a prediction image with reference to a decoded image of an encoded image stored in the frame memory 107, and The adder 101 calculates the difference between the predicted image and the input image, outputs a prediction residual, and the transform / quantization unit 102 performs orthogonal transform and quantization on the predicted residual, and the resulting transform The coefficient is output, and the transform coefficient is inversely quantized and inverse orthogonal transformed by the inverse quantization / inverse transform unit 104. The addition unit 105 adds the transform coefficient to the predicted image to generate a decoded image, and the frame memory 107 (Sa12). And it changes to above-mentioned step Sa10 and continues a process.

  FIG. 3 is a schematic block diagram showing the configuration of the intra prediction unit 106. The moving image encoding apparatus 100 according to the present embodiment operates for each block of 4 pixels in the vertical direction and 4 pixels in the horizontal direction for the I picture among the images constituting the moving image. 106 performs intra prediction for each block. 161 is a first reference pixel group acquisition unit that acquires the value of each pixel of the first reference pixel group, which is a group of pixels adjacent to a block to be subjected to intra prediction, from the decoded image. 162 is a second reference pixel group acquisition that acquires, from the decoded image, the value of each pixel of the second reference pixel group that is a group of pixels that are one pixel apart in the vertical or horizontal direction from the block that is the target of intra prediction. Part. Note that the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 162 indicate that when the block to be subjected to intra prediction is at the upper end or the left end of the image and there are no pixels to be acquired, the pixels are used. Outputs “Not possible”. Further, the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 162 start from the encoded pixels when the pixels to be acquired are not yet encoded (modes 3, 7, and 8 to be described later). Use the obtained value instead.

  Reference numeral 163 denotes intra prediction in the prediction mode specified by the reference pixel values acquired by the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 162 and the prediction information from the intra prediction mode selection unit 164. And a pixel value prediction unit that calculates a predicted value of each pixel of the target block. The intra prediction mode selection unit 164 outputs some temporary prediction information to the pixel value prediction unit 163 to try intra prediction, compares the prediction values obtained by these trials, for example, Prediction information that provides a prediction value with the best coding efficiency is selected, and the switch unit 165 is instructed to output a prediction value based on the prediction information as a prediction image. Also, the intra prediction mode selection unit 164 outputs the selected prediction information to the encoding unit 103.

  The pixel value refers to three values of the luminance value Y, the blue color difference value Cb, and the red color difference value Cr of the pixel. The first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 162 obtains the luminance value Y, the blue color difference value Cb, and the red color difference value Cr from the decoded image, and the pixel value prediction unit 163 obtains the luminance value Y and the blue color difference value Cb of each pixel of the target block. The red color difference value Cr is calculated from the luminance value Y of the reference pixel, the blue color difference value Cb, and the red color difference value Cr, respectively. Here, the luminance value Y and the color difference values Cb and Cr are each represented by 8 bits, the range of values that the luminance value Y can take is 0 to 255, and the values that the blue and red color difference values Cb and Cr can take. The range is -128 to 127.

FIG. 4 is a diagram for defining coordinates for explaining calculation of a prediction value performed by the pixel value prediction unit 163. In the following explanation of intra prediction, as shown in FIG. 3, an x coordinate is taken on the right horizontal axis and a y coordinate is taken on the downward vertical axis. The coordinates of the upper left pixel are the origin. The value of each pixel is represented by p [x, y], such that the value of the upper left pixel of the block is p [0, 0] and the value of the lower right pixel of the block is p [3, 3].
In the following, a pixel that is one pixel in the vertical direction from the block that is the target of intra prediction refers to a pixel whose y coordinate is −2, and one pixel in the horizontal direction from the block that is the target of intra prediction. An interleaved pixel is a pixel whose x coordinate is -2.

  FIG. 5 is a diagram illustrating the direction of intra prediction performed by the pixel value prediction unit 163 in each mode from prediction mode 0 to prediction mode 8 specified by the intra prediction mode selection unit 164 to the pixel value prediction unit 163. As shown in FIG. 5, the prediction mode 0 is from top to bottom, the prediction mode 1 is from left to right, the prediction mode 2 is not oriented, and the average value of the reference pixels is the prediction value of all pixels, and the prediction mode 3 Is from the upper right to the lower left, the prediction mode 4 is from the upper left to the lower right, the prediction mode 5 is from the upper left to the lower right to the lower, the prediction mode 6 is from the upper left to the lower right to the upper, the prediction mode The prediction value calculated from the first and second reference pixels is applied in the direction from the upper right to the lower left from the lower, and the prediction mode 8 from the lower left to the upper right from the lower. A specific calculation method is shown in FIGS.

FIG. 6 is a diagram illustrating a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 0. In the prediction mode 0, the prediction value pred includes the pixels [0, −1] to [3, −1] as the first reference pixels and the pixels [0, −1] as the second reference pixels above the intra prediction target block. -2] to [3, -2] is an average value of pixels having the same x coordinate, and is calculated using the following equation (1) using the reference pixel value p.
pred [x, y] = (p [x, -1] + p [x, -2] +1) / 2 (1)
However, x, y = 0 to 3 and the decimal part is rounded down. Note that “+1” at the end of the parenthesis on the right side of Equation (1) is divided by 2 when there is no “+1” when the decimal point is rounded off when the parenthesis is divided by 2. The number is divided by half to make it the same as rounding off. In the following formulas, when the result of division is rounded down to the nearest decimal point, it is added.
When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equation (1) ′.
pred [x, y] = p [x, −1] (1) ′
However, x, y = 0 to 3 and the decimal part is rounded down.

FIG. 7 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 1. In the prediction mode 1, the prediction value pred includes the pixels [−1, 0] to [−1, 3] that are the first reference pixels and the pixels [−2] that are the second reference pixels on the left side of the intra prediction target block. , 0] to [−2, 3] is an average value of pixels having the same y coordinate, and is calculated using the following equation (2) using the reference pixel value p.
pred [x, y] = (p [-1, y] + p [-2, y] +1) / 2 (2)
However, x, y = 0 to 3 and the decimal part is rounded down.
When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equation (2) ′.
pred [x, y] = p [-1, y] (2) ′
However, x, y = 0 to 3 and the decimal part is rounded down.

FIG. 8 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 2. In the prediction mode 2, the prediction value pred is the pixels [-1, 0] to [-1, 3], [0, -1] to [0] that are the first reference pixels on the left side and the upper side of the intra prediction target block. 3, -1] and the average value of the second reference pixels [-2, 0] to [-2, 3], [0, -2] to [3, -2], and the reference pixel value It is calculated using the following equation (3) using p.
pred [x, y] = (p [-1,0] + p [-1,1] + p [-1,2] + p [-1,3] + p [-2,0] + p [-2,1] + P [-2,2] + p [-2,3] + p [0, -1] + p [1, -1] + p [2, -1] + p [3, -1] + p [0, -2] + p [1, -2] + p [2, -2] + p [3, -2] +8) / 16 (3)
However, x, y = 0 to 3 and the decimal part is rounded down.

When the value of the reference pixel on the upper side cannot be used because the intra prediction target block is at the upper end of the image, it is calculated using the following equation (4).
pred [x, y] = (p [-1,0] + p [-1,1] + p [-1,2] + p [-1,3] + p [-2,0] + p [-2,1] + P [-2,2] + p [-2,3] +4) / 8 (4)
However, x, y = 0 to 3 and the decimal part is rounded down.

Similarly, when the value of the reference pixel on the left side cannot be used because the intra prediction target block is at the left end of the image, it is calculated using the following equation (5).
pred [x, y] = (p [0, -1] + p [1, -1] + p [2, -1] + p [3, -1] + p [0, -2] + p [1, -2] + P [2, -2] + p [3, -2] +4) / 8 (5)
Further, when the values of the left and upper reference pixels cannot be used because the intra prediction target block is at the upper left corner of the image, it is calculated using the following equation (6).
pred [x, y] = 128 (6)
However, x, y = 0 to 3 and the predicted value is the luminance value Y. "0" for color differences Cb and Cr

When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equations (3) ′ to (6) ′.
pred [x, y] = (p [0, -1] + p [1, -1] + p [2, -1] + p [3, -1] + p [-1,0] + p [-1,1] + P [−1,2] + p [−1,3] +4) / 8 (3) ′
When the value of the upper reference pixel cannot be used pred [x, y] = (p [−1,0] + p [−1,1] + p [−1,2] + p [−1,3] +2) / 4 ... (4) '
When the value of the left reference pixel cannot be used pred [x, y] = (p [0, −1] + p [1, −1] + p [2, −1] + p [3, −1] +2) / 4 ... (5) '
When both the upper and left reference pixel values cannot be used,
pred [x, y] = 128 (in the case of luminance), 0 (in the case of color difference) (6) ′
However, the decimal part is rounded down.

FIG. 9 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 3. In the prediction mode 3, the prediction value pred includes the pixels [0, −1] to [7, −1] and the second reference pixels that are the first reference pixels in the upper block and the right block following the intra prediction target block. Among the pixels [1, -2] to [7, -2], which are the weighted average values of the reference pixels in the vicinity of the root in the prediction direction, and the following expression (7) using the reference pixel value p is used. Is calculated.
pred [x, y] = (p [x + y + 1, −2] + 2 × p [x + y + 2, −2] + p [x + y + 3, −2] + p [x + y, −1] + 2 × p [x + y + 1, −1] + p [x + y + 2 , -1] +4) / 8 (7)
However, x, y = 0 to 3 and the decimal part is rounded down.

When [x, y] = [2, 3] or [3, 2], the predicted value pred is calculated using the following equation (8).
pred [x, y] = (p [x + y + 1, −2] + p [x + y + 2, −2] + p [x + y, −1] + 3 × p [x + y + 1, −1] + 2 × p [x + y + 2, −1] +4) / 8 (8)
However, the decimal part is rounded down.
When [x, y] = [3, 3], the predicted value pred is calculated using the following equation (9).
pred [x, y] = (p [6, -1] + 2 × p [7, -1] + p [7, -2] +2) / 4 (9)
However, the decimal part is rounded down.

When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equations (7) ′ and (9) ′.
When [x, y] = [3, 3] pred [x, y] = (p [6, -1] + 3 × p [7, -1] +2) / 4 (7) ′
Otherwise, pred [x, y] = (p [x + y, −1] + 2 × p [x + y + 1, −1] + p [x + y + 2, −1] +2) / 4 (9) ′
However, the decimal part is rounded down.

  FIG. 10 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 4. In the prediction mode 4, the prediction value pred is the pixels [−1, −1] to [3, −1], [−1, 0] to [−1] that are the first reference pixels on the upper and left sides of the intra prediction target block. −1, 3] and the second reference pixel [−2, −2] to [2, −2], [−2, −1] to [−2, 2], near the root of the prediction direction Is calculated by using any one of the following formulas (10) to (12) using the reference pixel value p.

When the x coordinate of the target pixel is larger than the y coordinate (when x> y)
pred [x, y] = (p [x−y−3, −2] + 2 × p [x−y−2, −2] + p [x−y−1, −2] + p [x−y−2] , -1] + 2 × p [xy-1, -1] + p [xy, -1] +4) / 8 (10)
When the x coordinate of the target pixel is smaller than the y coordinate (when x <y)
pred [x, y] = (p [−2, y−x−3] + 2 × p [−2, y−x−2] + p [−2, y−x−1] + p [−1, y−] x-2] + 2 × p [−1, y−x−1] + p [−1, y−x] +4) / 8 (11)
When the x coordinate of the target pixel is the same value as the y coordinate (when x = y)
pred [x, y] = (p [−2, −1] + 2 × p [−2, −2] + p [−1, −2] + p [0, −1] + 2 × p [−1, −1 ] + P [-1,0] +4) / 8 (12)
However, the decimal part is rounded down.

When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equations (10) ′ to (12) ′.
When x> y pred [x, y] = (p [x−y−2, −1] + 2 × p [x−y−1, −1] + p [x−y, −1] +2) / 4 ... (10) '
When x <y pred [x, y] = (p [−1, y−x−2] + 2 × p [−1, y−x−1] + p [−1, y−x] +2) / 4 ... (11) '
When x = y pred [x, y] = (p [0, −1] + 2 × p [−1, −1] + p [−1,0] +2) / 4 (12) ′
However, the decimal part is rounded down.

FIG. 11 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 5. In the prediction mode 5, the prediction values pred are the pixels [-1, -1] to [3, -1] and [-1, 0] to [-1, 0] that are the first reference pixels on the upper and left sides of the intra prediction target block. −1, 2] and the second reference pixel [−2, −2] to [3, −2], [−2, −1] to [−2,0], the vicinity of the root of the prediction direction Is calculated by using any one of the following formulas (13) to (16) using the reference pixel value p.
zVR = 2x−y (where x and y are the x and y coordinates of the target pixel), and when zVR = 0, 2, 4, and 6, pred [x, y] = (p [xy / 2−2, -2] + 2 * p [xy / 2-1, -2] + p [xy / 2, -2] + 2 * p [xy / 2-1, -1] + 2 * p [x- y / 2, -1] +4) / 8 (13)
When zVR = 1, 3, 5 pred [x, y] = (2 × p [xy−2-2, −2] + 2 × p [xy / 2-1, −2] + p [x −y / 2−2, −1] + 2 × p [xy / 2−1, −1] + p [xy−2, −1] +4) / 8 (14)
When zVR = −1 pred [x, y] = (2 × p [−2, −2] + 2 × p [−1, −2] + p [−1,0] + 2 × p [−1, −1] ] + P [0, -1] +4) / 8 (15)
When zVR = −2, -3 pred [x, y] = (3 × p [−2, y−4] + p [−2, y−3] + p [−1, y−3] + 2 × p [ −1, y−2] + p [−1, y−1] +4) / 8 (16)
However, the decimal part is rounded down.

When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equations (13) ′ to (16) ′.
When zVR = 0, 2, 4, 6 pred [x, y] = (p [xy−2-1, −1] + p [xy / 2, −1] +1) / 2... (13) '
When zVR = 1, 3, 5 pred [x, y] = (p [x−y / 2−2, −1] + 2 × p [xy / 2-1, −1] + p [xy] / 2, -1] +2) / 4 (14) '
When zVR = −1 pred [x, y] = (p [−1,0] + 2 × p [−1, −1] + p [0, −1] +2) / 4 (15) ′
When zVR = −2 or −3 pred [x, y] = (p [−1, y−3] + 2 × p [−1, y−2] + p [−1, y−1] +2) / 4 ... (16) '
However, the decimal part is rounded down.

  FIG. 12 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 6. In the prediction mode 6, the prediction values pred are the pixels [-1, -1] to [2, -1] and [-1, 0] to [-1, 0], which are the first reference pixels on the upper and left sides of the intra prediction target block. −1, 3] and the second reference pixel [−2, −2] to [0, −2], [−2, −1] to [−2,3], the vicinity of the root in the prediction direction Is calculated using any one of the following formulas (17) to (20) using the reference pixel value p.

zHD = 2y−x (x and y are the x and y coordinates of the target pixel), and when zHD = 0, 2, 4, and 6, pred [x, y] = (p [−2, y−x / 2) −2] + 2 × p [−2, y−x / 2-1] + p [−2, y−x / 2] + 2 × p [−1, y−x / 2-1] + 2 × p [−1 , Y−x / 2] +4) / 8 (17)
When zHD = 1, 3, 5 pred [x, y] = (2 × p [−2, y−x / 2-2] + 2 × p [−2, y−x / 2-1] + p [− 1, y−x / 2-2] + 2 × p [−1, y−x / 2-1] + p [−1, y−x / 2] +4) / 8 (18)
When zHD = −1 pred [x, y] = (2 × p [−2, −2] + 2 × p [−2, −1] + p [−1,0] + 2 × p [−1, −1 ] + P [0, -1] +4) / 8 (19)
When zHD = −2, -3 pred [x, y] = (3 × p [x−4, −2] + p [x−3, −2] + p [x−3, −1] + 2 × p [ x-2, -1] + p [x-1, -1] +4) / 8 (20)
However, the decimal part is rounded down.

When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equations (17) ′ to (20) ′.
When zHD = 0, 2, 4, 6 pred [x, y] = (p [−1, y−x / 2-1] + p [−1, y−x / 2] +1) / 2... (17) '
When zHD = 1, 3, 5 pred [x, y] = (p [−1, y−x / 2-2] + 2 × p [−1, y−x / 2-1] + p [−1, y−x / 2] +2) / 4 (18) ′
When zHD = −1 pred [x, y] = (p [−1,0] + 2 × p [−1, −1] + p [0, −1] +2) / 4 (19) ′
When zHD = −2 or −3 pred [x, y] = (p [x−3, −1] + 2 × p [x−2, −1] + p [x−1, −1] +2) / 4 ... (20) '
However, the decimal part is rounded down.

FIG. 13 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 7. In the prediction mode 7, the prediction value pred is the pixels [0, −1] to [6, −1] and the second reference pixels that are the first reference pixels in the upper block of the intra prediction target block and the subsequent right block. Of the certain pixels [0, -2] to [6, -2], the weighted average value of the reference pixels in the vicinity of the base in the prediction direction, and the following equations (21) to (22) using the reference pixel value p ).
When y = 0, 2, pred [x, y] = (p [x + y / 2, −2] + 2 × p [x + y / 2 + 1, −2] + p [x + y / 2 + 2, −2] + 2 × p [x + y / 2, −1] + 2 × p [x + y / 2 + 1, −1] +1) / 2 (21)
When y = 1, pred [x, y] = (2 × p [x + y / 2 + 1, −2] + 2 × p [x + y / 2 + 2, −2] + p [x + y / 2, −1] + 2 × p [ x + y / 2 + 1, -1] + p [x + y / 2 + 2, -1] +2) / 4 (22)
However, the decimal part is rounded down.

When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equations (21) ′ and (22) ′.
When y = 0 or 2 pred [x, y] = (p [x + y / 2, −1] + p [x + y / 2 + 1, −1] +1) / 2 (21) ′
Otherwise (y = 1 or 3) pred [x, y] = (p [x + y / 2, -1] + 2 * p [x + y / 2 + 1, -1] + p [x + y / 2 + 2, -1] +2) / 4 (22) '
However, the decimal part is rounded down.

FIG. 14 is a diagram for explaining a prediction value calculation method of the pixel value prediction unit 163 in the prediction mode 8. In the prediction mode 8, the prediction value pred includes the pixels [-1, 0] to [-1, 3] that are the first reference pixels on the left side of the intra prediction target block and the pixels [-2, 2] that are the second reference pixels. 0] to [−2, 3] is a weighted average value of the reference pixels near the root of the prediction direction, and any one of the following expressions (23) to (26) using the reference pixel value p Is calculated using
zHU = x + 2y (x and y are the x and y coordinates of the target pixel), and when zHU = 0, 2, 4 pred [x, y] = (p [−2, y + x / 2] + 2 × p [− 2, y + x / 2 + 1] + p [−2, y + x / 2 + 2] + 2 × p [−1, y + x / 2] + 2 × p [−1, y + x / 2 + 1] +4) / 8 (23)

When zHU = 1, 3, pred [x, y] = (2 × p [−2, y + x / 2 + 1] + p [−2, y + x / 2 + 2] + p [−1, y + x / 2] + 2 × p [−1 , Y + x / 2 + 1] + p [-1, y + x / 2 + 2] +4) / 8 (24)
When zHU = 5,
pred [x, y] = (p [−1,2] + 2 × p [−1,3] + p [−2,3] +2) / 4 (25)
When other than zHU = 0, 1, 2, 3, 4, 5 pred [x, y] = (p [−2, 3] + 7 × p [−1, 3] +4) / 8 (26)
However, the decimal part is rounded down.

When the second reference pixel cannot be referred to and only the first reference pixel can be referred to, the predicted value pred is calculated using the following equations (23) ′ to (26) ′.
When zHU = 0, 2, 4, pred [x, y] = (p [-1, y + x / 2] + p [-1, y + x / 2 + 1] +1) / 2 (23) '
When zHU = 1, 3 pred [x, y] = (p [−1, y + x / 2] + 2 × p [−1, y + x / 2 + 1] + p [−1, y + x / 2 + 2] +2) / 4・ (24) '
When zHU = 5 pred [x, y] = (p [−1,2] + 3 × p [−1,3] +2) / 4 (25) ′
At other times pred [x, y] = p [−1, 3] (26) ′

  FIG. 15 is a flowchart for explaining the operation of the intra prediction unit 106 in the present embodiment. In the intra prediction unit 106, first, the first reference pixel group acquisition unit 161 acquires the value of each pixel of the first reference pixel group from the block that continues to the left, upper, and upper right of the processing target block (Sb1). Next, the second reference pixel group acquisition unit 162 acquires the value of each pixel of the second reference pixel group in the same manner as the first reference pixel group acquisition unit 161 (Sb2). The intra prediction mode selection unit 164 selects one prediction mode and instructs the pixel value prediction unit 163 to generate a prediction image in the prediction mode (Sb3). The pixel value prediction unit 163 generates a prediction image with reference to the first reference pixel and the second reference pixel in the prediction mode specified by the intra prediction mode selection unit 164 in step Sb3 (Sb4). The intra prediction mode selection unit 164 determines whether there is a usable prediction mode other than the prediction mode specified in step Sb3 so far (Sb5). If it is determined that there is, the process returns to step Sb3. The prediction mode is selected from the unselected prediction modes, and the same processing as described above is performed. Here, the determination as to whether or not the prediction mode can be used is used by the first reference pixel group acquisition unit 161 and the second reference image group acquisition unit 162 to calculate a prediction value in the prediction mode. Judgment is made based on whether or not the value of the reference pixel has been acquired. On the other hand, when it is determined in step Sb5 that there is no usable prediction mode, the intra prediction mode selection unit 164 compares the prediction images generated by the pixel value prediction unit 163 in each prediction mode, for example, the most encoded code A prediction image with improved conversion efficiency is selected, the pixel value prediction unit 163 inputs the prediction image to the addition unit 101 via the switch unit 109, and the prediction mode used when generating the prediction image is selected. It outputs to the encoding part 103 as prediction information (Sb6).

  FIG. 16 is a flowchart for explaining the operation of the pixel value prediction unit 163 in the present embodiment. The pixel value prediction unit 163 determines whether or not the first reference pixel depends on whether the first reference pixel group acquisition unit 161 receives the value of the first reference pixel that is referred to in the prediction mode specified by the intra prediction mode selection unit 164. Whether or not can be used is determined (Sc1). If it is determined in step Sc1 that use is not possible, the process proceeds to step Sc5, and the pixel value predicting unit 163 ends the processing assuming that prediction is impossible. On the other hand, if it is determined in step Sc1 that use is possible, the process proceeds to step Sc2, and the pixel value prediction unit 163 sets the value of the second reference pixel to be referred to in the designated prediction mode as the second reference pixel. It is determined whether or not the second reference pixel is usable depending on whether or not it is received from the group acquisition unit 162 (Sc2). If it is determined in step Sc2 that use is not possible, the process proceeds to step Sc4, and calculation is performed when only the first reference pixel described in the prediction value calculation method in each prediction mode of FIGS. According to the method, a predicted pixel value in the designated prediction mode is calculated (Sc4). On the other hand, if it is determined in step Sc2 that use is possible, the process proceeds to step Sc3, and the specified prediction is performed according to the calculation method described in the prediction value calculation method in each prediction mode of FIGS. A predicted pixel value according to the mode is calculated (Sc3).

  As described above, in the moving image encoding apparatus 100 according to the present embodiment, the pixel value prediction unit 163 has not only the value of the pixel adjacent to the processing target block but also one pixel in the vertical or horizontal direction from the block. Since the value of each pixel is predicted using the average value of these pixels with reference to the value of the pixel at an interval, the value of the pixel can be predicted with high accuracy. For this reason, the prediction residual, which is the difference between the predicted image and the input image, becomes a small value. If the information to be encoded is concentrated on a small value, it can be encoded with a small amount of code by arithmetic encoding or variable length encoding, and the encoding efficiency of moving image encoding can be improved.

[Second Embodiment]
FIG. 17 is a schematic block diagram showing the configuration of the moving picture decoding apparatus 200 according to the second embodiment of the present invention. The video decoding device 200 in the present embodiment is a device that decodes the encoded image data generated by the video encoding device 100 in the first embodiment. Reference numeral 201 denotes a decoding unit that decodes entropy coding performed by the coding unit 103 according to the first embodiment on input coded image data, and acquires transform coefficients and prediction information. In the same way as the inverse quantization / inverse transform unit 104 in the first embodiment, 202 is an inverse quantization / inverse transform unit that performs inverse quantization and inverse orthogonal transform to obtain a prediction residual. Reference numeral 203 denotes an adding unit that adds a prediction residual output from the inverse quantization / inverse transform unit 202 and a prediction image generated by an intra prediction unit 204 described later to obtain a decoded image. The intra prediction unit 204 obtains a prediction mode from the prediction information received from the decoding unit 201, and targets a pixel of a decoded image of a block that has already been decoded among the images to be decoded as a reference pixel. Intra prediction is performed in the prediction mode acquired for each pixel of the block, and a prediction image as a result of the intra prediction is output.

  Reference numeral 205 denotes a frame memory that stores the decoded image calculated by the adding unit 203. 206 obtains a prediction mode from the prediction information received from the decoding unit 201, and is stored in the frame memory 205, referring to a decoded image of an image decoded before the image to be decoded. Then, prediction in the time direction is performed in the acquired prediction mode, and a prediction image as a result of the prediction is output. If the image to be decoded is an I picture that does not refer to a decoded image at another time, the predicted image output from the intra prediction unit 204 is output to the adding unit 203, and the decoded image at another time is output. Is a switch unit that outputs the prediction image output from the inter prediction unit 206 to the addition unit 203.

  FIG. 18 is a schematic block diagram illustrating a configuration of the intra prediction unit 204 in the present embodiment. As illustrated in FIG. 18, the intra prediction unit 204 removes the intra prediction mode selection unit 164 and the switch unit 165 from the intra prediction unit 106 illustrated in FIG. 3, and supplies the prediction information from the decoding unit 201 to the pixel value prediction unit 163. The configuration entered. Here, the first reference pixel group acquisition unit 161, the second reference pixel group acquisition unit 162, and the pixel value prediction unit 163 are included in the intra prediction unit 106 illustrated in FIG. 3 and have the same functions as the corresponding units. The pixel value prediction unit 163 that directly receives the prediction information from the decoding unit 201 is included in the prediction information using the reference pixels received from the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 162. A prediction value of each pixel is calculated in the prediction mode.

  As described above, in the moving picture decoding apparatus 200 according to the present embodiment, the pixel value prediction unit 163 includes not only the value of the pixel adjacent to the processing target block but also one pixel in the vertical or horizontal direction from the block. Since the value of each pixel is predicted using the average value of these pixels with reference to the value of the pixel having the above, the value of the pixel can be predicted with high accuracy. This is a prediction method corresponding to the prediction method of the moving image encoding device 100, and the encoded image data generated by the moving image encoding device 100 and having high encoding efficiency can be decoded.

[Third Embodiment]
FIG. 19 is a schematic block diagram showing a configuration of a moving picture coding apparatus 300 according to the third embodiment of the present invention. In the figure, portions (101 to 105, 107 to 109) corresponding to the respective portions in FIG. Similar to the intra prediction unit 106, 306 performs intra prediction on each pixel of the target block using the pixel of the decoded image of the already encoded block as the reference pixel among the images to be encoded. However, it differs from the intra prediction unit 106 in that extrapolation is used to calculate a predicted value.

  FIG. 20 is a schematic block diagram illustrating the configuration of the intra prediction unit 306 according to the present embodiment. In the same figure, parts (161, 162, 164, 165) corresponding to the respective parts in FIG. Similarly to the pixel value prediction unit 163, the reference numeral 363 denotes the reference pixel values acquired by the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 162 and the prediction specified by the intra prediction mode selection unit 164. A pixel value prediction unit that calculates a prediction value of each pixel in the mode, but extrapolation is performed instead of the average value of the reference pixels for calculation of the prediction value in prediction modes 0, 1, and 3 to 8 other than the prediction mode 2. It differs from the pixel value prediction unit 163 in that it is used. The calculation method of the predicted value in each mode other than the prediction mode 2 will be described below.

The pixel value prediction unit 363 calculates a predicted value by extrapolation according to (27) below.
pred [x, y] = p [xa, ya] + (p [xa, ya] −p [xb, yb]) × DR ([x, y], [xa, ya], [xb, yb]) ... (27)
Here, xa, ya, xb, yb are the coordinates of the average pixel obtained from the reference pixel or the reference pixel, and DR ([x1, y1], [x2, y2], [x3, y3]) is "Distance between pixel [x1, y1] and pixel [x2, y2]" / "Distance between pixel [x2, y2] and pixel [x3, y3]", | x1-x2 | / | x2-x3 Or | y1-y2 | / | y2-y3 |.
Xa, ya, xb, yb and p [xa, ya] and p [xb, yb in each prediction mode are shown below. When the second reference pixel cannot be acquired, the second reference pixel is calculated from the first reference pixel by the same calculation method as the pixel value prediction unit 163 in the first embodiment.

Xa, ya, xb, and yb for the pixel [x, y] in the prediction mode 0 are expressed by the following equations (28) and (29).
[Xa, ya] = [x, -1] (28)
[Xb, yb] = [x, -2] (29)
The xa, ya, xb, and yb for the pixel [x, y] in the prediction mode 1 are expressed by the following equations (30) and (31).
[Xa, ya] = [-1, y] (30)
[Xb, yb] = [− 2, y] (31)

Xa, ya, xb, yb and p [xa, ya], p [xb, yb] for the pixel [x, y] in the prediction mode 3 are expressed by the following equations (32) to (39). When [x, y] = [2, 3] or [3, 2] [xa, ya] = [6, −1] (32)
[Xb, yb] = [7, -2] (33)
p [xa, ya] = (p [5, -1] + 2 × p [6, -1] + p [7, -1] +2) / 4 (34)
p [xb, yb] = (p [6, −2] + 3 × p [7, −2] +2) / 4 (35)
In other cases [xa, ya] = [x + y, −1] (36)
[Xb, yb] = [x + y + 1, −2] (37)
p [xa, ya] = (p [x + y, −1] + 2 × p [x + y + 1, −1] + p [x + y + 2, −1] +2) / 4 (38)
p [xb, yb] = (p [x + y + 1, −2] + 2 × p [x + y + 2, −2] + p [x + y + 3, −2] +2) / 4 (39)
However, when [x, y] = [3, 3], since there is no second reference pixel, the predicted value is calculated not by extrapolation but by taking the average value using equation (40).
pred [3,3] = (p [6, -1] + 2 × p [7, -1] p [7, -2] +2) / 4 (40)

Xa, ya, xb, yb and p [xa, ya], p [xb, yb] for the pixel [x, y] in the prediction mode 4 are expressed by the following equations (41) to (52). When x> y [xa, ya] = [x−y−1, −1] (41)
[Xb, yb] = [x−y−2, −2] (42)
p [xa, ya] = (p [x−y−3, −2] + 2 × p [x−y−2, −2] + p [x−y−1, −2] +2) / 4... (43)
p [xb, yb] = (p [x−y−2, −1] + 2 × p [x−y−1, −1] + p [x−y, −1] +2) / 4 (44) )
When x <y [xa, ya] = [− 1, y−x−1] (45)
[Xb, yb] = [− 2, y−x−2] (46)
p [xa, ya] = (p [−1, y−x−2] + 2 × p [−1, y−x−1] + p [−1, y−x] +2) / 4 (47) )
p [xb, yb] = (p [−2, y−x−3] + 2 × p [−2, y−x−2] + p [−2, y−x−1] +2) / 4... (48)
When x = y [xa, ya] = [-1, -1] (49)
[Xb, yb] = [− 2, −2] (50)
p [xa, ya] = (p [0, −1] + 2 × p [−1, −1] + p [−1,0] +2) / 4 (51)
p [xb, yb] = (p [−2, −1] + 2 × p [−2, −2] + p [−1, −2] +2) / 4 (52)

Xa, ya, xb, yb and p [xa, ya], p [xb, yb] for the pixel [x, y] in the prediction mode 5 are expressed by the following equations (53) to (68). zVR = 2x−y.
When zVR = 0, 2, 4, 6 [xa, ya] = [xy / 2-0.5, -1] (53)
[Xb, yb] = [xy / 2-0.5, -2] (54)
p [xa, ya] = (p [xy / 2-1, -1] + p [x−y / 2, −1] +1) / 2 (55) p [xb, yb] = ( p [x−y / 2−2−2] + 2 × p [x−y / 2−1−2] + p [x−y / 2, −2] +2) / 4 (56)
When zVR = 1, 3, 5 [xa, ya] = [xy / 2-1, −1] (57)
[Xb, yb] = [xy / 2-1.5, -2] (58)
p [xa, ya] = (p [xy / 2-2, -1] + 2 * p [xy / 2-1, -1] + p [xy / 2, -1] +2) / 4 (59) p [xb, yb] = (p [x−y / 2−2−2] + p [x−y / 2−1−2] +1) / 2 (60) )

When zVR = −1 [xa, ya] = [− 1, −1] (61)
[Xb, yb] = [− 1.5, −2] (62)
p [xa, ya] = (p [-1,0] + 2 × p [-1, -1] + p [0, -1] +2) / 4 (63)
p [xb, yb] = (p [−2, −2] + p [−1, −2] +1) / 2 (64) When zVR = −2 or −3, [xa, ya] = [-1, y-2] (65)
[Xb, yb] = [− 2, y−3.5] (66)
p [xa, ya] = (p [−1, y−3] + 2 × p [−2, y−2] + p [−1, y−1] +2) / 4 (67) p [xb , Yb] = (3 × p [−2, y−4] + p [−2, y−3] +2) / 4 (68)

Xa, ya, xb, yb and p [xa, ya], p [xb, yb] for the pixel [x, y] in the prediction mode 6 are expressed by the following equations (69) to (84). Let zHD = 2y−x.
When zHD = 0, 2, 4, 6 [xa, ya] = [− 1, y−x / 2−0.5] (69)
[Xb, yb] = [− 2, y−x / 2-1] (70)
p [xa, ya] = (p [−1, y−x / 2−1] + p [−1, y−x / 2] +1) / 2 (71)
p [xb, yb] = (p [−2, y−x / 2-2] + 2 × p [−2, y−x / 2-1] + p [−2, y−x / 2] +2) / 4 ... (72)
When zHD = 1, 3, 5 [xa, ya] = [− 1, y−x / 2-1] (73)
[Xb, yb] = [− 2, y−x / 2−1.5] (74)
p [xa, ya] = (p [−1, y−x / 2-2] + 2 × p [−1, y−x / 2-1] + p [−1, y−x / 2] +2) / 4 (75)
p [xb, yb] = (p [−2, y−x / 2-2] + p [−2, y−x / 2-1] +1) / 2 (76)

When zHD = −1 [xa, ya] = [− 1, −1] (77)
[Xb, yb] = [− 2, −1.5] (78)
p [xa, ya] = (p [-1, 0] + 2 × p [-1, -1] + p [0, -1] +2) / 4 (79)
p [xb, yb] = (p [−2, −2] + p [−2, −1] +1) / 2 (80)
When zHD = -2 or -3 [xa, ya] = [x-2, -1] (81)
[Xb, yb] = [x-3.5, -2] (82)
p [xa, ya] = (p [x−3, −1] + 2 × p [x−2, −1] + p [x−1, −1] +2) / 4 (83)
p [xb, yb] = (3 × p [x−4, −2] + p [x−3, −2] +2) / 4 (84)

Xa, ya, xb, yb and p [xa, ya], p [xb, yb] for the pixel [x, y] in the prediction mode 7 are expressed by the following equations (85) to (92). When y = 0 or 2 [xa, ya] = [x + y / 2 + 0.5, −1] (85)
[Xb, yb] = [x + y / 2 + 1, -2] (86)
p [xa, ya] = (p [x + y / 2, -1] + p [x + y / 2 + 1, -1] +1) / 2 (87)
p [xb, yb] = (p [x + y / 2, −2] + 2 × p [x + y / 2 + 1, −2] + p [x + y / 2 + 2, −2] +2) / 4 (88)
In other cases (y = 1 or 3) [xa, ya] = [x + y / 2 + 1, −1] (89)
[Xb, yb] = [x + y / 2 + 1.5, −2] (90)
p [xa, ya] = (p [x + y / 2, −1] + 2 × p [x + y / 2 + 1, −1] + p [x + y / 2 + 2, −1] +2) / 4 (91)
p [xb, yb] = (p [x + y / 2 + 1, −2] + p [x + y / 2 + 2] +1) / 2 (92)

Xa, ya, xb, yb and p [xa, ya], p [xb, yb] for the pixel [x, y] in the prediction mode 8 are expressed by the following equations (93) to (100). Let zHU = x + 2y.
When zHU = 0, 2, 4 [xa, ya] = [-1, y + x / 2 + 0.5] (93)
[Xb, yb] = [− 2, y + x / 2 + 1] (94)
p [xa, ya] = (p [-1, y + x / 2] + p [-1, y + x / 2 + 1] +1) / 2 (95)
p [xb, yb] = (p [−2, y + x / 2] + 2 × p [−2, y + x / 2 + 1] + p [−2, y + x / 2 + 2] +2) / 4 (96)

When zHU = 1, 3 [xa, ya] = [− 1, y + x / 2 + 1] (97)
[Xb, yb] = [− 2, y + x / 2 + 1.5] (98)
p [xa, ya] = (p [-1, y + x / 2] + 2 * p [-1, y + x / 2 + 1] + p [-1, y + x / 2 + 2] +2) / 4 (99)
p [xb, yb] = (p [−2, y + x / 2 + 1] + p [−2, y + x / 2 + 2] +1) / 2 (100)
In the following cases, since there is no second reference pixel, the predicted value is calculated by taking the average value using Equations (101) and (102) instead of extrapolation.
When zHU = 5 pred [x, y] = (p [−1,2] + 2 × p [−1,3] + p [−2,3] +2) / 4 (101) In other cases pred [X, y] = (p [−2,3] + 7 × p [−1,3] +4) / 8 (102)

  As described above, in the moving image encoding apparatus 300 according to the present embodiment, the pixel value prediction unit 363 has not only the value of the pixel adjacent to the processing target block but also one pixel in the vertical or horizontal direction from the block. The values of the pixels are also referred to, and extrapolation is performed based on the values of these pixels to predict the value of each pixel, so that the value of the pixel can be predicted with high accuracy. For this reason, the prediction residual, which is the difference between the predicted image and the input image, becomes a small value. If the information to be encoded is concentrated on a small value, it can be encoded with a small amount of code by arithmetic encoding or variable length encoding, and the encoding efficiency of moving image encoding can be improved.

[Fourth Embodiment]
FIG. 21 is a schematic block diagram showing the configuration of the moving picture decoding apparatus 400 according to the fourth embodiment of the present invention. The video decoding device 400 in the present embodiment is a device that decodes the encoded image data generated by the video encoding device 300 in the third embodiment. In the figure, parts (201 to 203, 205 to 207) corresponding to the parts in FIG. 404 obtains the prediction mode from the prediction information received from the decoding unit 201, and among the images to be decoded, each pixel of the target block with the pixel of the decoded image of the already decoded block as a reference pixel It is an intra prediction part which performs intra prediction in the prediction mode acquired about the pixel, and outputs the prediction image which is the result of this intra prediction.

  FIG. 22 is a schematic block diagram illustrating a configuration of the intra prediction unit 404 in the present embodiment. As illustrated in FIG. 22, the intra prediction unit 404 removes the intra prediction mode selection unit 164 and the switch unit 165 from the intra prediction unit 306 illustrated in FIG. 20, and supplies the prediction information from the decoding unit 201 to the pixel value prediction unit 363. The configuration entered. Here, the first reference pixel group acquisition unit 161, the second reference pixel group acquisition unit 162, and the pixel value prediction unit 363 are included in the intra prediction unit 306 illustrated in FIG. 20 and have the same functions as the corresponding units. The pixel value prediction unit 363 that directly receives the prediction information from the decoding unit 201 is included in the prediction information using the reference pixels received from the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 162. A prediction value of each pixel is calculated in the prediction mode.

  As described above, in the video decoding device 400 according to the present embodiment, the pixel value prediction unit 363 has not only the value of the pixel adjacent to the processing target block but also one pixel in the vertical or horizontal direction from the block. Since the values of each pixel are also referred to and extrapolated based on the values of these pixels to predict the value of each pixel, the value of the pixel can be accurately predicted. This is a prediction method corresponding to the prediction method of the video encoding device 300, and the encoded image data generated by the video encoding device 300 with high encoding efficiency can be decoded.

[Fifth Embodiment]
FIG. 23 is a schematic block diagram showing a configuration of a moving image encoding apparatus 500 according to the fifth embodiment of the present invention. In the figure, portions (101 to 105, 107 to 109) corresponding to the respective portions in FIG. Similarly to the intra prediction unit 106, 506 performs intra prediction on each pixel of the target block with reference to the pixel of the decoded image of the already encoded block among the images to be encoded. However, it differs from the intra prediction unit 106 in that a pixel selected from pixels between 1 and 3 pixels in the vertical direction or the horizontal direction from the prediction target block is used as the second reference pixel.

  FIG. 24 is a schematic block diagram illustrating a configuration of the intra prediction unit 506 according to the present embodiment. In the figure, parts (161, 165) corresponding to the parts shown in FIG. Reference numeral 562 denotes a second reference pixel group acquisition unit that acquires, from the decoded image, the value of each pixel of the second reference pixel group that is a group of pixels between one pixel and three pixels from the block to be subjected to intra prediction. . Note that the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 562 indicate that when the block to be subjected to intra prediction is at the upper end or the left end of the image and there are no pixels to be acquired, the pixels are used. Outputs “Not possible”. Also, the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 562, when the pixel to be acquired has not been encoded yet (modes 3, 7, and 8 to be described later), Use the obtained value instead.

  Reference numeral 563 denotes a reference mode acquired by the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 562, a prediction mode specified by the prediction information from the intra prediction mode selection unit 564, and the second reference. This is a pixel value prediction unit that performs intra prediction according to the position of a pixel used as a reference pixel among pixels, and calculates a prediction value of each pixel of a target block. The pixel value prediction unit 563 may use the average value of the reference pixels as the prediction value as in the pixel value prediction unit 163, or may calculate a value calculated by extrapolation of the reference pixel as in the pixel value prediction unit 363. A predicted value may be used. The intra prediction mode selection unit 564 outputs some provisional prediction information (the position of the second reference pixel to be referred to as the prediction mode) to the pixel value prediction unit 563 and tries intra prediction. For example, the prediction information of the combination of the prediction mode in which the prediction value with the best coding efficiency is obtained and the position of the second reference pixel to be referenced is selected, and the switch unit 165 selects the prediction information. An instruction to output a prediction value based on prediction information as a prediction image is given. Here, in order to select a prediction mode with good coding efficiency, it is necessary to obtain the generated code amount. For this purpose, the generated code amount obtained by performing temporary encoding using temporary prediction information is used. There are a method of measuring and comparing, and a method of estimating and comparing by an approximate expression using prediction information as an input. Thereby, the prediction mode in which the best predicted image is obtained and the position of the second reference pixel to be referred to are selected. In addition, the intra prediction mode selection unit 564 outputs the selected prediction information (the position of the second reference pixel referred to as the prediction mode) to the encoding unit 103. For this reason, the encoding unit 103 also entropy-codes the position of the reference pixel and includes it in the encoded image data.

For example, the pixel value prediction unit 563 uses the average value as the prediction value, and when the prediction mode 0 and the pixel position n (n is 1 to 3) are designated from the intra prediction mode selection unit 564 intra, The unit 563 calculates the predicted value pred using the following equation (103). Here, the pixel position n is a number indicating how many pixels between the prediction target blocks are used as reference pixels.
pred [x, y] = (p [x, −1] + p [x, −1−n] +1) / 2 (103)
Further, when the prediction mode 1 and the pixel position n are designated, the pixel value prediction unit 563 calculates the prediction value pred by the following equation (104).
pred [x, y] = (p [−1, y] + p [−1−n, y] +1) / 2 (104)

  FIG. 25 is a flowchart for explaining the operation of the pixel value prediction unit 563. The pixel value prediction unit 563 determines whether or not the first reference pixel depends on whether or not the first reference pixel group acquisition unit 161 receives the value of the first reference pixel to be referred to in the prediction mode designated by the intra prediction mode selection unit 564. It is determined whether or not can be used (Sd1). If it is determined in step Sd1 that use is not possible, the process proceeds to step Sd6, and the pixel value prediction unit 563 terminates the processing because it cannot be predicted. On the other hand, when it is determined that it is usable in the determination of step Sd1, the process proceeds to step Sd2, and the pixel value prediction unit 563 is selected from the second reference pixels received from the second reference pixel group acquisition unit 562. A prediction mode and a pixel at a pixel position designated by the prediction information are selected (Sd2). The pixel value prediction unit 563 determines whether or not the second reference pixel can be used based on whether or not the pixel selected in step Sd2 is the pixel received from the second reference pixel group acquisition unit 562 (Sd3). . If it is determined in step Sd3 that use is not possible, the process proceeds to step Sd5, and calculation is performed when only the first reference pixel described in the prediction value calculation method in each prediction mode of FIGS. According to the method, a predicted pixel value in the designated prediction mode is calculated (Sd5). On the other hand, if it is determined in step Sd3 that use is possible, the process proceeds to step Sd4, and the average or extrapolation between the first reference pixel and the designated second reference pixel is performed for the designated prediction mode. The predicted pixel value is calculated (Sd4), and the process is terminated.

  As described above, in the moving picture encoding apparatus 500 according to the present embodiment, the pixel value prediction unit 563 has not only the value of the pixel adjacent to the processing target block but also 1 to 2 in the vertical direction or the horizontal direction from the block. Since the values of the pixels between the three pixels are also referred to, and the average value is calculated or extrapolated based on the values of these pixels to predict the value of each pixel, the pixel value can be accurately predicted. For this reason, the prediction residual which is the difference between the predicted image and the input image is a small value. If the information to be encoded is concentrated on a small value, it can be encoded with a small amount of code by arithmetic encoding or variable length encoding, and the encoding efficiency of moving image encoding can be improved.

  Also, the intra prediction mode selection unit 564 selects a combination of the prediction mode with the highest coding efficiency and the second reference pixel to be referenced from 1 to 3 pixels in the vertical direction or the horizontal direction from all the prediction modes and the processing target block. Since the selection is made from a combination of pixels with a space between them, the encoding efficiency of the moving image encoding can be improved.

[Sixth Embodiment]
FIG. 26 is a schematic block diagram showing a configuration of a video decoding device 600 according to the sixth embodiment of the present invention. The video decoding device 600 in the present embodiment is a device that decodes the encoded image data generated by the video encoding device 500 in the fifth embodiment. In the figure, parts (201 to 203, 205 to 207) corresponding to the parts in FIG. 604 acquires the prediction mode and the pixel position from the prediction information received from the decoding unit 201, and among the blocks already decoded in the image to be decoded, the pixel at the specified pixel position and the decoding target An intra prediction unit that performs intra prediction in the prediction mode acquired for each pixel of a target block using a pixel adjacent to the block being a reference pixel and outputs a prediction image as a result of the intra prediction is there.

  FIG. 27 is a schematic block diagram illustrating the configuration of the intra prediction unit 604 in the present embodiment. As illustrated in FIG. 27, the intra prediction unit 604 removes the intra prediction mode selection unit 564 and the switch unit 165 from the intra prediction unit 506 illustrated in FIG. 24, and supplies the prediction information from the decoding unit 201 to the pixel value prediction unit 563. The configuration entered. Here, the prediction information includes information designating a prediction mode and information indicating the position of a pixel to be referred to from pixels between 1 and 3 pixels in the vertical direction or the horizontal direction from the prediction target block. Also, the first reference pixel group acquisition unit 161, the second reference pixel group acquisition unit 562, and the pixel value prediction unit 563 are included in the intra prediction unit 506 illustrated in FIG. 24 and have the same functions as the corresponding units. The pixel value prediction unit 563 that directly receives the prediction information from the decoding unit 201 is included in the prediction information using the reference pixels received from the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 562. Using the prediction mode, the prediction value of each pixel is calculated with reference to the pixel at the pixel position included in the prediction information.

  As described above, in the moving picture decoding apparatus 600 according to the present embodiment, the pixel value prediction unit 563 has not only the value of the pixel adjacent to the block to be processed but also 1 to 3 in the vertical or horizontal direction from the block. Among the pixels between the pixels, the values of the pixels specified in the prediction information are also referred to, and the average value is calculated or extrapolated based on the values of these pixels to predict the value of each pixel. The value can be predicted with high accuracy. This is a prediction method corresponding to the prediction method of the moving image encoding device 500, and the encoded image data generated by the moving image encoding device 500 with high encoding efficiency can be decoded.

[Seventh Embodiment]
FIG. 28 is a schematic block diagram showing a configuration of a moving image encoding apparatus 700 according to the seventh embodiment of the present invention. In the figure, portions (101 to 105, 107 to 109) corresponding to the respective portions in FIG. Similarly to the intra prediction unit 106, 706 performs intra prediction on each pixel of the target block using the pixel of the decoded image of the already encoded block as a reference pixel among the images to be encoded. However, the intra prediction unit 106 uses a pixel selected based on a predetermined rule among the pixels between the first and third pixels in the vertical direction or the horizontal direction from the prediction target block as the second reference pixel. And different.

  FIG. 29 is a schematic block diagram showing the configuration of the intra prediction unit 706 according to this embodiment. In the figure, parts (161, 165) corresponding to the parts shown in FIG. 762 is a second reference in which the value of each pixel of the second reference pixel group, which is a group of pixels between 1 and 3 pixels in the vertical or horizontal direction from the block to be subjected to intra prediction, is obtained from the decoded image. It is a pixel group acquisition unit. Note that the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 762 indicate that when the block to be subjected to intra prediction is at the upper end or the left end of the image and there are no pixels to be acquired, the pixels are used. Outputs “Not possible”. Also, the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 762, from the encoded pixel, when the pixel to be acquired is not yet encoded (modes 3, 7, and 8 to be described later). Use the obtained value instead.

  763, intra prediction is performed according to the reference pixel values acquired by the first reference pixel group acquisition unit 161 and the second reference pixel group acquisition unit 762 and the prediction mode specified by the prediction information from the intra prediction mode selection unit 164. And a pixel value prediction unit that calculates a predicted value of each pixel of the target block. At this time, the pixel value prediction unit 763 receives pixels between 1 and 3 pixels in the vertical direction or the horizontal direction from the block to be predicted as the second reference pixel, but when calculating the predicted value of each pixel, The median value among these is selected as a pixel to be referred to. Also. The pixel value prediction unit 763 may use the average value of the reference pixels as a prediction value as in the pixel value prediction unit 163, or may calculate a value calculated by extrapolation of the reference pixel as in the pixel value prediction unit 363. A predicted value may be used.

For example, when the pixel value prediction unit 763 uses the average value as the prediction value and the prediction mode 0 is designated from the intra prediction mode selection unit 164 intra, the pixel value prediction unit 763 calculates the following equation (105). To calculate the predicted value pred.
pred [x, y] = (p [x, -1] + CT (p [x, -2], p [x, -3], p [x, -4]) + 1) / 2 (105) )
However, CT (a, b, c) is a function for selecting a median value from a, b, c.
When the prediction mode 1 is designated, the pixel value prediction unit 763 calculates the prediction value pred using the following equation (106).
pred [x, y] = (p [-1, y] + CT (p [-2, x], p [-3, x], p [-4, x]) + 1) / 2 (106) )

  FIG. 30 is a flowchart for explaining the operation of the pixel value prediction unit 763. The pixel value prediction unit 763 determines whether or not the first reference pixel value that is referred to in the prediction mode designated by the intra prediction mode selection unit 164 is received from the first reference pixel group acquisition unit 161. It is determined whether or not can be used (Se1). If it is determined in step Se1 that use is not possible, the process proceeds to step Se6, and the pixel value predicting unit 763 ends the processing assuming that prediction is impossible. On the other hand, when it is determined in step Se <b> 1 that use is possible, the process proceeds to step Se <b> 2, and the pixel value prediction unit 763 determines whether or not the second reference pixel is received from the second reference pixel group acquisition unit 762. Then, it is determined whether or not the second reference pixel can be used (Se2). If it is determined in step Se2 that use is not possible, the process proceeds to step Se5, and calculation is performed when only the first reference pixel described in the prediction value calculation method in each prediction mode of FIGS. According to the method, a predicted pixel value in the designated prediction mode is calculated (Se5). On the other hand, if it is determined in step Se2 that use is possible, the process proceeds to step Se3, and a median pixel is selected from the second reference pixels at the base of the prediction direction of the prediction target pixel ( Se3). The pixel value prediction unit 763 calculates the predicted pixel value by performing the average or extrapolation of the second reference pixel and the first reference pixel selected in Step Se3 for the designated prediction mode (Se4), and ends the process. To do.

  As described above, in the moving picture coding apparatus 700 according to the present embodiment, the pixel value prediction unit 763 has not only the value of the pixel adjacent to the block to be processed, but also 1 to 3 in the vertical direction or the horizontal direction from the block. A pixel selected according to a predetermined rule from among pixels between three pixels is also referred to, and an average value is calculated or extrapolated based on the value of these pixels to predict the value of each pixel. The value can be predicted with high accuracy. For this reason, the prediction residual, which is the difference between the predicted image and the input image, becomes a small value. If the information to be encoded is concentrated on a small value, it can be encoded with a small amount of code by arithmetic encoding or variable length encoding, and the encoding efficiency of moving image encoding can be improved.

[Eighth Embodiment]
FIG. 31 is a schematic block diagram showing the configuration of the moving picture decoding apparatus 800 according to the eighth embodiment of the present invention. A video decoding device 800 in the present embodiment is a device that decodes the encoded image data generated by the video encoding device 700 in the seventh embodiment. In the figure, parts (201 to 203, 205 to 207) corresponding to the parts in FIG. 804 obtains a prediction mode from the prediction information received from the decoding unit 201, and is selected according to a predetermined rule from pixels of a decoded image of a block that has already been decoded among images to be decoded. This is an intra-prediction unit that performs intra prediction in the prediction mode acquired for each pixel of the target block using the obtained pixel as a reference pixel, and outputs a predicted image that is a result of the intra prediction.

  FIG. 32 is a schematic block diagram illustrating a configuration of the intra prediction unit 804 in the present embodiment. As illustrated in FIG. 32, the intra prediction unit 804 removes the intra prediction mode selection unit 164 and the switch unit 165 from the intra prediction unit 706 illustrated in FIG. 29, and sends the prediction information from the decoding unit 201 to the pixel value prediction unit 763. It is the input configuration. Here, the prediction information includes information designating the prediction mode. The first reference pixel group acquisition unit 161, the second reference pixel group acquisition unit 762, and the pixel value prediction unit 763 are included in the intra prediction unit 706 illustrated in FIG. 29 and have the same functions as the corresponding units. The pixel value prediction unit 763 that directly receives the prediction information from the decoding unit 201 receives the pixel received from the first reference pixel group acquisition unit 161 and the median value of the reference pixels received from the second reference pixel group acquisition unit 762 Based on these pixels, the prediction value of each pixel is calculated using the prediction mode included in the prediction information.

  As described above, in the moving picture decoding apparatus 800 according to the present embodiment, the pixel value prediction unit 763 has not only the value of the pixel adjacent to the processing target block but also 1 to 3 in the vertical or horizontal direction from the block. Among the pixels between the pixels, a predetermined rule, that is, a median value pixel is referred to, and an average value is calculated or extrapolated based on the value of these pixels to predict the value of each pixel. Can be accurately predicted. This is a prediction method corresponding to the prediction method of the moving image encoding device 500, and the encoded image data generated by the moving image encoding device 500 with high encoding efficiency can be decoded.

In the first to eighth embodiments, the luminance value Y and the color difference values Cb and Cr are described as the pixel values. However, the pixel values to be used are the red value R, the green value G, and the blue value. B or other color systems such as hue H, lightness V, and saturation C may be used.
In the first to eighth embodiments described above, the inter prediction and the intra prediction are described as being switched in units of images. The intra prediction of the present invention may be applied to a moving picture coding method such as H.264 that switches in units of macroblocks, which are fixed-size blocks obtained by collecting several blocks. At this time, when a block including a reference pixel is encoded using inter prediction or belongs to an SI macroblock and constrained_intra_pred_flag = 1, the reference pixel cannot be used for intra prediction.

In each of the first to eighth embodiments described above, in each prediction mode, the first reference pixel adjacent to the prediction target block and the second reference pixel separated from the block by any one of 1 to 3 pixels. However, the prediction value may be calculated with reference to three or more pixels. For example, in the case of the prediction mode 0, xa, ya, xb, yb, p [xa, ya], and p [xb, yb] used in the equation (27) are changed to the following equations (107) to (110). May be.
[Xa, ya] = [x, −1.5] (107)
[Xb, yb] = [x, −3.5] (108)
p [xa, ya] = p [x, −1.5] = (p [x, −1] + p [x, −2] +1) / 2 (109)
p [xb, yb] = p [x, −3.5] = (p [x, −3] + p [x, −4] +1) / 2 (110)

In the prediction mode 1, xa, ya, xb, yb, p [xa, ya], and p [xb, yb] may be changed to the following formulas (111) to (114).
[Xa, ya] = [− 1.5, y] (111)
[Xb, yb] = [− 3.5, y] (112)
p [−1.5, y] = (p [−1, y] + p [−2, y] +1) / 2 (113)
p [−3.5, y] = (p [−3, y] + p [−4, y] +1) / 2 (114)

  In the first to eighth embodiments described above, the prediction mode has been described as specifying the prediction direction. However, in addition to the prediction direction, a calculation method, for example, an average value or extrapolation is specified. You may do it. In this case, the pixel value prediction unit calculates a predicted value using a calculation formula selected based on the calculation method (average value or extrapolation) designated in the prediction mode and the prediction direction.

  Also, the addition unit 101, transform / quantization unit 102, encoding unit 103, inverse quantization / inverse transform unit 104, addition unit 105, intra prediction unit 106, frame memory 107, inter prediction unit 108, switch unit in FIG. 109, the decoding unit 201 in FIG. 17, the inverse quantization / inverse conversion unit 202, the addition unit 203, the intra prediction unit 204, the frame memory 205, the inter prediction unit 206, the switch unit 207, or the addition unit 101 in FIG. Transformer / quantizer 102, encoder 103, inverse quantizer / inverse transformer 104, adder 105, intra predictor 306, frame memory 107, inter predictor 108, switch 109, or decoder 201 in FIG. , Inverse quantization / inverse transform unit 202, addition unit 203, intra prediction unit 404, frame memo 205, inter prediction unit 206, switch unit 207, or addition unit 101, transform / quantization unit 102, encoding unit 103, inverse quantization / inverse transform unit 104, addition unit 105, intra prediction unit 506 in FIG. , Frame memory 107, inter prediction unit 108, switch unit 109 or decoding unit 201, inverse quantization / inverse transform unit 202, adder unit 203, intra prediction unit 604, frame memory 205, inter prediction unit 206, switch in FIG. Unit 207, or addition unit 101, transform / quantization unit 102, encoding unit 103, inverse quantization / inverse transform unit 104, addition unit 105, intra prediction unit 706, frame memory 107, inter prediction unit 108 in FIG. , Switch unit 109, or decoding unit 201, inverse quantization / inverse transform unit in FIG. 202, an adder 203, an intra-prediction unit 804, a frame memory 205, an inter-prediction unit 206, and a program for realizing the functions of the switch unit 207 are recorded on a computer-readable recording medium and recorded on the recording medium. The program may be read by a computer system and executed to encode or decode a moving image, or may be realized by dedicated hardware. Here, the “computer system” includes an OS and hardware such as peripheral devices.

Further, the “computer system” includes a homepage providing environment (or display environment) if a WWW system is used.
The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory in a computer system serving as a server or a client in that case, and a program that holds a program for a certain period of time are also included. The program may be a program for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in a computer system.

  The embodiment of the present invention has been described in detail with reference to the drawings. However, the specific configuration is not limited to this embodiment, and includes design and the like within a scope not departing from the gist of the present invention.

  The present invention is suitable for use in a moving picture encoding apparatus in a television broadcast and a moving picture decoding apparatus included in a television apparatus that receives a television broadcast and displays a moving picture, but is not limited thereto.

It is a schematic block diagram which shows the structure of the moving image encoder 100 by 1st Embodiment of this invention. It is a flowchart explaining operation | movement of the moving image encoder 100 in the embodiment. It is a schematic block diagram which shows the structure of the intra estimation part 106 in the embodiment. It is a figure explaining the coordinate of the pixel in the embodiment. It is a figure explaining the prediction direction of each prediction mode in the embodiment. It is a figure explaining the predicted value calculation method of prediction mode 0 in the embodiment. It is a figure explaining the predicted value calculation method of prediction mode 1 in the embodiment. It is a figure explaining the predicted value calculation method of prediction mode 2 in the embodiment. It is a figure explaining the predicted value calculation method of prediction mode 3 in the embodiment. It is a figure explaining the predicted value calculation method of prediction mode 4 in the embodiment. It is a figure explaining the prediction value calculation method of prediction mode 5 in the embodiment. It is a figure explaining the predicted value calculation method of prediction mode 6 in the embodiment. It is a figure explaining the predicted value calculation method of prediction mode 7 in the embodiment. It is a figure explaining the prediction value calculation method of prediction mode 8 in the embodiment. It is a flowchart explaining operation | movement of the intra estimation part 106 in the embodiment. It is a flowchart explaining operation | movement of the pixel value estimation part 163 in the embodiment. It is a schematic block diagram which shows the structure of the moving image decoding apparatus 200 by 2nd Embodiment of this invention. It is a schematic block diagram which shows the structure of the intra estimation part 204 in the embodiment. It is a schematic block diagram which shows the structure of the moving image encoder 300 by 3rd Embodiment of this invention. It is a schematic block diagram which shows the structure of the intra estimation part 306 in the embodiment. It is a schematic block diagram which shows the structure of the moving image decoding apparatus 400 by 4th Embodiment of this invention. It is a schematic block diagram which shows the structure of the intra estimation part 404 in the embodiment. It is a schematic block diagram which shows the structure of the moving image encoder 500 by 5th Embodiment of this invention. It is a schematic block diagram which shows the structure of the intra estimation part 506 in the embodiment. It is a flowchart explaining operation | movement of the pixel value estimation part 563 in the same embodiment. It is a schematic block diagram which shows the structure of the moving image decoding apparatus 600 by 6th Embodiment of this invention. It is a schematic block diagram which shows the structure of the intra estimation part 604 in the embodiment. It is a schematic block diagram which shows the structure of the moving image encoder 700 by 7th Embodiment of this invention. It is a schematic block diagram which shows the structure of the intra estimation part 706 in the embodiment. It is a flowchart explaining operation | movement of the pixel value estimation part 763 in the embodiment. It is a schematic block diagram which shows the structure of the moving image decoding apparatus 800 by 8th Embodiment of this invention. It is a schematic block diagram which shows the structure of the intra estimation part 804 in the embodiment.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100, 300, 500, 700 ... Video coding apparatus 101, 105 ... Adder 102 ... Transformer / quantizer 103 ... Encoder 104 ... Inverse quantizer / inverse transformer 106, 306, 506, 706 ... Intra prediction Unit 107 ... frame memory 108 ... inter prediction unit 109 ... switch unit 161 ... first reference pixel group acquisition unit 162, 562, 762 ... second reference pixel group acquisition unit 163, 363, 563, 763 ... pixel value prediction unit 164, 564 ... Intra prediction mode selection unit 165 ... Switch unit 200, 400, 600, 800 ... Video decoding device 201 ... Decoding unit 202 ... Inverse quantization / inverse transform unit 203 ... Adder unit 204, 404, 604, 804 ... Intra prediction Unit 205 ... Frame memory 206 ... Inter prediction unit 207 ... Switch unit

Claims (16)

In a moving image encoding apparatus that encodes an image constituting a moving image by dividing it into a plurality of blocks,
An intra prediction unit that calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block;
A prediction residual calculator that calculates a difference between the predicted value of each pixel and the value of the corresponding pixel of the input image;
And a coding unit configured to code the difference and generate a coded moving image.   The moving image encoding apparatus according to claim 1, wherein the intra prediction unit calculates a prediction value by extrapolating a value of the adjacent pixel and a value of a pixel between one or more pixels. .   The moving image encoding apparatus according to claim 1, wherein the intra prediction unit uses a mean value of a value of the adjacent pixel and a value of a pixel between the one or more pixels as a predicted value.   The intra prediction unit calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one pixel away from the block in a vertical direction or a horizontal direction. The moving picture coding apparatus according to any one of claims 1 to 3, wherein The intra prediction unit selects a pixel from among pixels at least one pixel from the block, and configures each block constituting the block based on the value of the selected pixel and the value of a pixel adjacent to the block. Calculate the predicted value of the pixel,
The encoding unit includes, in addition to the difference, information indicating a pixel position between the one or more pixels used by the intra prediction unit in an encoded image. 4. The moving image encoding apparatus according to any one of items 3 to 4.   The intra prediction unit configures the block based on a value of a pixel selected based on a predetermined rule among pixels at least one pixel from the block and a value of a pixel adjacent to the block. The moving picture encoding apparatus according to claim 1, wherein a predicted value of each pixel is calculated. In a moving image decoding apparatus for decoding a moving image that is encoded by dividing an image constituting a moving image into a plurality of blocks,
A decoding unit that decodes the encoded moving image and calculates a prediction residual of each pixel constituting the block;
An intra prediction unit that calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block;
A moving picture decoding apparatus comprising: an adding unit that calculates the value of each pixel by adding the prediction residual and the prediction value for each pixel.   The moving picture decoding apparatus according to claim 7, wherein the intra prediction unit calculates a predicted value by extrapolating a value of the adjacent pixel and a value of a pixel between the one or more pixels.   The video decoding device according to claim 7, wherein the intra prediction unit uses an average value of a value of the adjacent pixel and a value of a pixel between the one or more pixels as a predicted value.   The intra prediction unit calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one pixel away from the block in a vertical direction or a horizontal direction. The moving picture decoding apparatus according to claim 7, wherein the moving picture decoding apparatus is provided. The decoding unit extracts, from the encoded moving image, information representing pixel positions between the one or more pixels used when calculating a predicted value by the intra prediction unit,
The intra-prediction unit selects a pixel from among the pixels between the one or more pixels according to the information indicating the pixel position extracted by the decoding unit, and the pixel value adjacent to the block and the value of the selected pixel 10. The moving picture decoding apparatus according to claim 7, wherein a predicted value of each pixel constituting the block is calculated based on the value of the video.   The intra prediction unit configures the block based on a value of a pixel selected based on a predetermined rule among pixels at least one pixel from the block and a value of a pixel adjacent to the block. The video decoding device according to claim 7, wherein a predicted value of each pixel is calculated. In a moving image encoding method in a moving image encoding apparatus that divides and encodes an image constituting a moving image into a plurality of blocks,
The moving image encoding device is configured to predict a predicted value of each pixel constituting the block on the basis of a value of a pixel adjacent to the block and a value of a pixel between the block in the vertical direction or the horizontal direction. A first process of calculating
A second process in which the moving image encoding device calculates a difference between the predicted value of each pixel and the value of the corresponding pixel of the input image;
A moving image encoding method comprising: a third process in which the moving image encoding device generates an encoded moving image by encoding the difference. In a moving image decoding method in a moving image decoding apparatus for decoding a moving image encoded by dividing an image constituting a moving image into a plurality of blocks,
A first process in which the video decoding device decodes the encoded video and calculates a prediction residual of each pixel constituting the block;
The moving picture decoding device calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block. A second process of calculating;
The moving picture decoding method comprising: a third step of calculating a value of each pixel by adding the prediction residual and the predicted value for each pixel. A computer that divides and encodes an image constituting a moving image into a plurality of blocks,
An intra prediction unit that calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block;
A prediction residual calculation unit that calculates a difference between the predicted value of each pixel and the value of the corresponding pixel of the input image;
A program that functions as an encoding unit that encodes the difference to generate an encoded moving image. A computer that decodes a moving image encoded by dividing an image constituting a moving image into a plurality of blocks,
A decoding unit that decodes the encoded moving image and calculates a prediction residual of each pixel constituting the block;
An intra prediction unit that calculates a predicted value of each pixel constituting the block based on a value of a pixel adjacent to the block and a value of a pixel that is one or more pixels in the vertical or horizontal direction from the block;
A program causing each pixel to function as an adding unit that calculates the value of each pixel by adding the prediction residual and the prediction value.

Images