JP2011142699A - Image encoding apparatus, and image encoding method - Google Patents

Abstract

<P>PROBLEM TO BE SOLVED: To provide an image encoding apparatus which performs parallel processing of intra-prediction and outputs encoded data which an image decoding apparatus compliant to the H.264 standard can decode. <P>SOLUTION: An image encoding apparatus is configured to perform orthogonal transformation, quantization, dequantization, inverse orthogonal transformation and intra-prediction on each of all the blocks divided from one macro block. The image encoding apparatus includes: a predictive block control section 22 which intra-predicts all the blocks in an order different from a raster scanning order specified by the H.264 standard (b) while using a part of a plurality of modes specified by the H.264 standard to be used for intra-prediction for at least a part of all the blocks (a); and an alignment buffer 23 which outputs all the intra-predicted blocks in accordance with the raster scanning order. <P>COPYRIGHT: (C)2011,JPO&INPIT

Description

  The present invention relates to an image coding apparatus that performs orthogonal transform, quantization, inverse quantization, inverse orthogonal transform, and intra prediction on a block divided from a macroblock.

  As a method for realizing high encoding efficiency nearly twice as high as conventional image encoding methods such as MPEG-2 and MPEG-4. H.264 (also referred to as “MPEG-4 AVC”) has been standardized (for example, see Non-Patent Document 1). H. H.264 is the same as the conventional method in that it is a hybrid image encoding method based on orthogonal transform and motion compensation. However, H.C. In H.264, the degree of freedom of the encoding tool of each element to be encoded is high, and high encoding efficiency is realized by their cumulative effect.

  FIG. 11 is a block diagram showing a configuration of a conventional image encoding device. In FIG. 11, only the components related to intra prediction are shown for ease of explanation. That is, in FIG. 11, a motion prediction unit, a selection unit that selects one of intra prediction and motion prediction, a deblocking filter, and the like are omitted.

  A conventional image coding apparatus includes a block dividing unit 11, a subtracting unit 12, an orthogonal transform unit (T) 13, a quantizing unit (Q) 14, a coding unit 15, and an inverse quantizing unit (iQ). 16, an inverse orthogonal transform unit (iT) 17, an addition unit 18, a frame memory 19, an intra prediction unit (IPD) 20, and a rate control unit 21.

  When each picture of a moving picture composed of continuous pictures (one coding unit including both frames and fields) is a picture in the “4: 2: 0” format, FIG. As shown in (A) to (C), it is composed of one luminance signal (Y signal 31) and two color difference signals (Cr signal 32 and Cb signal 33), and the image size of the color difference signal is It is 1/2 of the luminance signal in both the vertical and horizontal directions.

  Each picture is divided into blocks and encoded in units of blocks. This block is called a macroblock, and is a Y signal block 41 of 16 × 16 pixels shown in FIG. 13A, and a Y signal block 41 and a space shown in FIGS. 13B and 13C, respectively. 8 × 8 pixel Cr signal block 42 and Cb signal block 43 (see, for example, Non-Patent Document 1).

  Each picture is divided into input macroblocks by the block dividing unit 11. The input macroblock is input to the subtraction unit 12. The subtraction unit 12 subtracts the pixel value of the prediction macroblock generated by the intra prediction unit (IPD) 20 from the pixel value of the input macroblock for the pixel at each position, and outputs a difference macroblock. The difference macroblock is input to the orthogonal transformation unit (T) 13, and the orthogonal transformation unit (T) 13 orthogonally transforms the difference macroblock. The size of the block to be orthogonally transformed is 8 × 8 pixels in the MPEG system, but In H.264, 4 × 4 pixels are the basic size.

  The orthogonal transform unit (T) 13 first converts the difference macroblocks into 24 4 × 4 pixel blocks (“51-0” to “51-15”, “52−”) shown in FIGS. 0 ”to“ 52-3 ”, and“ 53-0 ”to“ 53-3 ”), and orthogonal transform is performed for each of them. If the difference macroblock is composed of intra 16 × 16 pixels, which will be described later, the orthogonal transform unit (T) 13 further collects only the DC components of each 4 × 4 orthogonal transform block (“ 51-16 "," 52-4 ", and" 53-4 ") are configured for each signal component and orthogonally transformed. Each transform coefficient in the orthogonal transform block is input to the quantization unit (Q) 14.

  The quantization unit (Q) 14 quantizes the transform coefficient in each orthogonal transform block according to the quantization parameter input from the rate control unit 21. The quantized orthogonal transform coefficient is input to the encoding unit 15 and encoded. H. In H.264, the encoding unit 15 performs variable-length encoding, that is, performs Quantization Conversion by using Context-based Adaptive Variable Length Coding (CAVLC) or Context-based Adaptive Binary Coding (CABAC). .

  The encoding unit 15 encodes the quantized orthogonal transform coefficient as described above, encodes macroblock type information and a prediction mode, which will be described later, further forms a stream, and outputs the result.

  The quantized orthogonal transform coefficient is supplied to the encoding unit 15 and input to the inverse quantization unit (iQ) 16. The inverse quantization unit (iQ) 16 performs inverse quantization on the quantized orthogonal transform coefficient according to the quantization parameter input from the rate control unit 21. As a result, the orthogonal transform block is restored. The restored orthogonal transformation block is subjected to inverse orthogonal transformation by an inverse orthogonal transformation unit (iT) 17 and restored to a differential macroblock. The restored differential macroblock is input to the adder 18 together with the predicted macroblock predicted by the intra prediction unit (IPD) 20.

  The adding unit 18 adds the restored pixel value of the difference macroblock and the pixel value of the prediction macroblock for the pixel at each position to generate a reproduction macroblock. The reproduced macroblock is stored in the frame memory 19 to be used for intra prediction.

  Next, a prediction method and a prediction mode when the intra prediction unit (IPD) 20 generates a prediction macroblock will be described.

  Intra prediction is a method for predicting pixel values in a macroblock using encoded pixels in a frame. H. In the H.264 encoding method, two types of block sizes are prepared as units for performing prediction. They are macroblock types called “intra 4 × 4 prediction” and “intra 16 × 16 prediction”.

  Further, nine types of prediction modes are prepared for the macro block type of intra 4 × 4 prediction, and four types of prediction modes are prepared for the macro block type of intra 16 × 16 prediction, and for each macro block (4 for the intra 4 × 4 prediction). A prediction mode can be selected for each block of × 4 pixels.

  FIG. 15A shows prediction target pixels (16 pixels from “a” to “p”) and pixels (reconstructed after being decoded and reproduced) used for prediction in intra 4 × 4 prediction. It is a figure which shows arrangement | positioning of the adjacent pixel; "A" to "L" 12 pixels). Here, the prediction target pixels (“a” to “p”) are pixels in the encoding target macroblock output from the block dividing unit 11, and the reconstructed adjacent pixels (“A” to “L”) are Macroblocks or block pixels reproduced after decoding and read from the frame memory 19.

  FIG. 15B is a diagram illustrating a prediction direction of intra 4 × 4 prediction. The pixel value of the prediction target pixel is calculated by using the pixel value of the reconstructed adjacent pixel and using a standard arithmetic expression (for example, see Non-Patent Document 1) along the prediction direction. The prediction direction is indicated by a mode number (mode 0 to mode 8). Each of FIG. 15C to FIG. 15K shows a mode number and a prediction direction. The prediction direction is vertical in mode 0 of block 60 in FIG. 15C, the prediction direction is horizontal in mode 1 of block 61 in FIG. 15D, and in mode 2 of block 62 in FIG. The prediction is average (DC). In mode 3 of block 63 in FIG. 15F, the prediction direction is diagonally lower left, in mode 4 of block 64 in FIG. 15G, the prediction direction is diagonally lower right, and block 65 in FIG. In mode 5, the prediction direction is vertical right. In mode 6 of block 66 in FIG. 15 (I), the prediction direction is horizontally downward, in mode 7 of block 67 in FIG. 15 (J), the prediction direction is vertically left, and in block 68 in FIG. 15 (K). In mode 8, the prediction direction is horizontally upward.

  Intra 4 × 4 prediction is applied to the luminance signal. For example, assuming that the predicted value of the pixel is “P”, the predicted value P of the pixel in each mode is as follows. Here, adjacent pixels “A” to “M” shown in FIGS. 15C to 15K used for prediction are reconstructed pixels that have already been reproduced after decoding. However, if the pixels “E” to “H” have not yet been reconfigured, or if they belong to a different slice or frame from the 4 × 4 block, the value of the pixel “D” is equal to the pixels “E” to “E”. Assigned to “H”.

  In mode 0 (vertical), as shown in block 60 of FIG. 15C, when reference pixels “A”, “B”, “C”, and “D” exist, each pixel in block 60 Can be predicted, and each predicted value P is calculated as follows.

a, e, i, m: P = A
b, f, j, n: P = B
c, g, k, o: P = C
d, h, l, p: P = D

  In mode 1 (horizontal), as shown in the block 61 of FIG. 15D, when the reference pixels “I”, “J”, “K”, and “L” exist, each pixel in the block 61 Can be predicted, and each predicted value P is calculated as follows.

a, b, c, d: P = I
e, f, g, h: P = J
i, j, k, l: P = K
m, n, o, p: P = L

  In mode 2 (DC), as shown in block 62 of FIG. 15E, reference pixels “A”, “B”, “C”, “D”, “I”, “J”, “K” , And “L”, the predicted value P of each pixel in the block 62 is as follows.

  P = (A + B + C + D + I + J + K + L + 4) >> 3

  When only the reference pixels “I”, “J”, “K”, and “L” are present, the predicted value P of each pixel in the block 62 is as follows.

  P = (I + J + K + L + 2) >> 2

  When only the reference pixels “A”, “B”, “C”, and “D” exist, the predicted value P of each pixel in the block 62 is as follows.

  P = (A + B + C + D + 2) >> 2

  Further, when none of the reference pixels “A”, “B”, “C”, “D”, “I”, “J”, “K”, and “L” exists, each pixel in the block 62 The predicted value P is as follows.

  P = 128

  In mode 3 (diagonal down-left), as indicated by block 63 in FIG. 15F, reference pixels “A”, “B”, “C”, “D”, “E”, “F”, When “G” and “H” exist, the predicted value P of each pixel in the block 63 is calculated as follows.

a: P = (A + 2B + C + 2) >> 2
b, e: P = (B + 2C + D + 2) >> 2
c, f, i: P = (C + 2D + E + 2) >> 2
d, g, j, m: P = (D + 2E + F + 2) >> 2
h, k, n: P = (E + 2F + G + 2) >> 2
l, o: P = (F + 2G + H + 2) >> 2
p: P = (G + 3H + 2) >> 2

  In mode 4 (diagonal down-right), as indicated by block 64 in FIG. 15G, reference pixels “A”, “B”, “C”, “D”, “I”, “J”, When “K”, “L”, and “M” exist, the predicted value P of each pixel in the block 64 is calculated as follows.

a, f, k, p: P = (A + 2M + I + 2) >> 2
b, g, l: P = (M + 2A + B + 2) >> 2
c, h: P = (A + 2B + C + 2) >> 2
d: P = (B + 2C + D + 2) >> 2
e, j, o: P = (M + 2I + J + 2) >> 2
i, n: P = (I + 2J + K + 2) >> 2
m: P = (J + 2K + L + 2) >> 2

  In mode 5 (vertical-right), as shown in block 65 of FIG. 15H, reference pixels “A”, “B”, “C”, “D”, “I”, “J”, “J” When K ”,“ L ”, and“ M ”exist, the predicted value P of each pixel in the block 65 is calculated as follows.

a, j: P = (M + A + 1) >> 1
b, k: P = (A + B + 1) >> 1
c, l: P = (B + C + 1) >> 1
d: P = (C + D + 1) >> 1
e, n: P = (I + 2M + A + 2) >> 2
f, o: P = (M + 2A + B + 2) >> 2
g, p: P = (A + 2B + C + 2) >> 2
h: P = (B + 2C + D + 2) >> 2
i: P = (J + 2I + M + 2) >> 2
m: P = (K + 2J + I + 2) >> 2

  In mode 6 (horizontal-down), as shown in block 66 of FIG. 15I, reference pixels “A”, “B”, “C”, “D”, “I”, “J”, “J” When K ”,“ L ”, and“ M ”exist, the predicted value P of each pixel in the block 66 is calculated as follows.

a, g: P = (M + I + 1) >> 1
e, k: P = (I + J + 1) >> 1
i, o: P = (J + K + 1) >> 1
m: P = (K + L + 1) >> 1
f, l: P = (M + 2I + J + 2) >> 2
j, p: P = (I + 2J + K + 2) >> 2
n: P = (J + 2K + L + 2) >> 2
b, h: P = (I + 2M + A + 2) >> 2
c: P = (B + 2A + M + 2) >> 2
d: P = (C + 2B + A + 2) >> 2

  In mode 7 (vertical-left), as indicated by block 67 in FIG. 15J, reference pixels “A”, “B”, “C”, “D”, “E”, “F”, “ When G ”and“ H ”exist, the predicted value P of each pixel in the block 67 is calculated as follows.

a: P = (A + B + 1) >> 1
b, i: P = (B + C + 1) >> 1
c, j: P = (C + D + 1) >> 1
d, k: P = (D + E + 1) >> 1
l: P = (E + F + 1) >> 1
e: P = (A + 2B + C + 2) >> 2
f, m: P = (B + 2C + D + 2) >> 2
g, n: P = (C + 2D + E + 2) >> 2
h, o: P = (D + 2E + F + 2) >> 2
p: P = (E + 2F + G + 2) >> 2

  In mode 8 (horizontal-up), when reference pixels “I”, “J”, “K”, and “L” exist, as shown in block 68 of FIG. The predicted value P of each pixel is calculated as follows.

a: P = (I + J + 1) >> 1
e, c: P = (J + K + 1) >> 1
i, g: P = (K + L + 1) >> 1
b: P = (I + 2J + K + 2) >> 2
f, d: P = (J + 2K + L + 2) >> 2
j, h: P = (K + 3L + 2) >> 2
k, l, m, n, o, p: P = L

  As for the luminance signal, as shown in FIGS. 16A to 16D, four prediction modes ((A) mode 0: vertical, (B) mode 1: horizontal, (C) mode 2: average DC, (D) mode 3: plane) are defined in the standard (see, for example, Non-Patent Document 1), and a total of 13 prediction modes including intra 4 × 4 prediction are included. From these, the optimal prediction mode is selected and used for encoding.

  As for the color difference signal, four kinds of prediction modes (prediction modes with prediction directions similar to the intra 16 × 16 prediction of the luminance signal are used for the 8 × 8 pixel block. However, mode 0: DC, mode 1: horizontal, mode 2 : Vertical: mode 3: plane) is defined, and can be encoded independently of the luminance signal.

  By the way, in intra prediction, intra 8 * 8 prediction was added about the luminance signal as Fidelity Range Extension. Intra 8 × 8 prediction has been introduced along with the addition of an 8 × 8 pixel orthogonal transform encoding tool for the purpose of improving the encoding efficiency of high-definition moving images. In intra 8 × 8 prediction, a macroblock is divided into four, each block is smoothed with a 3-tap low-pass filter, and prediction is performed in nine prediction modes as in intra 4 × 4 prediction (see Non-Patent Document 1).

  Note that the prediction block predicted in each mode of each intra prediction type is compared with that of the target block output from the block division unit 11 corresponding to the position and size of the block, and the sum of absolute differences between the blocks, etc. The evaluation value of each prediction block is calculated based on the evaluation function that uses. Based on the calculated evaluation values, the prediction block of the best prediction mode that is estimated to have the smallest code amount is selected, and the prediction block is output to the subtraction unit 12 and the addition unit 18.

  Further, the intra prediction unit (IPD) 20 outputs information on the mode number of the selected prediction mode to the encoding unit 15.

  H. Each block of 4 × 4 pixels included in a macroblock in H.264 encoding is encoded in the zigzag raster scan order indicated by the numbers in the block of FIG. 2 by default (standard). In intra prediction, in order to predict a certain block, it is necessary to encode and decode the image of the surrounding block in advance. For example, in order to intra-predict the block of number 6 in FIG. 2 in all intra prediction modes (9 modes for intra 4 × 4 prediction), number 3 (left), number 1 (upper left), number 4 (upper ) And the reference pixel of the decoded image of the block of number 5 (upper right). That is, in order to intra-predict a certain block, intra prediction (IPD), orthogonal transform (T), quantization (Q), inverse quantization (iQ), and inverse orthogonal transform ( iT) must be completed.

  However, if encoding is performed in the order of the default zigzag raster scan of FIG. 2, the blocks with numbers shaded at points in FIG. 17A, that is, blocks 1, 2, 3, 5, 6, 7, 9, 10, 11, 13, 14, and 15 cannot start intra prediction until decoding of all the blocks in the vicinity is completed. For the blocks of numbers 3, 11, 7, 13 and 15, the upper right block cannot be referred to originally (because the upper right block encoding and decoding processes are later in time), so The rightmost pixel value is allowed to be used as a reference pixel for the upper right block.

  Therefore, as shown in the processing time series of each prediction block in FIG. 17B, IPD (intra prediction processing series) block processing, TQiQiT (orthogonal change (T), quantization (Q), and inverse quantization (IQ) and inverse orthogonal transform (iT) processing sequence) block processing each has a start waiting time. This waiting time hinders parallel processing (pipelining) of the IPD processing system and the TQiQiT processing system. This is a problem of speeding up the H.264 encoding.

  On the other hand, Patent Document 1 discloses the following technique. That is, two or more blocks before and after the prediction block that are located on the left and top of the prediction block used for intra prediction without sequentially processing the blocks in the order of the default zigzag raster scan shown in FIG. Process in this order. This enables pipeline processing.

  For example, in patent document 1, the example shown to FIG. 4 (A) is given as an order of the process of a prediction block. As shown in FIG. 3, when a macroblock composed of 16 pixels × 16 lines is divided into 16 blocks composed of 4 pixels × 4 lines for processing, the position (address) of each block is ( X, Y) {X, Y = 0, 1, 2, 3}, FIG. 4A shows (0, 0), (1, 0), (0, 1), (2, 0), (1,1), (3,0), (2,1), (0,2), (3,1), (1,2), (0,3), (2,2) , (1,3), (3,2), (2,3), and (3,3). However, even if processing is performed in the order shown in FIG. 4A, intra prediction is still started for the six blocks numbered 1, 2, 6, 10, 14, and 15 as shown in FIG. 4B. There is a waiting time. Further, even in other processing orders, if the intra-prediction mode is executed in all intra prediction modes with reference to neighboring blocks according to the rules of the processing order of Patent Document 1, at least six blocks are parallelized (pipe). Line) will hinder processing.

  Therefore, in Patent Document 1, among the default (standard) prediction modes for intra prediction, a prediction mode using reference pixels “E” to “H” at the position of the upper right block shown in FIG. Is changed to a prediction mode in which these reference pixels “E” to “H” are not used, thereby avoiding a hindrance to parallel (pipelining) processing of blocks of numbers 2, 6, 10 and 14. That is, H.I. By using a prediction mode that is not defined by the H.264 standard, an obstacle to parallelization (pipelining) processing is avoided.

  Further, for the remaining blocks numbered 1 and 15, as shown in FIG. 18A, a fixed value (for example, “128”) is used as a predicted value. Alternatively, as shown in FIG. 18B, the pixel value of the decoded image in the block located on the left side of the left adjacent block is used. This avoids an obstacle to parallelization (pipelining) processing.

  FIG. 19 is a block diagram illustrating a configuration of an image encoding device disclosed in Patent Document 1. In FIG. The prediction block control unit 192 performs the H.264 operation on the intra prediction unit (IPD) 20 in the order shown in FIG. Intra prediction is performed using a prediction mode or the like not defined by the H.264 standard.

  In Patent Document 1, the image encoding device is different from the default zigzag raster scan order in the order of H.264. Data is output by sequentially processing blocks using a prediction mode not defined by the H.264 standard. Therefore, the image decoding apparatus includes means for restoring the data of each block in the order of the default zigzag raster scan, and the above-described H.264. And means for decoding data of a block predicted intra by a prediction mode not defined by the H.264 standard.

JP 2004-140473 A

Draft of Version 4 of H. H.264 / AVC (ITU-T Recommendation H.264 and ISO / IEC 14496-10 (MPEG-4 part 10) Advanced Video Coding), Joint Video Team (JVT) of ISO / IEC MPEG T & V N050d1, 2005-01-28

  In Patent Document 1, H.P. In the intra prediction of the H.264 encoding method, the prediction blocks are not sequentially processed in the default order, and the blocks positioned on the left and the top of the prediction block used for the intra prediction are in the order of two or more before the prediction block. In this way, intra prediction is pipelined to speed up the processing. Furthermore, a prediction mode for intra prediction that does not use the reference pixel in the upper right block of the prediction block is set, and for a block that cannot be predicted yet, the reference pixel is set to a fixed value, or a referenceable pixel that is two blocks apart is copied. The method used for prediction is proposed.

  As described above, the image encoding apparatus disclosed in Patent Document 1 processes blocks in an order different from the default standard, uses an intra prediction prediction mode different from the standard, and inserts reference values not allowed by the standard. Etc., the block is encoded. For this reason, the image decoding apparatus requires a separate means for decoding encoded data that deviates from the standard. For this reason, H.264 conforming to the standard. In the H.264 image decoding apparatus, data output from the image encoding apparatus disclosed in Patent Document 1 cannot be decoded.

  The present invention enables parallel (pipelining) processing of intra prediction, and It is an object of the present invention to provide an image encoding device that outputs encoded data that can be decoded by an image decoding device compliant with the H.264 standard.

  In order to solve the above-described problems and achieve the above object, the image coding apparatus of the present invention performs orthogonal transform, quantization, inverse quantization, inverse orthogonal for each of a plurality of blocks obtained by dividing one macroblock. An image encoding apparatus that performs conversion and intra prediction based on a predetermined encoding standard, wherein the macroblock is composed of pixels of 16 pixels × 16 lines, and the block is 4 pixels × 4 lines The block located at the uppermost left of the macroblock is defined with reference to the X axis in the horizontal direction and the Y axis in the vertical direction, and the position of the block is (X, Y) {X, Y = 0, 1, 2, 3}, the intra prediction modes to be applied to all the blocks in the macroblock are modes 0, 1, 2, 4, 5, 6 defined in the predetermined standard. , 8 out of 8 The prediction block control unit set from any one mode and the block are defined as (0, 0), (1, 0), (0, 1), (2, 0), (1, 1), (3 , 0), (2, 1), (0, 2), (3, 1), (1, 2), (0, 3), (2, 2), (1, 3), (3, 2 ), (2, 3), (3, 3) in that order, the prediction block control unit performs intra prediction encoding using the intra prediction mode set in each block, and the intra prediction code. (0,0), (1,0), (0,1), (2,0), (1,1), (3,0), (2,1), (0 , 2), (3, 1), (1, 2), (0, 3), (2, 2), (1, 3), (3, 2), (2, 3), (3, 3 ) In that order, a conversion unit that performs inverse quantization and inverse orthogonal transform, and the image encoding unit and And the transform unit is located in (1, 0) of the blocks and in-plane predictive coding from the block located in (0, 1) to the block located in (2, 3). This is an image coding apparatus that processes inverse quantization and inverse orthogonal transform from a block to a block located at (3, 2) in parallel at the same timing.

  Note that, for example, the present image coding apparatus performs orthogonal transform, quantization, inverse quantization, inverse orthogonal transform, and intra prediction on a predetermined coding standard for each of a plurality of blocks obtained by dividing one macroblock. A prediction block control unit that sets an intra prediction mode to be applied to a block to be encoded among a plurality of blocks from an intra prediction mode defined in the encoding standard; and An intra prediction unit that intra-predicts a block to be encoded based on an intra prediction mode set by a prediction block control unit, wherein the prediction block control unit conforms to the encoding standard This is an intra prediction mode that is specified, and the pixel data that has been inversely quantized and inversely orthogonal transformed at the time of encoding the block to be encoded. One intra prediction mode included in an intra prediction mode (hereinafter referred to as an intra prediction mode group) that uses the intra prediction mode is set, and the intra prediction unit includes the intra prediction mode group set by the prediction block control unit. An image in which a plurality of blocks obtained by dividing a macroblock are encoded in an encoding order in which the number of modes is equal to or greater than the number of intra prediction modes included in the intra prediction mode group when encoding is performed in the order of raster scanning. An encoding device may be used.

  The present image encoding apparatus is an apparatus that performs orthogonal transform, quantization, inverse quantization, inverse orthogonal transform, and intra prediction for each of all blocks divided into a plurality of macro blocks. (A) At least some of all the blocks are used for intra prediction. Using a part of a plurality of modes defined by the H.264 standard, H. The control means for intra-predicting all the blocks in an order different from the order of raster scanning defined by the H.264 standard, and all the blocks intra-predicted based on the control by the control means. Output means for outputting in accordance with the order.

  For example, the control means includes intra prediction for a second block that is a block on the left side of the first block that is one of the blocks, and a block that is an upper block of the first block. The intra prediction for the third block is performed in a processing order two or more times before the intra prediction for the first block.

  Further, for example, the macro block is configured by pixels of 16 pixels × 16 lines, the block is configured by pixels of 4 pixels × 4 lines, and the positions of the blocks in the macro block are When expressed as (X, Y) {X, Y = 0, 1, 2, 3} with reference to the left and top of the macroblock, the control means (A) position (0, 1), For the blocks (2, 1), (0, 3), and (2, 3), the H. Intra prediction is performed in modes other than mode 3 and mode 7 of modes 0 to 8 defined by the H.264 standard, and (B) the blocks at positions (1, 0) and (3, 3) H. (C) position (0,0), (2,0), (3,0), (1,1), (3 , 1), (0, 2), (1, 2), (2, 2), and (1, 3), the H. Intra prediction is performed from mode 0 to mode 8 defined by the H.264 standard, and (D) the 16 blocks are assigned positions (0, 0), (1, 0), (0, 1), (2, 0 ), (1,1), (3,0), (2,1), (0,2), (3,1), (1,2), (0,3), (2,2), Intra prediction is performed in the order of (1, 3), (3, 2), (2, 3), and (3, 3).

  Further, for example, the macro block is composed of pixels of 16 pixels × 16 lines, the block is composed of pixels of 4 pixels × 4 lines, and the control means includes (A) all the blocks For H. Intra prediction is performed in modes other than mode 3 and mode 7 of mode 0 to mode 8 defined by the H.264 standard, and (B) 16 blocks are located at positions (0, 0), (1, 0). , (0,1), (2,0), (1,1), (3,0), (2,1), (0,2), (3,1), (1,2), ( Intra prediction is performed in the order of 0,3), (2,2), (1,3), (3,2), (2,3), and (3,3).

  An image encoding apparatus of the present invention is an apparatus that performs orthogonal transform, quantization, inverse quantization, inverse orthogonal transform, and intra prediction for each of all blocks divided into a plurality of macroblocks. (A) H. is used for intra prediction for all the blocks. Using a part of a plurality of modes defined by the H.264 standard, And a control unit for intra-predicting all the blocks in the order of raster scanning defined by the H.264 standard.

  For example, the macro block is composed of pixels of 16 pixels × 16 lines, the block is composed of pixels of 4 pixels × 4 lines, and the position of each block in the macro block is the macro block. When expressed as (X, Y) {X, Y = 0, 1, 2, 3} with reference to the left and top of the block, the control means (A) position (0, 1), ( 2, 1), (0, 3), and (2, 3), the H. (B) position (1, 0), (3, 0), (1, 1), and intra prediction is performed in modes other than mode 3 and mode 7 of mode 0 to mode 8 defined by the H.264 standard. For the blocks (3, 1), (1, 2), (3, 2), (1, 3), and (3, 3), the H. Intra prediction is performed by mode 0, mode 3 and mode 7 defined by the H.264 standard, and (C) the positions (0, 0), (2, 0), (0, 2), and (2, 2) Regarding the block, the above-mentioned H.P. Intra prediction is performed from mode 0 to mode 8 defined by the H.264 standard, and (D) 16 blocks are assigned to positions (0, 0), (1, 0), (0, 1), (1, 1 ), (2,0), (3,0), (2,1), (3,1), (0,2), (1,2), (0,3), (1,3), Intra prediction is performed in the order of (2, 2), (3, 2), (2, 3), and (3, 3).

  The present invention can be realized as an image encoding method including steps as characteristic constituent means of the image encoding device of the present invention, or realized as a program for causing a computer to execute these steps, or the characteristic constituent means described above. It can also be realized as an integrated circuit including. The above program can also be distributed via a recording medium such as a CD-ROM or a transmission medium such as a communication network.

  The present invention enables parallel (pipelining) processing of intra prediction, and It is possible to provide an image encoding device that outputs encoded data that can be decoded by an image decoding device compliant with the H.264 standard.

  That is, according to the present invention, the intra prediction process can be speeded up.

It is a block diagram which shows the structure of the image coding apparatus of embodiment. H. It is a figure for demonstrating the order which encodes the block in a macroblock in H.264 image coding. It is a figure for demonstrating the case where the positional information on the block of intra 4x4 prediction within a macroblock is represented by a coordinate (X, Y). It is a figure for demonstrating the order of the process of the block in a macroblock, etc. FIG. It is a figure for demonstrating the method of extracting and outputting the output data at the time of encoding in the default order of the standard at the time of encoding in the order in a non-standard in the block in a macroblock by the image encoding in embodiment. It is a figure for demonstrating the process conditions etc. of the 1st intra prediction in embodiment. It is a flowchart for demonstrating the process of the 1st intra prediction in embodiment. It is a figure for demonstrating the processing conditions of the 2nd intra prediction in embodiment, etc. FIG. It is a flowchart for demonstrating the process of the 2nd intra prediction in embodiment. It is a figure for demonstrating the process conditions of the 3rd intra prediction in embodiment, etc. FIG. H. 1 is a block diagram illustrating a partial configuration of an image encoding device according to the H.264 standard. It is a figure for demonstrating the video signal of 1 picture of 4: 2: 0 format. It is a figure for demonstrating the video signal of 1 macroblock of 4: 2: 0 format. It is a figure for demonstrating the orthogonal transformation block of 1 macroblock of 4: 2: 0 format. H. It is a figure for demonstrating the intra 4x4 prediction of the luminance signal in H.264 image coding. H. It is a figure for demonstrating the intra 16x16 prediction of the luminance signal in H.264 image coding. H. The figure for demonstrating the order which processes the block in the macroblock which carries out intra prediction in H.264 image coding, and the figure for demonstrating the block which blocks pipelining in the case of intra prediction, and its process time series is there. In conventional image coding, it is a figure for demonstrating substituting the pixel value which cannot be referred at the time of intra prediction with a fixed value, or copying from a block two blocks away. FIG. 10 is a block diagram illustrating a configuration of an image encoding device disclosed in Patent Document 1.

  The best mode for carrying out the present invention will be described below with reference to the drawings.

  FIG. 1 is a block diagram illustrating a configuration of an image encoding device according to an embodiment. In FIG. 1, only the components related to intra prediction are shown for ease of explanation. That is, in FIG. 1, a motion prediction unit, a selection unit that selects one of intra prediction and motion prediction, a deblocking filter, and the like are omitted.

  As shown in FIG. 1, the image encoding device according to the embodiment includes a block dividing unit 11, a subtracting unit 12, an orthogonal transform unit (T) 13, a quantizing unit (Q) 14, and an encoding unit 15. An inverse quantization unit (iQ) 16, an inverse orthogonal transform unit (iT) 17, an addition unit 18, a frame memory 19, an intra prediction unit (IPD) 20, a rate control unit 21, and a prediction block control A unit 22 and an alignment buffer 23 are provided.

  The image coding apparatus according to the embodiment has the same configuration as that of the conventional image coding apparatus shown in FIG. 11 except that a prediction block control unit 22 and an alignment buffer 23 are added. Therefore, among the constituent elements that constitute the image encoding apparatus according to the embodiment, the constituent elements that have the same functions as the constituent elements that constitute the conventional image encoding apparatus are configured to constitute the conventional image encoding apparatus. The same code as that assigned to the element is assigned. In the present embodiment, the functions and operations of the prediction block control unit 22 and the sorting buffer 23 will be mainly described.

  The block dividing unit 11 divides each picture of the input moving image into macro blocks (input macro blocks), and the input macro blocks are input to the subtracting unit 12. The subtraction unit 12 subtracts the pixel value of the macroblock (predicted macroblock) generated by the intra prediction unit 20 from the pixel value of the input macroblock for the pixel at each position, and the macro configured by the obtained value Output block (difference macroblock). The difference macroblock is input to the orthogonal transformation unit (T) 13, and the orthogonal transformation unit (T) 13 orthogonally transforms the difference macroblock.

  The quantization unit (Q) 14 quantizes each coefficient obtained by the orthogonal transform unit (T) 13 according to the quantization parameter input from the rate control unit 21, and the obtained value (quantized orthogonal transform coefficient) Are output to the alignment buffer 23 and the inverse quantization unit (iQ) 16. The encoding unit 15 encodes the quantized orthogonal transform coefficient from the alignment buffer 23 and information on the prediction mode number selected by the intra prediction unit (IPD) 20 described later, and outputs the streamed data.

  The inverse quantization unit (iQ) 16 inversely quantizes the quantized orthogonal transform coefficient from the quantization unit (Q) 14 according to the quantization parameter input from the rate control unit 21, and the inverse orthogonal transform unit (iT) 17 To supply. The inverse orthogonal transform unit (iT) 17 performs inverse orthogonal transform on the coefficient supplied from the inverse quantization unit (iQ) 16 to restore the differential macroblock. The restored difference macroblock is input to the adding unit 18 together with the prediction macroblock generated by the intra prediction unit 20.

  The adder 18 adds the restored difference macroblock pixel value and the predicted macroblock pixel value supplied from the intra prediction unit (IPD) 20 to the macroblock (reproduced macroblock). ) Is generated. This reproduced macroblock is stored in the frame memory 19. The reproduced macroblock stored in the frame memory 19 is supplied to an intra prediction unit (IPD) 20.

  The intra prediction unit (IPD) 20 reads a playback macroblock and a playback block (a block played back in a macroblock that is currently being played back) from the frame memory 19 and is supplied from a prediction block control unit 22 described later. A prediction block is generated according to the position information of the predicted block and the prediction mode used in the prediction block. The intra prediction unit (IPD) 20 calculates a prediction evaluation value for the block of the input signal corresponding to the prediction block (the input block at the same position of the corresponding original image) for the generated prediction block of each prediction mode, A prediction block is determined by selecting the best prediction mode from the prediction evaluation value. The determined prediction block is supplied to the adding unit 18.

  As a prediction evaluation value, in intra prediction, a distortion by a sum of absolute values or square values of error signals of a prediction block and an input block in each prediction mode, or a sum of absolute values of error signals after Hadamard transform, and the like, A value calculated by a Rate-Distortion (RD) function that optimizes the balance with the Rate, which is the bit amount when the prediction mode is encoded, is used.

  The prediction block control unit 22 includes information indicating which block of the macroblock to be predicted has been intra-predicted by the intra prediction unit (IPD) 20 (information regarding the position of the block and whether the intra processing of the block has been completed). Information indicating whether or not). In addition, the prediction block control unit 22 outputs to the intra prediction unit (IPD) 20 a control signal that specifies which block of the macroblocks to be predicted should be intra-predicted in which prediction mode. Further, the prediction block control unit 22 supplies position information of each block in the macroblock to the alignment buffer 23.

  The alignment buffer 23 encodes the processed blocks in the macroblock in a default (standard) order (see FIG. 2) based on the block position information supplied from the prediction block control unit 22. This is a buffer to be output to the unit 15. FIG. It is a figure which shows the order of the process of each block when a certain macroblock defined by H.264 standard is divided | segmented into 16 blocks. This order is called “raster scan order”. Here, as shown in FIG. 3, 16 blocks in the macro block are set to 4 × 4_block (X, Y), and the position (X, Y) of each block is (0, 0) to (3, 3). It is defined as At this time, for example, if the prediction block control unit 22 designates the blocks in the macro block to be processed in the order shown in FIG. 4A, as shown in 4 × 4_block (X, Y) in FIG. In addition, the intra prediction information and the quantized orthogonal transform coefficient of each block are stored in the alignment buffer 23 from the quantization unit (Q) 14 in the order of the processed blocks.

  The alignment buffer 23 includes the position information of each block in the macroblock supplied from the prediction block control unit 22 and the position information of each default block incorporated in advance in the alignment buffer 23 (FIGS. 2 and 5 ( B) 4 × 4_block (X, Y)), the intra prediction information and quantized orthogonal transform coefficient of each block accumulated in units of macroblocks are shown in FIG. ) Are extracted in the order of default (standard) and output to the encoding unit 15 as indicated by the arrow to).

  Hereinafter, several intra prediction encoding processes will be described, focusing on the order in which the prediction block control unit 22 processes each block in the macroblock and in which prediction mode. .

(First intra prediction)
First, as shown in FIG. 6A, a case where intra processing (IPD) and processing from orthogonal transform to decoding are executed in the default order (raster scan order: see FIG. 2) will be described. . In this specification, processing from orthogonal transformation to decoding, that is, orthogonal transformation performed by the orthogonal transformation unit (T) 13, quantization performed by the quantization unit (Q) 14, and inverse quantization unit (iQ) 16 And the inverse orthogonal transform performed by the inverse orthogonal transform unit (iT) 17 are described as “TQiQiT”.

  In the first intra prediction, as shown in FIG. 6C, the IPD and TQiQiT are pipelined (so that there is no space in the processing), and the intra prediction accuracy is not lowered as much as possible. In addition, a prediction mode for intra prediction is set. That is, as shown in FIG. 6B, the prediction mode of each block is set so that pipelining is possible and as many prediction modes as possible are used. Each block with a number in FIG. 6C indicates a block with the same number in FIG. 6A.

  Note that which prediction mode within the standard is selected as a candidate (option) is a problem in mounting the image coding apparatus. The image decoding apparatus performs inverse intra prediction according to the prediction mode finally selected by the image encoding apparatus. Therefore, H.264 is included in the image decoding apparatus. It is not necessary to add a function outside the H.264 standard.

  When each of the 16 blocks is intra-predicted in the default order (raster scan order), prediction mode candidates (options) in each block are set as shown in FIG. As shown in FIG. 3, when the positions of the 16 blocks are represented as (0, 0) to (3, 3), in the first intra prediction, as shown in FIG. The blocks that can be intra-predicted by the mode are four blocks (0, 0), (0, 2), (2, 0), and (2, 2). The blocks that can be intra-predicted by the seven prediction modes except mode 3 and mode 7 are the four blocks (0, 1), (0, 3), (2, 1), and (2, 3). is there. The remaining 8 blocks: (1, 0), (1, 1), (1, 2), (1, 3), (3, 0), (3, 1), (3, 2) and For eight blocks (3, 3), intra prediction is possible only in mode 0, mode 3, and mode 7.

  The operation of the image coding apparatus according to the present embodiment in the above-described default order and when the prediction mode for intra prediction is limited will be described with reference to the flowchart of FIG.

  In step S201, the prediction block control unit 22 selects a block to be predicted in the default order, and proceeds to step 202 for intra prediction and step S203 for reading the pixel value of the original image corresponding to the selected block.

  In the intra prediction in step S202, the intra prediction unit 20 first generates a prediction block for each block in the order shown in FIG. 6A in each prediction mode shown in FIG. 6B. Then, the intra prediction unit 20 calculates the prediction evaluation value using the pixel value of the prediction block in each prediction mode and the pixel value of the original image block of the input signal corresponding to the prediction block read in step S203, Determine the best prediction mode. In step S203, the intra prediction unit 20 reads the pixel value of the original image.

  Next, in step S204, the subtraction unit 12 subtracts the pixel value of the prediction block obtained by the best prediction mode determined by the intra prediction unit 20 from the pixel value of the original image block for the pixel at each position. The orthogonal transform unit (T) 13 performs orthogonal transform on the difference block obtained by the subtraction unit 12 in step S205, and the quantization unit (Q) 14 is obtained by the orthogonal transform unit (T) 13 in step S206. Quantize the orthogonal transform coefficient. The quantized orthogonal transform coefficient is input to the encoding unit 15 via the alignment buffer 23, and in step 207, the encoding unit 15 encodes the quantized orthogonal transform coefficient.

  On the other hand, the quantized orthogonal transform coefficient obtained in step S206 is input to the inverse quantization unit (iQ) 16, and in step S208, the inverse quantization unit (iQ) 16 dequantizes the quantized orthogonal transform coefficient. . In step S209, the inverse orthogonal transform unit (iT) 17 performs inverse orthogonal transform on the orthogonal transform coefficient obtained by the inverse quantization unit (iQ) 16. In step S210, the addition unit 18 adds the difference value of the prediction block subjected to inverse orthogonal transform and the prediction block obtained in step S202. Thereby, the decoding block is reconstructed. The reconstructed decoded block is stored in the frame memory 19 in step 211. The decoded block stored in the frame memory 19 is supplied to the intra prediction unit 20 and used as a reference pixel for intra prediction in step S202.

  In step S212, the intra prediction unit 20 determines whether all the blocks of the macroblock to be predicted have been processed (reconstructed). If all the blocks have not been processed (No in step S212), the process returns to step S201. The image coding apparatus according to the present embodiment performs the same processing as described above for the remaining blocks. On the other hand, if all the blocks have been processed (Yes in step S212), the image coding apparatus according to the present embodiment performs the above processing for the next macroblock.

  In the first intra prediction, since the blocks are processed in the default order, the alignment buffer 23 is not necessary. However, even if the sorting buffer 23 is provided, the default order information is supplied from the prediction block control unit 22 to the sorting buffer 23, so that the data of each block is processed in the order of processing (default order). It is output to the encoding unit 15 as it is, and no problem occurs.

(Second intra prediction)
Next, as shown in FIG. 8A, a case will be described in which intra processing (IPD) and processing from orthogonal transform to decoding (TQiQiT) are executed in an order different from the default order.

  Also in the second intra prediction, as shown in FIG. 8C, the IPD and TQiQiT are pipelined (so that there is no space in processing), and the intra prediction accuracy is not lowered as much as possible. In addition, a prediction mode for intra prediction is set. That is, as shown in FIG. 8B, the prediction mode of each block is set so that pipelining is possible and as many prediction modes as possible are used. Each block with a number in FIG. 8C indicates a block with the same number in FIG. 8A.

  When each of the 16 blocks is intra-predicted in an order different from the default order (an order different from the raster scan order), prediction mode candidates (options) in each block are set as shown in FIG. 8B. The That is, in the second intra prediction, as shown in FIG. 8B, the blocks that can be intra-predicted by all nine prediction modes are (0, 0), (0, 2), (1, 1 ), (1,2), (1,3), (2,0), (2,2), (3,0), (3,1) and (3,2) . In addition, four blocks (0, 1), (0, 3), (2, 1), and (2, 3) can be intra-predicted in seven prediction modes except for mode 3 and mode 7. It is a block. The remaining two blocks, ie, the two blocks (1, 0) and (3, 3), can be intra predicted only in mode 0, mode 3, and mode 7.

  When the second intra prediction is executed, the number of blocks that can be intra-predicted in all prediction modes is greatly increased from four to ten compared to the case where the first intra prediction is executed. That is, when the second intra prediction is executed, the accuracy of the intra prediction is greatly increased as compared with the case where the first intra prediction is executed.

  Hereinafter, the operation of the image coding apparatus according to the present embodiment when the prediction mode for intra prediction is limited in an order different from the default order will be described with reference to the flowchart of FIG. 9.

  In step S301, the prediction block control unit 22 selects a prediction target block in an order different from the default order, and proceeds to step 302 for intra prediction and step S303 for reading the pixel value of the original image corresponding to the selected block. move on. The prediction block control unit 22 outputs the position information of the block to be processed to the alignment buffer 23.

  In the intra prediction in step S302, the intra prediction unit 20 first generates a prediction block for each block in the order shown in FIG. 8A in each prediction mode shown in FIG. 8B. Then, the intra prediction unit 20 calculates the prediction evaluation value using the pixel value of the prediction block of each prediction mode and the pixel value of the original image block of the input signal corresponding to the prediction block read in step S303, Determine the best prediction mode. In step S303, the intra prediction unit 20 reads the pixel value of the original image.

  Next, in step S304, the subtraction unit 12 subtracts the pixel value of the prediction block obtained by the best prediction mode determined by the intra prediction unit 20 from the pixel value of the original image block for the pixel at each position. In step S305, the orthogonal transform unit (T) 13 performs orthogonal transform on the difference block obtained by the subtraction unit 12, and in step S306, the quantization unit (Q) 14 is obtained by the orthogonal transform unit (T) 13. Quantize the coefficients. In step S307, the quantized orthogonal transform coefficients are stored in the alignment buffer 23 with a block as a unit.

  On the other hand, the orthogonal transform coefficient (quantized orthogonal transform coefficient) of the difference value of the prediction block obtained in step S306 is input to the inverse quantization unit (iQ) 16, and in step S308, the inverse quantization unit (iQ) 16 Dequantizes the quantized orthogonal transform coefficients. In step S309, the inverse orthogonal transform unit (iT) 17 performs inverse orthogonal transform on the orthogonal transform coefficient obtained by the inverse quantization unit (iQ) 16. In step S310, the addition unit 18 adds the difference value of the prediction block subjected to inverse orthogonal transform and the prediction block obtained in step S302. Thereby, the decoding block is reconstructed. The reconstructed decoded block is stored in the frame memory 19 in step 311. The decoded blocks stored in the frame memory 19 are supplied to the intra prediction unit 20 and used as reference pixels for intra prediction in step S302.

  In step S312, the intra prediction unit 20 determines whether or not all blocks of the macroblock to be predicted have been processed (reconstructed). If all blocks have not been processed (No in step S312), the process returns to step S301. The image coding apparatus according to the present embodiment performs the same processing as described above for the remaining blocks. On the other hand, if all the blocks have been processed (Yes in step S312), the process proceeds to step S313, and the image coding apparatus according to the present embodiment performs the above-described process for the next macroblock.

  In step S313, the alignment buffer 23 extracts the quantized orthogonal transform coefficients of the difference values of the prediction blocks accumulated in units of blocks in step S307 in the default order, and outputs them to the encoding unit 15. In step S314, the encoding unit 15 encodes the quantized orthogonal transform coefficients in the default order.

  By the way, in said 1st intra prediction and 2nd intra prediction, the number and content of prediction modes differ for every block. From the viewpoint of processing uniformity, when the same prediction mode is set for all blocks, the following third intra prediction can be used.

(Third intra prediction)
In the third intra prediction, each block is intra-predicted in an order different from the default order shown in FIG. 10A, and as shown in FIG. 10C, IPD and TQiQiT are pipelined. (There are two empty places in the process), and the prediction mode of each block is set as shown in FIG.

  When each of the 16 blocks is intra-predicted by the prediction mode shown in FIG. 10B in an order different from the default order, there are no blocks that can be intra-predicted by all nine prediction modes. However, intra prediction can be performed for all blocks in seven prediction modes except for mode 3 and mode 7. However, the two blocks (0, 1) and (3, 3) hinder pipelining.

  The third intra prediction is executed by causing the sorting buffer 23 to intra predict each block in an order different from the default order shown in FIG. 10A in the prediction mode shown in FIG. . The sorting buffer 23 sorts the processed blocks in a default order and outputs the sorted blocks to the encoding unit 15.

  The image encoding device of the present invention is suitable for an image encoding device that encodes a picture in units of blocks. Specifically, a web server that distributes moving image data, a terminal device that receives moving image data, and a moving image It is suitable for a digital camera capable of recording / reproducing image data, a mobile phone with a camera, a DVD recorder / player, a PDA, a personal computer, and the like.

11 Block division unit 12 Subtraction unit 13 Orthogonal transformation unit (T)
14 Quantizer (Q)
15 Encoding unit 16 Inverse quantization unit (iQ)
17 Inverse orthogonal transform unit (iT)
18 Adder 19 Frame memory 20 Intra prediction (IPD)
21 Rate control unit 22 Predictive block control unit 23 Sorting buffer

Claims (3)

An image encoding device that performs orthogonal transform, quantization, inverse quantization, inverse orthogonal transform, and intra prediction based on a predetermined coding standard for each of a plurality of blocks obtained by dividing one macroblock. ,
The macro block is composed of 16 pixels × 16 lines of pixels,
The block is composed of 4 pixels × 4 lines of pixels,
Taking the block located at the upper left of the macro block as a reference, taking the X axis in the horizontal direction and the Y axis in the vertical direction, the position of the block is (X, Y) {X, Y = 0, 1, 2, 3 },
A prediction block in which an intra prediction mode to be applied to all blocks in the macroblock is set from any one of modes 0, 1, 2, 4, 5, 6, and 8 defined in the predetermined standard A control unit;
The block is divided into (0,0), (1,0), (0,1), (2,0), (1,1), (3,0), (2,1), (0,2 ), (3, 1), (1, 2), (0, 3), (2, 2), (1, 3), (3, 2), (2, 3), (3, 3) In order, an image encoding unit that performs in-plane prediction encoding using the intra prediction mode set by each block by the prediction block control unit,
The intra-prediction encoded blocks are denoted by (0,0), (1,0), (0,1), (2,0), (1,1), (3,0), (2, 1), (0, 2), (3, 1), (1, 2), (0, 3), (2, 2), (1, 3), (3, 2), (2, 3) , (3, 3) in the order of inverse quantization and inverse orthogonal transform,
The image encoding unit and the conversion unit include intra prediction encoding from a block located at (0, 1) to a block located at (2, 3) among the blocks, and (1, An image coding apparatus that performs inverse quantization and inverse orthogonal transformation from a block located at 0) to a block located at (3, 2) in parallel at the same timing. The image encoding unit performs intra prediction prediction encoding a block located at (0, 0) at a first timing,
The transform unit performs inverse quantization and inverse orthogonal transform on a block located at (0, 0) at a second timing that is a timing next to the first timing,
Further, the image encoding unit performs intra prediction encoding at a third timing that is a timing next to the second timing, for the block located at (1,0).
The image encoding unit and the conversion unit are located at (2, 3) from the block located at (0, 1) among the blocks after the fourth timing that is the timing next to the third timing. In-plane predictive coding up to the block, and inverse quantization and inverse orthogonal transformation from the block located at (1,0) to the block located at (3,2) among the blocks are performed in parallel at the same timing. The image encoding device according to claim 1, wherein the image encoding device is processed. An image encoding method for performing orthogonal transform, quantization, inverse quantization, inverse orthogonal transform, and intra prediction based on a predetermined coding standard for each of a plurality of blocks obtained by dividing one macroblock. ,
The macro block is composed of 16 pixels × 16 lines of pixels,
The block is composed of 4 pixels × 4 lines of pixels,
Taking the block located at the upper left of the macro block as a reference, taking the X axis in the horizontal direction and the Y axis in the vertical direction, the position of the block is (X, Y) {X, Y = 0, 1, 2, 3 },
The block located at (0,0) is subjected to in-plane predictive coding at the first timing,
A block located at (0,0) is inversely quantized and inversely orthogonal transformed at a second timing that is a timing next to the first timing;
The block located at (1, 0) is subjected to in-plane predictive coding at a third timing that is a timing next to the second timing,
After the fourth timing, which is the next timing after the third timing, intra prediction encoding from the block located at (0, 1) to the block located at (2, 3) among the blocks, The inverse quantization and inverse orthogonal transformation from the block located at (1,0) to the block located at (3,2) are processed in parallel at the same timing,
An image encoding method in which the block is always set from any one of modes 0, 1, 2, 4, 5, 6, and 8 defined in the predetermined standard when performing intra-frame encoding.

Images