JP5194039B2 - Intra prediction apparatus, encoder, decoder, and program - Google Patents

Description

  The present invention relates to an apparatus and a program for predicting a pixel value to be encoded from decoded pixel values by extrapolation, and more particularly to an intra prediction apparatus, an encoder, and an intra prediction apparatus that can perform prediction without impairing the continuity of alternating current components of an image. The present invention relates to a decoder and a program.

  MPEG-4 AVC / H. In the H.264 system (ISO / IEC 14496-10 / ITU-T Rec. H.264), in addition to inter prediction (inter-screen prediction), the pixel value of an adjacent block that has been encoded is determined for an image block to be encoded. Intra-prediction coding (intra-screen prediction coding) that generates a prediction image and encodes a difference from the prediction image is employed.

  FIG. 9 shows a conventional MPEG-4 AVC / H. The block diagram of the encoder 4 in a H.264 system is shown. The encoder 4 includes a rearrangement unit 31, a subtraction unit 32, an orthogonal transformation unit 33, a quantization unit 34, a variable length coding unit 35, an inverse quantization unit 36, and an inverse orthogonal transformation unit 37. , A changeover switch 38, an intra prediction unit 39, a frame memory 40, a motion compensation prediction unit 41, and an addition unit 42.

  The rearrangement unit 31 temporarily stores the input video signal, rearranges the order of the frame images, and outputs the frame image necessary for the encoding process to the subtraction unit 32 and the motion compensation prediction unit 41.

  The motion compensation prediction unit 41 performs motion vector detection on the input image supplied from the rearrangement unit 31 using a reference image acquired from the frame memory 40, and performs motion compensation using the obtained motion vector. The predicted image obtained as a result is output to the subtraction unit 32 and the addition unit 42 via the changeover switch 38. The motion vector information is output to the variable length encoding unit 35.

  The subtraction unit 32 generates a difference image between the input image from the rearrangement unit 31 and the prediction image from the motion compensation prediction unit 41 or the intra prediction unit 39 and outputs the difference image to the orthogonal transformation unit 33.

  The orthogonal transform unit 33 performs orthogonal transform (for example, DCT; Discrete Cosine Transform) on the difference image supplied from the subtraction unit 32 and outputs the orthogonal transform coefficient to the quantization unit 34. .

  The quantization unit 34 selects a quantization table for the orthogonal transform coefficient input from the orthogonal transform unit 33, performs a quantization process, and outputs the quantization table 35 and the inverse quantization unit 36.

  The variable length encoding unit 35 scans the quantized orthogonal transform coefficient input from the quantization unit 34 to perform variable length encoding processing to generate a bit stream, and also receives the input from the motion compensation prediction unit 41. The motion vector information is also subjected to variable length coding and output.

  The inverse quantization unit 36 performs an inverse quantization process on the quantized orthogonal transform coefficient input from the quantization unit 34 and outputs the result to the inverse orthogonal transform unit 37.

  The inverse orthogonal transform unit 37 performs inverse orthogonal transform (for example, IDCT; Inverse Discrete Cosine Transform) on the orthogonal transform coefficient input from the inverse quantization unit 36 and outputs the result to the addition unit 42.

  The adding unit 42 adds the image obtained by the inverse orthogonal transform obtained from the inverse orthogonal transform unit 37 and the prediction image obtained from the motion compensation prediction unit 41 or the intra prediction unit 39 to generate a decoded image, and generates an intra prediction unit. 39 and the frame memory 40.

  The changeover switch 38 switches between motion compensation prediction and intra prediction.

  The intra prediction unit 39 generates a prediction image predicted intra from the image obtained by decoding the already-encoded block (the output image of the addition unit 42), and outputs the prediction image to the subtraction unit 32 and the addition unit 42. Here, the subtraction unit 32 outputs the difference image between the predicted image and the original image to the orthogonal transformation unit 33 and encodes it via the quantization unit 34 and the variable length coding unit 35.

  Intra prediction is performed in units of 4 pixels × 4 lines, 8 pixels × 8 lines, or 16 pixels × 16 lines, and a plurality of types of prediction modes (prediction directions) (for example, prediction in units of 4 pixels × 4 lines). Selects the optimum prediction direction from among nine types. FIG. 10 is a diagram illustrating a prediction mode when prediction is performed in units of 4 pixels × 4 lines. In the figure, hatched circles indicate decoded pixels, white circles indicate pixels to be predicted, and arrows indicate prediction directions. (A) Prediction mode 0 is vertical direction prediction, (b) Prediction mode 1 is horizontal direction prediction, (c) Prediction mode 2 is DC prediction, (d) Prediction mode 3 is diagonal lower left direction prediction, ( e) Prediction mode 4 in diagonal lower right direction prediction, (f) prediction mode 5 in vertical right direction prediction, (g) prediction mode 6 in horizontal down direction prediction, and (h) prediction mode 7 in vertical left direction. In the prediction mode 8 of (i), horizontal upward prediction is performed. The above is MPEG-4 AVC / H. This is a technique of intra prediction in the H.264 system.

  In addition, in order to increase the accuracy of intra prediction, a technique of encoding a difference value between adjacent pixels by always referring to adjacent pixels instead of referring to spatially separated pixels (for example, Patent Document 1). And a technique for performing intra prediction based on the value of a pixel adjacent to a prediction block and the value of a pixel between the prediction block and one or more pixels is known (see, for example, Patent Document 2). Also known is a technique for generating a value that takes into account the frequency characteristics of adjacent decoded images as a prediction value in intra coding in order to improve the prediction performance for a picture image whose pixel value has a certain change tendency. (For example, see Patent Document 3). In addition, a technique for obtaining an orthogonal transform coefficient for a pixel group defined only in a part of a rectangular block for encoding an arbitrarily shaped image is also known (see, for example, Patent Document 4).

JP 2009-049969 A JP 2008-271371 A JP 2008-245088 A Japanese Patent No. 3907724

  However, the conventional intra prediction method with directionality assumes that the value of the decoded adjacent pixel and the value of the prediction target pixel are continuous in a DC sense, and is continuous in an AC sense. Sex was not assumed. Further, although the technique described in Patent Document 3 performs prediction including an AC component, there is no AC (phase) continuity between the reference pixel and the prediction pixel. For this reason, when the reference image is a part of a periodic pattern, it has not been possible to perform advanced prediction considering the AC component. In addition, there is a problem that the coding distortion included in the decoded image and the noise component included in the original image directly affect the predicted value. Patent Document 4 is intended to obtain an orthogonal transform coefficient of a defined pixel value, and is not intended to estimate an undefined pixel value.

  In order to solve the above-described problem, an object of the present invention is to provide an intra-prediction device that performs prediction that does not impair the continuity of the AC component between the reference pixel and the prediction pixel, and performs prediction with less influence of the noise component, and encoding. A decoder, a decoder, and a program.

  In order to solve the above-described problem, an intra prediction apparatus according to the present invention includes a processing block generation unit that generates a processing block including a reference block and a prediction block, and the orthogonal transform basis vector corresponding to the entire region of the processing block. A first transformation basis vector (main transformation basis) extracted according to the shape of the block, and a second transformation basis vector (sub transformation basis) obtained by extracting the orthogonal transformation basis vector according to the shape of the prediction block. A transform basis generation means for generating, and a weighted linear sum operation using a transform coefficient sequence corresponding to the transform basis vector corresponding to the predetermined order as a weight for the transform basis vector corresponding to the predetermined order in the first transform basis vector A reference image approximation means for approximating the reference image with a vector, and a predetermined order of the second transformed basis vectors To transform basis vector, by performing a weighted linear sum operation to weight the transform coefficient sequence used in the approximation, characterized in that it comprises a prediction image generation unit that generates a predicted image.

  In the intra prediction apparatus according to the present invention, the reference image approximating unit calculates a third transformed basis vector (corrected main transformation basis) obtained by correcting a norm with respect to the first transformed basis vector (main transformation basis). The conversion coefficient sequence is obtained using the method.

  In the intra prediction apparatus according to the present invention, the third transformed basis vector is a vector obtained by dividing the first transformed basis vector by the square of the norm of the vector and correcting the norm. Features.

  In the intra prediction apparatus according to the present invention, the reference image approximating unit performs an inner product operation of the transform base vector and the vector of the reference image in order from the transform base vector of the lower order among the third transform base vectors. Thus, a conversion coefficient for the conversion base vector is obtained, and a vector obtained by subtracting a vector obtained by multiplying the conversion coefficient by the conversion base vector of the order from the first conversion base vector from the vector of the reference image is newly added. A transform coefficient of the next order is obtained as a reference image vector, and the transform coefficient sequence is determined by repeating the process of obtaining the transform coefficient a predetermined number of times.

  In the intra prediction apparatus according to the present invention, a plurality of parameters including the reference block and the configuration information of the prediction block and the predetermined order information are prepared in advance, and a plurality of prediction images are generated using the plurality of parameters. And a parameter transmission unit that selects one prediction image from the plurality of prediction images and transmits a parameter used to generate the selected prediction image.

  Furthermore, an encoder according to the present invention includes the intra prediction device described above.

  Furthermore, a decoder according to the present invention includes the intra prediction apparatus described above.

  Furthermore, the present invention provides (a) a processing block comprising a reference block and a prediction block to a computer that functions as an intra prediction device that generates a prediction image of a prediction block from an image of a decoded reference block adjacent to the prediction block. (B) a first transformed basis vector obtained by extracting an orthogonal transformation basis vector corresponding to the entire area of the processing block in accordance with the shape of the reference block; and the orthogonal transformation basis vector of the prediction block. A step of generating a second conversion basis vector extracted in accordance with the shape; and (c) a conversion basis vector corresponding to the predetermined order with respect to the conversion basis vector corresponding to the predetermined order of the first conversion basis vectors. The reference image is approximated by a vector by performing a weighted linear sum operation with the transformation coefficient sequence corresponding to And (d) performing a weighted linear sum operation using the transformation coefficient sequence used for the approximation as a weight on the transformation basis vectors of a predetermined order among the second transformation basis vectors, And a step for generating the program.

  According to the present invention, it is possible to perform prediction that does not impair the continuity of the AC component between the reference pixel and the prediction pixel. In addition, it is possible to perform prediction with less influence of coding distortion included in a decoded image and noise components included in an original image.

It is a figure for demonstrating the concept of the intra prediction system by this invention. It is a figure explaining derivation | leading-out of the transformation coefficient using the non-orthogonal transformation base of the intra prediction system by this invention. It is a figure which illustrates the processing block of the intra prediction apparatus of one Example by this invention. It is a block diagram of the intra prediction apparatus by the side of the encoding of one Example by this invention. It is a block diagram of a process of the intra prediction apparatus by the side of the decoding of one Example by this invention. It is a flowchart which shows operation | movement of the intra prediction apparatus of one Example by this invention. It is a block diagram of an encoder provided with the intra prediction apparatus of one Example by this invention. It is a block diagram of a decoder provided with the intra prediction apparatus of one Example by this invention. Conventional MPEG4 AVC / H. 2 is a configuration diagram of an encoder in the H.264 scheme. Conventional MPEG4 AVC / H. It is a figure which shows the prediction mode in the case of estimating by 4 pixel x4 line unit in a H.264 system.

  First, the concept of the intra prediction method according to the present invention will be described with reference to FIGS.

[Intra prediction method concept]
The intra prediction apparatus according to the present invention performs prediction with little influence of noise components without impairing continuity with decoded images including AC components. For this purpose, not only adjacent pixels but also decoded pixels included in a certain region are used, and orthogonal transform such as DCT or FFT (Fast Fourier Transform) is used. Then, a predicted image is generated by deriving an estimated value of the transform coefficient of the decoded image. Since prediction is performed using a large number of reference pixels and emphasizing low frequency components, the influence of noise can be reduced.

  A method for generating a predicted image by deriving an estimated value of a transform coefficient of a decoded image will be described with reference to FIGS. 1 and 2. FIG. 1 is a diagram for explaining the concept of an intra prediction method according to the present invention. Considering a one-dimensional region for simplicity, a case of generating a predicted image of 3 pixels from a decoded image of 5 pixels will be described with reference to FIG.

  When orthogonal transformation such as DCT is performed on a one-dimensional image of eight pixels, eight orthogonal transformation coefficients corresponding to the eight pixels are uniquely obtained, and when inverse orthogonal transformation is performed on this, the original image is restored. Predicting a predicted image of 3 pixels is equivalent to estimating 8 orthogonal transform coefficients from 5 decoded reference pixels.

  The orthogonal transformation is a process for obtaining an orthogonal transformation coefficient that is a component sequence of each orthogonal transformation base included in the signal sequence, and can also be referred to as signal analysis using the orthogonal transformation base. Here, the eight orthogonal transform bases are redundant to represent five pixel values. There are an infinite number of orthogonal transform coefficient sets that reproduce a five-pixel reference image. Since the high-frequency component has a relatively large noise compared to the signal, it is preferable to obtain a transform coefficient set that reproduces the original image as much as possible using the transform base corresponding to the low-frequency component. As a result of such estimation, a prediction image having alternating continuity can be generated for the low-frequency component.

  As will be described later, transform bases that are not orthogonal to each other are used for coefficient estimation according to the present invention. Therefore, non-orthogonal transform bases are distinguished from orthogonal transform bases and referred to as transform bases, and coefficients derived using transform bases are referred to as transform coefficients and distinguished from orthogonal transform coefficients. Note that the orthogonal transform base and the transform base are represented by vector notation for convenience of calculation.

ここで、特許文献４に記載した技法を応用して分析を行う。

In the orthogonal transform, an orthogonal transform coefficient corresponding to the orthogonal transform base can be obtained by calculating the inner product of the image vector and the orthogonal transform base vector.

Here, analysis is performed by applying the technique described in Patent Document 4.

図２は、非直交の変換基底を用いた変換係数の導出を説明する図である。

このように、変換基底が直交していない場合には、変換係数は同時に求めるのではなく、連続近似によって順次求める必要がある。

ただし、非直交の変換基底を用いているため、ｃ₁以上の係数を導出した後、再び低次の成分が出る場合もある。

FIG. 2 is a diagram illustrating the derivation of transform coefficients using non-orthogonal transform bases.

As described above, when the transform bases are not orthogonal, transform coefficients need not be obtained simultaneously but sequentially obtained by continuous approximation.

However, since a non-orthogonal transform basis is used, a low-order component may be output again after deriving a coefficient of c ₁ or more.

  Next, the case where the image area is two-dimensional will be described, but the idea is the same as that in the case of one-dimensional. FIG. 3 is a diagram illustrating an example of an image block serving as a processing unit performed by the intra prediction apparatus 1. The image blocks D, B, and A are image blocks that have been encoded and then decoded, and are image blocks that are referred to in order to predict pixel values of the prediction block (hereinafter referred to as “reference blocks”). An image in the reference block is referred to as a reference image, and an image in the prediction block is referred to as a predicted image. The block sizes of the reference blocks D, B, A and the prediction block X are arbitrary. Here, the size of the reference block D is k pixels × m lines, the size of the reference block B is l pixels × m lines, and the reference block A Is the size of k pixels × n lines, and the size of the prediction block X is 1 pixel × n lines. Further, in the following description, an image block of (k + 1) pixels × (m + n) lines made up of the reference blocks D, B, A and the prediction block X is referred to as a “processing block”.

この演算を所定の次数（ｈ次）まで繰り返し、ｃ₀，ｃ₁，・・・，ｃ_ｈを得る。ｃ₁以上の係数を導出した後、再び低次の成分が出ることがあるが、上記手順では、再出現した低次の成分を拾うことはできない。従って、上記の手順は数次程度で行い、高周波成分の変換係数を求める前に近似を終えるのが好適である。また、残差参照画像ベクトルの大きさが所定の値より小さくなった時点で近似を終えることもできる。あるいはまた、再度現れた低次の成分を拾うために、上記演算を所定の次数まで繰り返した後、得られた残差参照画像ベクトルについて再び最低次数の変換係数から順に変換係数を導出することもできる。このような手順は、予め定めておくことで送受で同一動作を行うようにすることもできるし、パラメータとして伝送することもできる。

Repeat this operation until a predetermined degree (h following) to obtain c _0, c _1, · · ·, a _{c h.} After deriving a coefficient of c ₁ or higher, low-order components may come out again, but the above procedure cannot pick up low-order components that reappear. Therefore, it is preferable to perform the above procedure in several orders and finish the approximation before obtaining the conversion coefficient of the high frequency component. The approximation can also be finished when the size of the residual reference image vector becomes smaller than a predetermined value. Alternatively, in order to pick up low-order components that reappear, the above calculation is repeated to a predetermined order, and then the transform coefficients are derived in order from the transform coefficient of the lowest order again for the obtained residual reference image vector. it can. Such a procedure can be determined in advance so that the same operation is performed by transmission and reception, or can be transmitted as a parameter.

  FIG. 4 is a block diagram of the intra prediction apparatus 1 on the encoding side according to the first embodiment of the present invention. The intra prediction apparatus 1 includes a processing block generation unit 11, a reference image holding unit 12, an orthogonal transform base providing unit 13, an orthogonal transform base decomposition unit 14, a sub-transform base holding unit 15, and a main transform base holding unit 16. A main conversion base correction unit 17, a correction main conversion base holding unit 18, an inner product calculation unit 19, a predicted image generation unit 20, a reference image correction unit 21, a repetition setting unit 22, a parameter transmission unit 23, Is provided.

  The processing block generation unit 11 generates a processing block including the reference blocks D, B, and A and the prediction block X and outputs the processing block to the reference image holding unit 12. In addition, information on the configuration (block size) of the reference blocks D, B, A and the prediction block X is output to the parameter transmission unit 23 as a parameter.

  The repetition setting unit 22 sets the number of repetitions of the arithmetic processing in the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21 to the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21. And the repetition count information is output to the parameter transmission unit 23 as a parameter. The repetition number information is, for example, a predetermined number of times when the calculation is repeated a predetermined number of times and the approximation is completed, and when the approximation is finished when the size of the residual reference image vector becomes smaller than a predetermined value. When the calculation is repeated for the second predetermined number of times from the transform coefficient of the lowest order with respect to the residual reference image vector obtained after repeating the calculation for the first predetermined number of times, the first predetermined number of times and The second predetermined number of times.

Next, the second process in the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21 will be described.

In the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21, the above procedure is repeated from the low-frequency conversion base to the high-frequency conversion base and the number of times set by the repetition setting unit 22. Then, the predicted image generation unit 20 outputs the predicted image that has been added so far.

なお、ｎ＝ｈのとき、式（８）の演算結果は式（６）の演算結果と等しくなる。

When n = h, the calculation result of Expression (8) is equal to the calculation result of Expression (6).

  The parameter transmission unit 23 is the parameters input from the processing block generation unit 11, the information about the configuration (block size) of the reference blocks D, B, A and the prediction block X, and the parameters input from the repetition setting unit 22. The information on the number of repetitions is encoded and transmitted to the outside. These parameters can be specified in advance and shared by transmission and reception. In this case, the parameter transmission unit 23 is not required. In addition, a plurality of types of these parameters are tried on the encoding side to generate a plurality of prediction images, an optimum prediction image is selected in terms of prediction encoding performance, and the parameters used for generation of the selected prediction image are set. It can also be transmitted.

  FIG. 5 is a configuration diagram of the decoding-side intra prediction device 1 according to the first embodiment of the present invention. In the case of transmitting parameters using the parameter transmission unit 23 in the encoding-side intra prediction device 1, the parameter reception unit 24 is provided. The parameter reception unit 24 receives parameters from the parameter transmission unit 23 of the encoding-side intra prediction device 1 and outputs the received parameters to the processing block generation unit 11 and the repetition setting unit 22. The processing block generation unit 11 generates a processing block based on the parameters input from the parameter reception unit 24, and the repetition setting unit 22 converts the parameters input from the parameter reception unit 24 into the inner product calculation unit 19 and the predicted image generation. To the unit 20 and the reference image correction unit 21. The other processes are the same as those of the encoding-side intra apparatus 1, and thus the description thereof is omitted.

[Operation of intra prediction device]
FIG. 6 is a flowchart showing the operation of the intra prediction apparatus 1 according to the present invention. First, in step S101, the processing block generation unit 11 generates a processing block.

  In step S <b> 106, the repetition setting unit 22 sets the number of repetitions of the calculation process in the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21.

  In step S108, the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21 assign 1 to the variable n in order to perform the calculation processing in step S109, step S110, and step S111, respectively.

In step S109, the inner product calculation unit 19 derives a conversion coefficient cn _-1 based on the equation (7).

  In step S112, the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21 determine whether the variable n has reached the number of repetitions set by the repetition setting unit 22. If the variable n has not reached the number of repetitions, the process proceeds to step S113, and if the variable n has reached the number of repetitions, the process ends.

  In step S113, the inner product calculation unit 19, the predicted image generation unit 20, and the reference image correction unit 21 substitute n + 1 for the variable n, and perform steps S109, S110, and S111 for the number of times set by the repetition setting unit 22. Repeat the process.

[Encoder and decoder]
FIG. 7 is a configuration diagram of the encoder 2 including the intra prediction apparatus according to the first embodiment of the present invention. The encoder 2 is a conventional MPEG4 AVC / H. Instead of the intra prediction unit in the H.264 encoder (see FIG. 9), the intra prediction apparatus 1 according to the first embodiment of the present invention is provided. The intra prediction apparatus 1 inputs an image obtained by decoding an already-encoded block through the inverse quantization unit 36 and the inverse orthogonal transform unit 37 to the reference image holding unit 12 and subtracts the generated prediction image via the changeover switch 38. To the unit 32 and the adder 42. The parameter transmission unit 23 outputs the set parameters to the variable length coding unit 35. As described above, when a fixed parameter is used, it is not necessary to provide the parameter transmission unit 23 in the intra prediction apparatus 1.

  FIG. 8 is a configuration diagram of the decoder 3 including the intra prediction apparatus according to the first embodiment of the present invention. The decoder 3 switches between a variable length decoding unit 51, an inverse quantization unit 52, an inverse orthogonal transform unit 53, an addition unit 54, an intra prediction device 1, a frame memory 55, and a motion compensation prediction unit 56. A switch 57 and a rearrangement unit 58 are provided. The decoder 3 is a conventional MPEG4 AVC / H. An intra prediction device 1 according to the present invention is provided instead of the intra prediction unit of the H.264 decoder.

  The variable length decoding unit 51 receives a bit stream encoded by inter-frame prediction, performs variable length decoding processing, outputs the bit stream to the inverse quantization unit 52, and decodes motion vector information to perform motion compensation prediction unit To 56. Further, when parameter information is transmitted from the encoder side, the parameter information is decoded and output to the parameter receiving unit 24.

  The inverse quantization unit 52 performs an inverse quantization process on the quantized orthogonal transform coefficient input from the variable length decoding unit 51 to obtain an orthogonal transform coefficient of the motion compensated difference image, and performs an inverse orthogonal transform To the unit 53.

  The inverse orthogonal transform unit 53 performs inverse orthogonal transform (for example, IDCT) on the orthogonal transform coefficient of the difference image input from the inverse quantization unit 52, and outputs the obtained difference image to the addition unit 54.

  The addition unit 54 adds the difference image obtained from the inverse orthogonal transform unit 53 and the prediction image input from the motion compensation prediction unit 56 or the prediction image input from the intra prediction device 1 to restore the image, The data is output to the rearrangement unit 58, the intra prediction apparatus 1, and the frame memory 55.

  The motion compensation prediction unit 56 generates a prediction image using the reference image obtained from the frame memory 55 and the motion vector obtained from the variable length decoding unit 51, and outputs the prediction image to the addition unit 54 via the changeover switch 57.

  The rearrangement unit 58 rearranges the frames of the restored decoded image input from the addition unit 54 in the order of display frames and outputs the rearranged frames.

  The intra prediction apparatus 1 inputs the image decoded through the inverse quantization unit 52 and the inverse orthogonal transform unit 53, and outputs the generated predicted image to the addition unit 54 via the changeover switch 57.

  Here, in order to function as the intra prediction apparatus 1, a computer can be preferably used. Such a computer includes a processing block generation unit 11, a reference image holding unit 12, an orthogonal transform base providing unit 13, and the like. Orthogonal transform base decomposition unit 14, sub-transform base holding unit 15, main transform base holding unit 16, main transform base correcting unit 17, corrected main transform base holding unit 18, inner product calculating unit 19, and prediction image generation The control unit for causing the unit 20, the reference image correction unit 21, the repetition setting 21, and the parameter transmission unit 23 or the parameter reception unit 24 to function is configured by a CPU (Central Processing Unit) and at least one memory. This can be realized with a storage unit.

  Further, by causing such a computer to execute a predetermined program by the CPU, a processing block generation unit 11, a reference image holding unit 12, an orthogonal transformation base providing unit 13, an orthogonal transformation base decomposition unit 14, A conversion base holding unit 15, a main conversion base holding unit 16, a main conversion base correction unit 17, a corrected main conversion base holding unit 18, an inner product calculation unit 19, a predicted image generation unit 20, and a reference image correction unit 21 And the function which the repeat setting 21 and the parameter transmission part 23 or the parameter receiving part 24 have can be implement | achieved.

  Similarly, in order to function as the encoder 2, a computer can be suitably used. Such a computer includes an intra prediction device 1, a rearrangement unit 31, a subtraction unit 32, and an orthogonal transformation unit 33. A quantization unit 34, a variable length coding unit 35, an inverse quantization unit 36, an inverse orthogonal transform unit 37, a changeover switch 38, a frame memory 40, a motion compensation prediction unit 41, and an addition unit 42. A control unit for functioning can be realized by a CPU and a storage unit including at least one memory. Further, by causing such a computer to execute a predetermined program by the CPU, the intra prediction device 1, the rearrangement unit 31, the subtraction unit 32, the orthogonal transformation unit 33, the quantization unit 34, the variable length The functions of the encoding unit 35, the inverse quantization unit 36, the inverse orthogonal transform unit 37, the changeover switch 38, the frame memory 40, the motion compensation prediction unit 41, and the addition unit 42 can be realized.

  Similarly, in order to function as the decoder 3, a computer can be preferably used. Such a computer includes an intra prediction device 1, a variable length decoding unit 51, an inverse quantization unit 52, and an inverse orthogonal transform. The control unit for causing the unit 53, the addition unit 54, the frame memory 55, the motion compensation prediction unit 56, the changeover switch 57, and the rearrangement unit 58 to function is configured by a CPU and at least one memory. This can be realized with a storage unit. In addition, by causing such a computer to execute a predetermined program by the CPU, the intra prediction device 1, the variable length decoding unit 51, the inverse quantization unit 52, the inverse orthogonal transform unit 53, and the addition unit 54 The functions of the frame memory 55, the motion compensation prediction unit 56, the changeover switch 57, and the rearrangement unit 58 can be realized.

  Although the above embodiments have been described as representative examples, it will be apparent to those skilled in the art that many changes and substitutions can be made within the spirit and scope of the invention. Therefore, the present invention should not be construed as being limited by the above-described embodiments, and various modifications and changes can be made without departing from the scope of the claims. For example, when performing intra prediction, prediction based on the above-described method and conventional MPEG-4 AVC / H. The prediction by the H.264 system may be switched, or a prediction image by both methods may be generated to determine and output a prediction image with good coding efficiency.

  As described above, according to the present invention, it is possible to perform intra prediction suitable for various images, which is useful for any application for encoding or decoding an image.

DESCRIPTION OF SYMBOLS 1 Intra prediction apparatus 11 Processing block production | generation part 12 Reference image holding | maintenance part 13 Orthogonal transformation base provision part 14 Orthogonal transformation base decomposition | disassembly part 15 Subtransformation base holding part 16 Main conversion base holding part 17 Main conversion base correction | amendment part 18 Correction | amendment main conversion base holding Unit 19 inner product calculation unit 20 prediction image generation unit 21 reference image correction unit 22 repetition setting unit 23 parameter transmission unit 24 parameter reception unit 2 encoder 3 decoder

Claims (8)

An intra prediction device that generates a prediction image of a prediction block from a reference image of a decoded reference block adjacent to the prediction block,
Processing block generating means for generating a processing block including a reference block and a prediction block;
A first transformed basis vector obtained by extracting an orthogonal transformation basis vector corresponding to the entire area of the processing block according to the shape of the reference block, and a second obtained by extracting the orthogonal transformation basis vector according to the shape of the prediction block. Conversion basis generation means for generating a conversion basis vector of
The reference image is obtained by performing a weighted linear sum operation using a transform coefficient sequence corresponding to the transform basis vector corresponding to the predetermined order as a weight for the transform basis vector corresponding to the predetermined order in the first transform basis vector. A reference image approximation means for approximating a vector with a vector,
Predicted image generating means for generating a predicted image by performing a weighted linear sum operation using the conversion coefficient sequence used in the approximation as a weight for the conversion base vectors of a predetermined order among the second conversion base vectors. When,
An intra prediction apparatus comprising:   2. The intra prediction according to claim 1, wherein the reference image approximation unit obtains the transform coefficient sequence using a third transform basis vector obtained by correcting a norm with respect to the first transform basis vector. apparatus.   3. The intra of claim 2, wherein the third transformed basis vector is a vector obtained by dividing the first transformed basis vector by the square of the norm of the vector and correcting the norm. Prediction device. The reference image approximating means calculates a conversion coefficient for the conversion base vector by performing an inner product operation of the conversion base vector and the vector of the reference image in order from a low-order conversion base vector of the third conversion base vectors. Asking
A conversion coefficient of the next order is obtained by using a vector obtained by subtracting a vector obtained by multiplying the conversion base vector of the first conversion base vector and the conversion base vector of the order from the vector of the reference image as a new reference image vector. ,
4. The intra prediction apparatus according to claim 2, wherein the transform coefficient sequence is determined by repeating the process of obtaining the transform coefficient a predetermined number of times. 5.   A plurality of parameters including the configuration information of the reference block and the prediction block and the information of the predetermined order are prepared in advance, a plurality of prediction images are generated using the plurality of parameters, and 1 is generated from the plurality of prediction images. 5. The intra prediction apparatus according to claim 1, further comprising a parameter transmission unit that selects one prediction image and transmits a parameter used to generate the selected prediction image. 6.   An encoder comprising the intra prediction apparatus according to any one of claims 1 to 5.   A decoder comprising the intra prediction device according to any one of claims 1 to 4. A computer functioning as an intra prediction device that generates a predicted image of a prediction block from an image of a decoded reference block adjacent to the prediction block,
(A) generating a processing block comprising a reference block and a prediction block;
(B) A first transformed basis vector obtained by extracting an orthogonal transformation basis vector corresponding to the entire area of the processing block according to the shape of the reference block, and an orthogonal transformation basis vector extracted according to the shape of the prediction block. Generating a second transformed basis vector,
(C) performing a weighted linear sum operation on the conversion basis vectors of a predetermined order among the first conversion basis vectors, with a weighting of a conversion coefficient sequence corresponding to the conversion basis vector of the predetermined order, Approximating the reference image with a vector;
(D) A step of generating a prediction image by performing a weighted linear sum operation using the transform coefficient sequence used for the approximation as a weight for the transform basis vectors of a predetermined order among the second transform basis vectors. When,
A program for running

Images