JP2007096479A - Inter-layer prediction encoding method and apparatus, inter-layer prediction decoding method and apparatus, and program and recording medium thereof - Google Patents

Abstract

In an inter-layer predictive coding method, coding efficiency is improved by dynamically changing a DC block size to be subjected to Hadamard transform.
A Hadamard mode setting unit 204 sets a Hadamard mode according to a magnitude relationship between a DC block size of a current layer and a DC block size of a lower layer of an orthogonal transformation result. The Hadamard transform selection unit 206 causes the Hadamard transform units 207 to 209 to perform Hadamard transform according to the set DC block size. Information indicating the set Hadamard mode is added as additional information to the encoding information of the transform coefficient.
[Selection] Figure 11

Description

  The present invention relates to a high-efficiency image signal coding method, and in particular, in inter-layer prediction coding, intra-frame prediction is performed on frames of each layer, orthogonal transformation is performed on a prediction error, and a DC component of the obtained orthogonal transformation is obtained. The present invention relates to an inter-layer prediction processing method that performs Hadamard transform on.

  In recent years, scalable coding to deal with diversifying network environments and terminal environments has attracted attention.

  As an encoding method corresponding to time, space, and SNR scalability, Scalable Video Coding (SVC) has attracted attention. The Joint Scalable Video Model (JSVM) shown in Non-Patent Document 1 is a coding method based on AVC. One-way prediction in the time direction, two-way prediction, spatial prediction in a frame, and lower layer interpolation. Inter-layer prediction using signals is used. JSVM follows the AVC method for orthogonal transform, quantization, and entropy coding for prediction errors. For this reason, in the case of 16 × 16 intra prediction, the direct current component obtained by orthogonal transform with respect to the prediction error is subjected to Hadamard transform for each 4 × 4 component.

  FIG. 14 is a diagram illustrating an example of block division in JSVM. As shown in FIG. 14A, in the case of 16 × 16 intra prediction, a macroblock is divided into 16 macroblocks of 4 × 4 pixels. When orthogonal transformation is applied to each divided 4 × 4 pixel block, a direct current component appears in the hatched portion in FIG. 14B. Therefore, these direct current components are combined to form a direct current block of 4 × 4 pixels. This is the target of Hadamard transform.

Hereinafter, the block size (4 × 4 in the above example) of the Hadamard transform for the DC component is referred to as a DC block size. If the direct current component is generated by orthogonal transformation of 4 × 4 size, the direct current component is extracted from 16 × 16 elements. In this way, a block (16 × 16 in the above example) in the transform coefficient region to which the DC component within the DC block size belongs is called a Hadamard block.
J. Reichel, M. Wien and H. Schwarz, “Joint Scalable Video Model (JSVM) 2.0 Reference Encoding Algorithm Description”, ISO / IEC JTC 1 / SC 29 / WG 11N 7084, April 2005, Buzan, Korea

  The Hadamard transform for the DC component described above is a process performed for the purpose of removing the correlation between the DC components. Here, it should be noted that when the angle of view is the same, the range showing the same degree of correlation differs depending on the resolution of the image. For example, in a high-definition video such as 4096 × 2180 pixels and QCIF (176 × 144 pixels), if the angle of view is the same, the range showing the same degree of correlation is likely to be different. However, in JSVM, the DC block size is fixed at 4 × 4 as shown in FIG. 14, leaving room for improvement. Therefore, a method for dynamically changing the DC block size according to the resolution of the image is studied. As a result, further improvement in coding efficiency can be expected.

  The present invention has been made in view of such circumstances, and an object of the present invention is to establish an efficient encoding method for a DC component after orthogonal transform in scalable encoding.

In the following, the coordinates in the frame of time t of the j hierarchical (x, y) the pixel value at the _{f j (x, y, t} ), f j (x, y, t) of the decoded signal for g _j ( x, y, t). f _{j + 1} (x, y, t) has a spatial resolution twice that of f _j (x, y, t). For example, if f ₀ (x, y, t) is a QCIF size, f ₁ (x, y, t) is a CIF size.

Let the size of the Hadamard transform in the frame at time t in the j-th layer be L _j (t) × L _j (t). In the present invention, a mechanism for predicting the size of the Hadamard transform of the (j + 1) th layer from the size of the Hadamard transform of the jth layer is introduced.

  That is, the inter-layer predictive encoding method of the present invention divides an image signal to be encoded into layers having different spatial resolutions, performs intra-frame prediction on frames of each layer, performs orthogonal transformation on prediction errors, In the encoding method in which the Hadamard transform is performed on the DC component of the obtained orthogonal transform, the size of the Hadamard transform is adaptively changed, and mode information indicating the size of the Hadamard transform is added to the encoded information as additional information. It is characterized by.

  In addition, the inter-layer prediction encoding method of the present invention divides an image signal to be encoded into layers having different spatial resolutions, performs intra-frame prediction on frames of each layer, performs orthogonal transformation on prediction errors, In the encoding method for performing the Hadamard transform on the obtained DC component of the orthogonal transform, the Hadamard mode is set according to the magnitude relationship between the DC block size of the current layer and the DC block size of the lower layer of the orthogonal transform result. The information indicating the set Hadamard mode is added to the encoded information as additional information.

  Also, the inter-layer prediction image decoding method of the present invention, when decoding a signal obtained by dividing an image signal into layers having different spatial resolutions and performing inter-layer prediction encoding, predictive error signals for intra-frame prediction for frames of each layer. In the image decoding method that performs inverse Hadamard transform to obtain the DC component of orthogonal transform for, set the size of inverse Hadamard transform based on the additional information of the coding information, and perform inverse Hadamard transform according to the set size It is characterized by that.

  Also, the inter-layer prediction image decoding method of the present invention, when decoding a signal obtained by dividing an image signal into layers having different spatial resolutions and performing inter-layer prediction encoding, predictive error signals for intra-frame prediction for frames of each layer. In the image decoding method that performs inverse Hadamard transform to obtain the DC component of orthogonal transform for, the size of inverse Hadamard transform is set based on the size of the additional information and the inverse Hadamard transform of the lower layer, and according to the set size Inverse Hadamard transform is performed.

  The present invention can improve coding efficiency for prediction errors in intra prediction coding by adaptively changing the size of Hadamard transform for each layer having different spatial resolutions in scalable coding. . The present invention can be said to be an approach in which the coding efficiency is improved by utilizing the correlation between layers, with respect to the intra prediction coding having a low coding efficiency as compared with the inter prediction coding.

  1 to 3 are diagrams showing the relationship between the size of Hadamard transform between layers and the DC block size in the embodiment of the present invention.

  In the present embodiment, the DC block size is dynamically changed according to the frame resolution in each layer. At this time, the DC block size of the encoding target layer is expressed using the DC block size of the lower layer. This will be described with reference to the examples of FIGS. In any of the figures, orthogonal transformation is performed for each rectangular area indicated by a broken line. The size of the square is s × s (s = 4, 8). The size of the rectangular area indicated by the solid line is the size of the Hadamard transform.

  FIG. 1 shows an example in which the size of the Hadamard transform does not change in the upper and lower layers. In the figure, the size of Hadamard transform is 4 s × 4 s and the corresponding DC block size is 4 × 4 in both the j-th layer and the j + 1-th layer. In this case, the Hadamard mode is added to the encoded information as “1”. On the other hand, on the decoding side, when “1” is received as the Hadamard mode, the size of the Hadamard transform and the DC block size of the decoding target layer are set to the same size as the same size of the lower layer.

  FIG. 2 shows an example in which the size of the Hadamard transform increases in the upper layer. In the figure, the sizes of the Hadamard transforms of the j-th layer and the j + 1-th layer are 4s × 4s (j-th layer) and 8s × 8s (j + 1-th layer), respectively, and the corresponding DC block size is 4 × 4 (th-th layer). j hierarchy), 8 × 8 (j + 1 hierarchy). In this case, the Hadamard mode is set to “0” and added to the encoded information. On the other hand, on the decoding side, when “0” is received as the Hadamard mode, the size of the Hadamard transform and the DC block size of the decoding target layer are set to twice the same size of the lower layer. When the range showing the same degree of correlation expands in proportion to the resolution of the image, the use of this mode can improve the coding efficiency.

  FIG. 3 shows an example in which the size of Hadamard transform is reduced in the upper layer. In the figure, the sizes of the Hadamard transforms of the j-th layer and the j + 1-th layer are 4s × 4s (j-th layer) and 2s × 2s (j + 1-th layer), respectively, and the corresponding DC block size is 4 × 4 (th-layer). j hierarchy), 2 × 2 (j + 1 hierarchy). In this case, the Hadamard mode is added to the encoded information as “2”. On the other hand, on the decoding side, when “2” is received as the Hadamard mode, the size of the Hadamard transform and the DC block size of the decoding target layer are set to ½ times the same size of the lower layer. Use of this mode can be expected to improve the coding efficiency when fine textures and the like become apparent as the resolution increases and it is necessary to divide the area into smaller areas.

In the present embodiment, it is assumed that the above three modes are selected in the encoding of each layer. The method of selection shall be given separately. For example, the Lagrangian cost function
J = D + λR
To select the mode that minimizes the cost. Here, D is a distortion amount (for example, square error sum) after quantization of the Hadamard transform coefficient, and R is a code amount of the Hadamard transform coefficient after quantization. Λ is a mode given according to the value of the quantization parameter or based on a preliminary experiment. In this embodiment, three types of modes are used. However, in the embodiment of the present invention, only two types of modes, for example, Hadamard mode 0 and Hadamard mode 1, may be used.

FIG. 4 shows the definition of Hadamard mode in this embodiment. The Hadamard mode and its conditions are as follows.
Hadamard mode 0: The DC block size of the current layer is twice the DC block size of the lower layer.
Hadamard mode 1: The DC block size of the current layer is equal to the DC block size of the lower layer.
Hadamard mode 2: The DC block size of the current layer is ½ times the DC block size of the lower layer.

  FIG. 5 shows an example in which a different Hadamard mode is specified for each slice. As shown in FIG. 5, the size of the Hadamard transform can be changed in units of slices according to the local nature in the frame. It is also possible to preset a minimum value of the DC block size. For example, when the minimum value of the DC block size is set to 4 × 4, a DC block size of 4 × 4 or more is always set.

[Encoding process]
The present invention is used in SVC intra coding. FIG. 6 shows the flow of the encoding process according to the embodiment of the present invention. Hereinafter, the flow of the encoding process of the present embodiment will be described with reference to FIG.
[Step S1]: The prediction error is read from the register. If the prediction error is not stored in the register, this process is skipped.
[Step S2]: The prediction error read in step S1 is input, orthogonal transform is performed, and a transform coefficient is output.
[Step S3]: The conversion coefficient calculated in Step S2 is input, and a determination process is performed to determine whether the conversion coefficient is a DC component. If true, the process proceeds to Step S5. If false, Step S4 is performed. Move on to processing.
[Step S4]: If the conversion coefficient is not a DC component, the conversion coefficient is written to the register, and the process proceeds to Step S7.
[Step S5]: The direct current component of the transform coefficient is input, Hadamard transform processing is performed, and the Hadamard transform coefficient is calculated.
[Step S6]: The Hadamard transform coefficient calculated in step S5 is written to the register.
[Step S7]: The transform coefficient and Hadamard transform coefficient written in the register in steps S4 and S6 are input, quantization processing is performed, and the quantized value is written in the register.
[Step S8]: The quantized value output in step S7 is input, local decoding is performed, and the decoded signal is written to the register. The details of this process are the same as step S32 in FIG. 8, and will be described later with reference to FIG.
[Step S9]: The local decoded signal and the encoding target signal output in step S8 are input, intra prediction is performed, and a prediction signal is calculated.
[Step S10]: A prediction error and an encoding target signal are input, a prediction error is calculated, and the calculated prediction error is written to a register.
[Step S11]: The prediction error output in step S10 is input, entropy encoding is performed, and the encoding result is written to a register.
[Step S12]: The above steps S1 to S11 are performed for all macroblocks.

Next, details of the Hadamard transform process in step S5 in FIG. 6 will be described with reference to FIG. FIG. 7 shows the flow of Hadamard transform processing according to the embodiment of the present invention.
[Step S21]: A Hadamard transform size and a DC block size are set. It should be noted that a specific setting method is given from the outside.
[Step S22]: The DC block size set in step S21 is input, and a Hadamard transform conversion formula corresponding to the DC block size is read.
[Step S23]: The Hadamard transform conversion equation set in Step S22 and the prediction error are input, Hadamard transform is performed on the prediction error, and the transform coefficient is written to the register.
[Step S24]: On the other hand, the DC block size of the lower layer is read.
[Step S25]: Input the DC block size of the lower layer and the DC block size of the current layer, compare both, and check whether the DC block size of the current layer is twice the DC block size of the lower layer If it is true, the process proceeds to step S29. If it is false, the process proceeds to step S26.
[Step S26]: The DC block size of the lower layer and the DC block size of the current layer are input, and comparison processing is performed between them to determine whether the DC block size of the current layer and the DC block size of the lower layer are equal. If true, the process proceeds to step S28. If false, the process proceeds to step S27.
[Step S27]: If the DC block size of the current layer is smaller than the DC block size of the lower layer, Hadamard mode is set to 2 and written to the register.
[Step S28]: If the DC block size of the current layer is equal to the DC block size of the lower layer, the Hadamard mode is set to 1 and written to the register.
[Step S29]: If the DC block size of the current layer is larger than the DC block size of the lower layer, the Hadamard mode is set to 0 and written to the register.

[Decryption process]
The decoding process of the present invention is used when decoding an SVC intra-coded frame. FIG. 8 shows a flow of decoding processing according to the embodiment of the present invention. Hereinafter, the flow of the decoding process of the present embodiment will be described with reference to FIG.
[Step S31]: An encoded bit stream is input, entropy decoding is performed, and the decoded quantized value is written to a register.
[Step S32]: A local decoding process including the following steps S33 to S38 is executed.
[Step S33]: The quantized value output in step S31 is input, inverse quantization processing is performed, and the transform coefficient is written to the register.
[Step S34]: Using the transform coefficient output in step S33 as input, it is determined whether or not the transform coefficient is a Hadamard transform coefficient for the DC component of the orthogonal transform. If true, the process proceeds to step S36. If false, the process proceeds to step S35.
[Step S35]: If the transform coefficient is not a Hadamard transform coefficient for the DC component, the transform coefficient (decoded AC coefficient) is written to the register, and the process proceeds to step S38.
[Step S36]: When the transform coefficient is a Hadamard transform coefficient with respect to a DC component, the transform coefficient is input, an inverse Hadamard transform process is performed, and a DC component of orthogonal transform is calculated.
[Step S37]: The decoded DC coefficient of the DC component calculated in step S36 is input and written to the register.
[Step S38]: The transform coefficient of the orthogonal transform is input from steps S35 and S37, the inverse orthogonal transform process is performed, the decoding prediction error is calculated, and written to the register.
[Step S39]: The decoded signal to be referred to at the time of prediction is input, intra prediction processing is performed, and the prediction signal is written to the register.
[Step S40]: The decoding prediction error calculated in step S38 and the prediction signal calculated in step S39 are input, and both are added to obtain a decoded signal, and the decoded signal is written to the register.
[Step S41]: The above steps S31 to S40 are repeated for all macroblocks.

The flow of the inverse Hadamard transform process in step S36 in FIG. 8 will be described with reference to FIG. FIG. 9 shows the flow of inverse Hadamard transform processing according to the embodiment of the present invention.
[Step S51]: Read Hadamard mode.
[Step S52]: The DC block size of the lower layer is read.
[Step S53]: The Hadamard mode and the DC block size of the lower layer are input, and a determination process is performed to determine whether the Hadamard mode is 0. If true, the process proceeds to Step S57. If false, the process proceeds to Step S54. Move on.
[Step S54]: The Hadamard mode and the DC block size of the lower layer are input, and it is determined whether or not the Hadamard mode is 1. If true, the process proceeds to Step S56, and if false, the process of Step S55 is performed. Move on.
[Step S55]: When the Hadamard mode is 2, a value obtained by halving the DC block size of the lower layer is written in the register as the DC block size of the current layer.
[Step S56]: When the Hadamard mode is 1, the DC block size of the lower layer is written to the register as the DC block size of the current layer.
[Step S57]: When the Hadamard mode is 0, a value obtained by doubling the DC block size of the lower layer is written to the register as the DC block size of the current layer.
[Step S58]: The DC block size output in steps S55, S56, and S57 is input, and an inverse Hadamard transform equation corresponding to the DC block size is selected.
[Step S59]: The direct current component included in the direct current block is input, the inverse Hadamard transform process is performed using the inverse Hadamard transform formula selected in Step S58, and the conversion result is written to the register.

[Encoder]
FIG. 10 shows a configuration example of the encoding device according to the embodiment of the present invention. Hereinafter, the configuration of the encoding apparatus of the present embodiment will be described with reference to FIG.

  The orthogonal transform processing unit 101 receives a prediction error that is a difference between the encoded signal and the prediction signal, performs orthogonal transform, divides the transform coefficient into an AC component and a DC component, and each AC component storage unit 102 , And write to the DC component storage unit 103. The Hadamard transform processing unit 104 receives the DC component read from the DC component storage unit 103 as input, performs Hadamard transform processing, and writes the transform coefficient in the transform coefficient storage unit 105. Details of the Hadamard transform processing unit 104 will be described later with reference to FIG.

  The quantization processing unit 106 receives the AC component read from the AC component storage unit 102 and the Hadamard transform coefficient read from the transform coefficient storage unit 105 as input, performs quantization processing, and writes the quantization value to the quantization storage unit 107. . The entropy encoding unit 108 receives the quantized value read from the quantization storage unit 107 as input, performs entropy encoding processing, and writes the encoded result to the encoded bit storage unit 109.

  On the other hand, the local decoding processing unit 110 receives the quantized value read from the quantization storage unit 107, performs local decoding processing, and writes the obtained local decoded signal to the decoded signal storage unit 111. The local decoding processing unit 110 is the same as the local decoding processing unit 313 in FIG. 12, and details will be described in the description of FIG. The intra-screen prediction processing unit 112 receives the local decoded signal read from the decoded signal storage unit 111 as input, performs an intra-screen prediction process, and writes the prediction signal to the prediction signal storage unit 113.

  Next, details of the Hadamard transform processing unit 104 in FIG. 10 will be described with reference to FIG. FIG. 11 shows a configuration example of the Hadamard transform processing unit according to the embodiment of the present invention.

  The DC block size setting unit 201 sets a DC block size and writes it to the DC block size storage unit 202. The DC block size may be given in advance from the outside, or may be adaptively determined by predicting / estimating the amount of generated code. The lower layer DC block size storage unit 203 stores the written lower layer DC block size.

  The Hadamard mode setting unit 204 inputs the current block DC block size read from the DC block size storage unit 202 and the lower layer DC block size storage unit 203 and the DC block size of the lower layer, and sets the Hadamard mode. Processing is performed and the result is written in the Hadamard mode storage unit 205. The Hadamard mode is set according to the definition shown in FIG.

  The Hadamard transform selection unit 206 receives the DC block size read from the DC block size storage unit 202 as input, performs a process of selecting the Hadamard transform corresponding to the DC block size, and has a corresponding size Hadamard transform unit 207, 208, 209. To perform Hadamard transform processing and calculate Hadamard transform coefficients.

[Decoding device]
FIG. 12 shows a configuration example of a decoding device according to the embodiment of the present invention. Hereinafter, the configuration of the decoding apparatus of the present embodiment will be described with reference to FIG.

  The coded bit storage unit 301 stores the written coded bit stream. The entropy decoding unit 302 receives the encoded bit stream read from the encoded bit storage unit 301 as input, performs entropy decoding processing, and writes the calculated quantization value to the quantization storage unit 303. Note that the following inverse quantization processing unit 304 to inverse orthogonal transform processing unit 309 are the local decoding processing unit 313.

  The inverse quantization processing unit 304 receives the quantized value read from the quantization storage unit 303, performs inverse quantization processing, writes Hadamard transform coefficients for the DC component of orthogonal transform to the decoded transform coefficient storage unit 305, and performs orthogonal processing. The AC component of the conversion is written in the decoded AC component storage unit 308.

  The inverse Hadamard transform processing unit 306 receives the Hadamard transform coefficient for the DC component of the orthogonal transform read from the decoded transform coefficient storage unit 305 as input, performs the inverse Hadamard transform process, and stores the obtained DC component of the orthogonal transform in the decoded DC component storage. Write to part 307. Details of the inverse Hadamard transform processing unit 306 will be described later with reference to FIG.

  The inverse orthogonal transform processing unit 309 receives the DC component read from the decoded DC component storage unit 307 and the AC component read from the decoded AC component storage unit 308, performs inverse orthogonal transform processing, and obtains the obtained decoding prediction error. Write to the decoded signal storage unit 310. The intra-screen prediction processing unit 311 receives the decoded prediction error read from the decoded signal storage unit 310, performs an intra-screen prediction process, and writes the obtained prediction signal to the prediction signal storage unit 312.

  The inverse Hadamard transform processing unit 306 in FIG. 12 will be described with reference to FIG. FIG. 13 shows a configuration example of the inverse Hadamard transform processing unit according to the embodiment of the present invention.

  The Hadamard mode storage unit 401 stores the written Hadamard mode. The lower layer DC block size storage unit 402 stores the written lower layer DC block size.

  The DC block size setting unit 403 receives the Hadamard mode read from the Hadamard mode storage unit 401 and the lower layer DC block size read from the lower layer DC block size storage unit 402, and performs a process for obtaining the DC block size. The obtained DC block size is written in the DC block size storage unit 405. The DC block size is set according to the definition shown in FIG.

  The Hadamard transform selection unit 406 receives the DC block size read from the DC block size storage unit 405 as input, performs a process of selecting an inverse Hadamard transform corresponding to the DC block size, and performs inverse Hadamard transform units 407 and 408 having the corresponding sizes. , 409. The selected inverse Hadamard transform units 407, 408, and 409 perform inverse Hadamard transform processing and calculate a decoded signal of the DC component.

  In the above embodiment, the example in which the Hadamard mode is defined and set according to the magnitude relationship between the DC block size of the current hierarchy and the DC block size of the lower hierarchy as a result of the orthogonal transformation has been described. Independent of the size, the size of the Hadamard transform is adaptively converted based on prediction and estimation of the generated code amount and the distortion amount after quantization of the Hadamard transform coefficient, and the size information is encoded as mode information. Implementation in addition to information is also possible. In this case, the size of the inverse Hadamard transform is set from the decoding result of the mode information on the decoding side.

  The above inter-layer prediction encoding and decoding processes can be realized by a computer and a software program, and the program can be provided by being recorded on a computer-readable recording medium or via a network. It is.

It is a figure which shows the example in case the size of Hadamard transform does not change in the upper and lower layers. It is a figure which shows the example in case the size of a Hadamard transform expands in an upper hierarchy. It is a figure which shows the example in case the size of a Hadamard transform reduces in an upper hierarchy. It is a definition explanatory view of Hadamard mode. It is a figure which shows the example at the time of designating different Hadamard modes for every slice. It is a figure which shows the flow of the encoding process which concerns on embodiment of this invention. It is a figure which shows the flow of the Hadamard transformation process which concerns on embodiment of this invention. It is a figure which shows the flow of the decoding process which concerns on embodiment of this invention. It is a figure which shows the flow of the inverse Hadamard transformation process which concerns on embodiment of this invention. It is a figure which shows the structural example of the encoding apparatus which concerns on embodiment of this invention. It is a figure which shows the structural example of the Hadamard transformation process part which concerns on embodiment of this invention. It is a figure which shows the structural example of the decoding apparatus which concerns on embodiment of this invention. It is a figure which shows the structural example of the inverse Hadamard transformation process part which concerns on embodiment of this invention. It is a figure which shows the example of the block division in JSVM.

Explanation of symbols

DESCRIPTION OF SYMBOLS 101 Orthogonal transformation process part 102 AC component memory | storage part 103 DC component memory | storage part 104 Hadamard transform process part 105 Transform coefficient memory | storage part 106 Quantization process part 107 Quantization memory | storage part 108 Entropy encoding part 109 Encoding bit memory | storage part 110 Local decoding process Unit 111 decoded signal storage unit 112 intra prediction processing unit 113 prediction signal storage unit 201 DC block size setting unit 202 DC block size storage unit 203 lower hierarchy DC block size storage unit 204 Hadamard mode setting unit 205 Hadamard mode storage unit 206 Hadamard transform Selection unit 207, 208, 209 Hadamard transform unit 301 Encoded bit storage unit 302 Entropy decoding unit 303 Quantization storage unit 304 Inverse quantization processing unit 305 Decoded transform coefficient storage unit 306 Inverse Hadamard transform processing unit 307 Decoded DC component storage unit 308 Decoded AC component storage unit 309 Inverse orthogonal transform processing unit 310 Decoded signal storage unit 311 In-screen prediction processing unit 312 Prediction signal storage unit 313 Local decoding processing unit 401 Hadamard mode storage unit 402 Lower hierarchical DC block size storage 403 DC block size setting unit 405 DC block size storage unit 406 Hadamard transform selection unit 407, 408, 409 Inverse Hadamard transform unit

Claims (12)

The image signal is divided into layers with different spatial resolutions, intra-frame prediction is performed on the frames of each layer, orthogonal transformation is performed on the prediction error, and the Hadamard transform is performed on the DC component of the obtained orthogonal transformation. In the inter prediction coding method,
Setting the size of the Hadamard transform;
Encoding mode information indicating the size of the set Hadamard transform and adding to the encoded information,
An inter-layer predictive coding method characterized by adaptively changing the size of a Hadamard transform. The image signal is divided into layers with different spatial resolutions, intra-frame prediction is performed on the frames of each layer, orthogonal transformation is performed on the prediction error, and the Hadamard transform is performed on the DC component of the obtained orthogonal transformation. In the inter prediction coding method,
Setting Hadamard mode according to the magnitude relationship between the DC block size of the current layer and the DC block size of the lower layer of the orthogonal transformation result;
Encoding information indicating the set Hadamard mode and adding the encoded information to the encoded information,
The DC block size is dynamically changed. An inter-layer predictive encoding method characterized by: The image signal is divided into layers with different spatial resolutions, intra-frame prediction is performed on the frames of each layer, orthogonal transformation is performed on the prediction error, and the Hadamard transform is performed on the DC component of the obtained orthogonal transformation. In the inter prediction encoding device,
Means for setting the size of the Hadamard transform;
Means for encoding mode information indicating the size of the set Hadamard transform and adding it to the encoded information;
An inter-layer prediction encoding apparatus characterized by adaptively changing the size of a Hadamard transform. The image signal is divided into layers with different spatial resolutions, intra-frame prediction is performed on the frames of each layer, orthogonal transformation is performed on the prediction error, and the Hadamard transform is performed on the DC component of the obtained orthogonal transformation. In the inter prediction encoding device,
Means for setting the Hadamard mode according to the magnitude relationship between the DC block size of the current layer and the DC block size of the lower layer of the orthogonal transformation result;
Means for encoding information indicating the set Hadamard mode and adding it to the encoded information;
The DC block size is dynamically changed. An inter-layer predictive coding apparatus characterized by: In the inter-layer predictive decoding method for decoding the encoded information encoded by the inter-layer predictive encoding method according to claim 1,
Setting the size of the inverse Hadamard transform based on the mode information included in the encoding information to be decoded;
A method of performing inter-layer prediction decoding, comprising: performing inverse Hadamard transform to obtain an orthogonal transform DC component for a prediction error signal of intra-frame prediction for each frame according to the size. In the inter-layer predictive decoding method for decoding the encoded information encoded by the inter-layer predictive encoding method according to claim 2,
Setting the size of the inverse Hadamard transform based on the information indicating the Hadamard mode included in the encoded information to be decoded and the size of the inverse Hadamard transform in the lower layer;
A method of performing inter-layer prediction decoding, comprising: performing inverse Hadamard transform to obtain an orthogonal transform DC component for a prediction error signal of intra-frame prediction for each frame according to the size. In the inter-layer prediction decoding apparatus for decoding the encoded information encoded by the inter-layer prediction encoding apparatus according to claim 3,
Means for setting the size of the inverse Hadamard transform based on the mode information included in the encoding information to be decoded;
Means for performing inverse Hadamard transform in order to obtain a DC component of orthogonal transform for a prediction error signal of intra prediction for a frame of each layer according to the size. In the inter-layer prediction decoding apparatus for decoding the encoded information encoded by the inter-layer prediction encoding apparatus according to claim 4,
Means for setting the size of the inverse Hadamard transform based on the information indicating the Hadamard mode included in the encoded information to be decoded and the size of the inverse Hadamard transform in the lower layer;
Means for performing inverse Hadamard transform in order to obtain a DC component of orthogonal transform for a prediction error signal of intra prediction for a frame of each layer according to the size.   An inter-layer prediction encoding program for causing a computer to execute the inter-layer prediction encoding method according to claim 1 or 2.   An inter-layer predictive decoding program for causing a computer to execute the inter-layer predictive decoding method according to claim 5 or 6.   A computer-readable recording medium on which an inter-layer prediction encoding program for causing a computer to execute the inter-layer prediction encoding method according to claim 1 or 2 is recorded.   A computer-readable recording medium storing an inter-layer predictive decoding program for causing a computer to execute the inter-layer predictive decoding method according to claim 5 or 6.

Images