JP2012134869A - Image encoding method, image decoding method, image encoding apparatus and image decoding apparatus - Google Patents

Abstract

PROBLEM TO BE SOLVED: To realize an intra-prediction encoding method which reduces intra-prediction errors for increased efficiency.SOLUTION: An intra-prediction processing unit 101 generates an intra-prediction residual signal. A conversion processing unit 102 exerts conversion processing on the intra-prediction residual signal to generate a conversion coefficient, and an inverse conversion processing unit 105 exerts inverse conversion processing on a decode value of the conversion coefficient to generate a decode value of the intra-prediction residual signal. A residual prediction processing unit 110 generates prediction processing for the intra-prediction residual signal, and an adder A3 adds a decode signal of a residual prediction error signal and a prediction value of the intra-prediction residual signal together to generate a decode value of the intra-prediction residual signal. Unless residual prediction is performed, the residual prediction processing unit 110 is not used, while if residual prediction is performed, a decode signal is generated by the residual prediction processing unit 110 using a residual prediction error signal which has had residual prediction made.

Description

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

  The moving image coding algorithms are roughly classified into interframe coding (inter coding) and intraframe coding (intra coding). Interframe coding is an approach for compressing information by using the correlation in the time direction in a moving image. A typical example is inter-frame prediction using motion compensation. On the other hand, intra-frame coding is an approach for compressing information by using intra-frame correlation. In JPEG and MPEG-2, an approach using discrete cosine transform (DCT) was adopted, and in JPEG 2000, an approach using discrete wavelet transform was adopted. H. In H.264 / AVC, prediction in the spatial direction is performed in addition to the above-described transform coding (see, for example, Non-Patent Document 1).

  In general, the coding efficiency of intra coding is lower than that of inter coding, and further improvement in coding efficiency is required. H. As an approach for improving the H.264 / AVC intra prediction, a method that subdivides eight prediction directions and supports 33 prediction directions has been proposed. This method aims to reduce the prediction error due to the coarseness of the granularity in the prediction direction.

  FIG. 11 is a block diagram illustrating a configuration of an encoding device using conventional intra prediction. The encoding device includes an intra prediction processing unit 101, a transformation processing unit 102, a quantization processing unit 103, an inverse quantization processing unit 104, an inverse transformation processing unit 105, an in-loop filter processing unit 106, a frame memory 107, an intra prediction information code. And an entropy encoding processing unit 109.

  The adder A1 generates a difference signal (intra prediction residual signal) between an intra prediction mode from the intra prediction processing unit 101 described later and an input signal. The conversion processing unit 102 performs conversion processing on the intra prediction residual signal and generates a conversion coefficient. The quantization processing unit 103 performs a quantization process on the transform coefficient and generates a quantization index.

  The inverse quantization processing unit 104 performs an inverse quantization process on the quantization index, and generates a decoded value of the transform coefficient. The inverse transform processing unit 105 performs an inverse transform process on the decoded value of the change coefficient to generate a decoded signal of the intra prediction residual signal. Adder A2 adds the intra prediction signal produced | generated by the intra prediction by the intra prediction process part 101 to the decoded value of an intra prediction residual signal, and produces | generates a decoded signal. The in-loop filter processing unit 106 performs filter processing on the decoded signal. The frame memory 107 holds the decoded signal after the filter processing.

  The intra prediction processing unit 101 sets an intra coding method (intra coding mode), and generates an intra prediction signal based on the set intra prediction mode. The intra prediction information encoding unit 108 encodes intra prediction information. The entropy encoding processing unit 109 performs entropy encoding on the quantization index.

  FIG. 12 is a flowchart for explaining the operation of a coding apparatus using conventional intra prediction. The following steps S101 to S105 are executed for each encoding unit. First, an intra coding scheme (intra coding mode) is set, an intra prediction signal is generated based on the set intra prediction mode, and a difference signal (intra prediction residual) between the intra prediction mode and the input signal is generated. Signal) is generated (step S102).

  Next, a conversion process is performed on the intra prediction residual signal to generate a conversion coefficient (step S103-1), a quantization process is performed on the conversion coefficient, and a quantization index is generated (step S103-2). Next, the quantization index is entropy-encoded (step S103-3), the inverse quantization process is performed on the quantization index, and a decoded value of the transform coefficient is generated (step S103-4). Then, an inverse transform process is performed to generate a decoded signal of the intra prediction residual signal (step S103-5). Hereinafter, the processing of steps S103-1 to S103-5 is referred to as step S103.

  Then, an intra prediction signal generated by intra prediction (generated in step S102) is added to the decoded value of the intra prediction residual signal to generate a decoded signal (step S104).

  FIG. 13 is a block diagram illustrating a configuration of a decoding device using conventional intra prediction. The decoding apparatus includes an entropy decoding processing unit 201, an inverse quantization processing unit 202, an inverse transformation processing unit 203, an intra prediction information storage unit 204, an intra prediction processing unit 205, an in-loop filter processing unit 206, and a frame memory 207.

  The entropy decoding processing unit 201 receives the encoded data of the encoding unit as input and performs entropy decoding processing. The inverse quantization processing unit 202 performs an inverse quantization process using the quantization index as an input, and generates a decoded value of the transform coefficient. The inverse transform processing unit 203 receives the decoded value of the transform coefficient as an input, performs an inverse transform process, and generates a decoded signal of the intra prediction residual signal.

  The intra prediction information storage unit 204 stores the intra prediction information (intra coding mode) decoded by the entropy decoding processing unit 201. The intra prediction processing unit 205 reads an intra coding scheme (intra coding mode) from the decoded encoded data, and generates an intra prediction signal based on the intra prediction mode. The adder B1 receives the decoded value of the intra prediction residual signal and the intra prediction signal, adds them, and generates a decoded signal. The in-loop filter processing unit 206 performs filter processing on the decoded signal. The frame memory 207 holds the decoded signal after the filter processing.

  FIG. 14 is a flowchart for explaining the operation of a decoding apparatus using conventional intra prediction. The processing of steps S201 to S206 is executed for each encoding unit. First, entropy decoding processing is performed with the encoded data of the encoding unit as input (step S202). Next, an intra coding method (intra coding mode) is read from the decoded encoded data, and an intra prediction signal is generated based on the intra prediction mode (step S203).

  Next, an inverse quantization process is performed using the quantization index as an input to generate a decoded value of the transform coefficient (step S204-1), an inverse transform process is performed using the decoded value of the transform coefficient as an input, and an intra prediction residual is obtained. A decoded signal of the difference signal is generated (step S204-2). Hereinafter, the processing of steps S204-1 and S204-2 is referred to as step S204.

  Then, the decoded value of the intra prediction residual signal and the intra prediction signal (generated in step S203) are input, and both are added to generate a decoded signal.

“H.264 / AVC textbook”, Takuya Kakuno et al., Impress, 2008.

  Since all the conventional intra predictions are extrapolation predictions, the prediction error tends to increase as the distance between the reference pixel and the pixel to be predicted increases. For this reason, although there is a difference in the granularity of the prediction direction, the problem inherent in such extrapolation prediction is unsolved, and there is a problem of leaving room for improvement in improving the coding efficiency.

  The present invention has been made in view of such circumstances, and its purpose is to focus on the correlation between intra prediction errors, and to establish a highly efficient intra coding method by reducing intra prediction errors. The present invention provides an image encoding method, an image decoding method, an image encoding device, and an image decoding device.

  The present invention is an image encoding method for generating a prediction signal using spatial inter-pixel prediction and generating a residual signal between the prediction signal and an original signal, and for generating an intra prediction residual signal. A prediction processing step, performing inter-transform processing on the intra prediction residual signal to generate transform coefficients, performing quantization processing and inverse quantization processing on the transform coefficients, and decoding the transform coefficients Generating a value, performing an inverse transform process on the decoded value to generate a decoded value of the intra prediction residual signal, a prediction direction of intra prediction, and a distance between a prediction target block and a neighboring block Based on an identification step for identifying a reference block from neighboring blocks and a decoded value of the intra prediction residual signal in the reference block, a prediction process for the intra prediction residual signal of the prediction target block Performed, characterized in that it comprises a residual prediction error signal generating step of generating residual prediction error signals.

  The present invention is an image encoding method for generating a prediction signal using spatial inter-pixel prediction and generating a residual signal between the prediction signal and an original signal, and for generating an intra prediction residual signal. A prediction processing step, performing inter-transform processing on the intra prediction residual signal to generate transform coefficients, performing quantization processing and inverse quantization processing on the transform coefficients, and decoding the transform coefficients A generation step for generating a value, an identification step for identifying a reference block from the neighboring block based on a prediction direction of intra prediction and a distance between the prediction target block and the neighboring block, and conversion of an intra prediction residual signal in the reference block Residue for generating a prediction error signal for the transform coefficient by performing a prediction process on the transform coefficient of the intra prediction residual signal of the prediction target block using the decoded value of the coefficient Characterized in that it comprises a measuring error signal generating step.

  The present invention is obtained by performing transform processing on the residual prediction error signal generated in the residual prediction error signal generating step to generate transform coefficients, and performing quantization processing on the transform coefficients. A first cost value generation step for obtaining a weighted sum of a quantization error due to the quantization processing and a code amount for expressing the first quantization transform coefficient as a cost value for one quantized transform coefficient; The second quantized transform coefficient obtained by performing transform processing on the intra prediction residual signal generated in the intra prediction processing step to generate a transform coefficient, and performing quantization processing on the transform coefficient A second cost value generating step for obtaining a weighted sum of a quantization error due to the quantization process and a code amount for expressing the second quantization transform coefficient as a cost value, and the first cost Value generation step The cost value obtained is compared with the cost value obtained in the second cost value generation step, and when the cost value obtained by the first cost value generation step is smaller, the intra prediction residual An intra prediction residual flag indicating that the prediction process is applied is output as encoded data. In other cases, an intra prediction residual flag indicating that the prediction process for the intra prediction residual is not applied is encoded data. An output step of outputting is included.

  The present invention is obtained by performing transform processing on the residual prediction error signal generated in the residual prediction error signal generating step to generate transform coefficients, and performing quantization processing on the transform coefficients. A first cost value generation step for obtaining a weighted sum of a quantization error due to the quantization processing and a code amount for expressing the first quantization transform coefficient as a cost value for one quantized transform coefficient; , The second quantization transform coefficient obtained by performing the quantization process on the intra prediction residual signal for the transform coefficient generated in the intra prediction processing step, and the quantization error due to the quantization process and the quantization error A second cost value generating step for obtaining a weighted sum of code amounts for expressing the second quantized transform coefficient as a cost value; a cost value obtained in the first cost value generating step; and the second cost When the cost value obtained in the first cost value generation step is smaller than the cost value obtained in the generation step, the intra prediction residual indicating that the prediction process for the intra prediction residual is applied. A difference flag is output as encoded data; otherwise, an output step of outputting an intra prediction residual flag indicating that the prediction process for the intra prediction residual is not applied as encoded data is included. .

  The present invention relates to an image decoding method for decoding an encoded stream output from an encoder, wherein an entropy decoding step for entropy decoding the encoded stream, and an intra prediction mode from the encoded data subjected to the entropy decoding process. Reading, based on the intra prediction mode, an intra prediction processing step for generating an intra prediction signal, and identifying a reference block from the neighboring block based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, Using the decoded value of the intra prediction residual signal in the reference block, a prediction process is performed on the intra prediction residual signal of the decoding target block to generate a prediction value of the intra prediction residual signal, and the quantization index An inverse quantization process is performed to generate a decoded value of the transform coefficient, and the transform coefficient Inverse transform processing is performed on the signal value to generate a decoded signal of the residual prediction error signal, the decoded signal of the residual prediction error signal and the prediction value of the intra prediction residual signal are added, and an intra prediction residual is obtained. A decoding step of an intra prediction residual signal for generating a decoded signal of the signal, a decoded signal of the intra prediction residual signal generated in the decoding step of the intra prediction residual signal, and an intra prediction generated in the intra prediction processing step. And a generation step of generating a decoded signal by adding the signals.

  The present invention is an image decoding method for decoding an encoded stream output from an encoder, wherein an entropy decoding step for entropy decoding the encoded stream, and reading an intra prediction mode from the encoded data that has been entropy decoded An intra prediction processing step for generating an intra prediction signal based on the intra prediction mode, and reading an intra prediction residual flag indicating whether or not a prediction process for the intra prediction residual is present from the entropy-decoded encoded data. When it is determined that the prediction residual flag indicates that the prediction process is not applied to the intra prediction residual, the quantized index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the transform coefficient A decoded signal of the intra prediction residual signal is generated by performing an inverse transformation process on the decoded value of A first decoding step of an intra prediction residual signal, and when it is determined that the intra prediction residual flag indicates application of prediction processing to the intra prediction residual, the prediction direction of the intra prediction and the prediction target block and the neighborhood Based on the distance between the blocks, a reference block is identified from neighboring blocks, and the intra prediction residual signal of the decoding target block is predicted using the decoded value of the intra prediction residual signal in the reference block. A prediction value of the prediction residual signal is generated, an inverse quantization process is performed on the quantization index to generate a decoded value of the transform coefficient, and an inverse transform process is performed on the decoded value of the transform coefficient to obtain a residual A decoded signal of the prediction error signal is generated, the decoded signal of the residual prediction error signal and the prediction value of the intra prediction residual signal are added, and the intra prediction residual signal is restored. A second decoding step of the intra prediction residual signal for generating a signal and a decoded signal of the intra prediction residual signal generated in the first decoding step of the intra prediction residual signal, or the intra prediction residual signal. And a generation step of generating a decoded signal by adding the decoded signal of the intra prediction residual signal generated in the second decoding step to the intra prediction signal generated in the intra prediction processing step. And

  The present invention relates to an image decoding method for decoding an encoded stream output from an encoder, wherein an entropy decoding step for entropy decoding the encoded stream, and an intra prediction mode from the encoded data subjected to the entropy decoding process. Reading, based on the intra prediction mode, an intra prediction processing step for generating an intra prediction signal, and identifying a reference block from the neighboring block based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, Using the decoded value of the transform coefficient in the reference block, the prediction process for the transform coefficient of the decoding target block is performed to generate the predicted value of the transform coefficient of the decoding target block, and the inverse quantization process is performed on the quantization index. To generate a decoded value of the residual prediction error signal for the transform coefficient, the residual The decoding value of the measurement error signal and the prediction value of the transform coefficient of the decoding target block are added to generate a decoded signal of the transform coefficient of the intra prediction residual signal, and the transform coefficient decoding step and the transform coefficient decoding step Generated in the decoding step of the intra prediction residual signal and the decoding step of the intra prediction residual signal for performing the inverse transform process on the decoded value of the transform coefficient to be generated to generate the decoded signal of the intra prediction residual signal It includes a generation step of adding a decoded signal of the intra prediction residual signal and the intra prediction signal generated in the intra prediction processing step to generate a decoded signal.

  The present invention is an image decoding method for decoding an encoded stream output from an encoder, wherein an entropy decoding step for entropy decoding the encoded stream, and reading an intra prediction mode from the encoded data that has been entropy decoded An intra prediction processing step for generating an intra prediction signal based on the intra prediction mode, and reading an intra prediction residual flag indicating whether or not a prediction process for the intra prediction residual is present from the entropy-decoded encoded data. When it is determined that the prediction residual flag indicates that the prediction process is not applied to the intra prediction residual, the quantized index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the transform coefficient A decoded signal of the intra prediction residual signal is generated by performing an inverse transformation process on the decoded value of A first decoding step of an intra prediction residual signal, and when it is determined that the intra prediction residual flag indicates application of prediction processing to the intra prediction residual, the prediction direction of the intra prediction and the prediction target block and the neighborhood Based on the distance between the blocks, the reference block is identified from the neighboring blocks, the prediction value for the transform coefficient of the decoding target block is predicted using the decoded value of the transform coefficient in the reference block, and the transform coefficient of the decoding target block is determined. Generate a prediction value, perform inverse quantization processing on the quantization index to generate a decoded value of the residual prediction error signal for the transform coefficient, and convert the decoded value of the residual prediction error signal and the block to be decoded A transform coefficient decoding step for adding a prediction value of the coefficient to generate a decoded signal of the transform coefficient of the intra prediction residual signal; and a decoding step of the transform coefficient A second decoding step of the intra prediction residual signal for performing a reverse transform process on the decoded value of the transform coefficient generated in step 1 to generate a decoded signal of the intra prediction residual signal, and the intra prediction residual signal A decoded signal of the intra prediction residual signal generated in the first decoding step, or a decoded signal of the intra prediction residual signal generated in the second decoding step of the intra prediction residual signal, and the intra prediction. And a generation step of generating a decoded signal by adding the intra-prediction signal generated in the processing step.

  The present invention is an image encoding device that generates a prediction signal using spatial inter-pixel prediction and generates a residual signal between the prediction signal and an original signal, and includes an intra prediction residual signal. A prediction processing means, performing inter-transform processing on the intra prediction residual signal to generate transform coefficients, performing quantization processing and inverse quantization processing on the transform coefficients, and decoding the transform coefficients Generating a value, performing an inverse transform process on the decoded value to generate a decoded value of the intra prediction residual signal, a prediction direction of intra prediction, and a distance between a prediction target block and a neighboring block Based on the identification means for identifying the reference block from the neighboring blocks and the decoded value of the intra prediction residual signal in the reference block, a prediction process is performed on the intra prediction residual signal of the prediction target block, and the residual is determined. Characterized in that it comprises a residual prediction error signal generating means for generating a measurement error signal.

  The present invention is an image encoding device that generates a prediction signal using spatial inter-pixel prediction and generates a residual signal between the prediction signal and an original signal, and includes an intra prediction residual signal. A prediction processing means, performing inter-transform processing on the intra prediction residual signal to generate transform coefficients, performing quantization processing and inverse quantization processing on the transform coefficients, and decoding the transform coefficients A generating unit that generates a value; an identifying unit that identifies a reference block from a neighboring block based on a prediction direction of intra prediction and a distance between a prediction target block and the neighboring block; and conversion of an intra prediction residual signal in the reference block Residual prediction error signal for generating a prediction error signal for the transform coefficient by performing prediction processing on the transform coefficient of the intra prediction residual signal of the prediction target block using the decoded value of the coefficient Characterized in that it comprises a forming unit.

  The present invention is an image decoding apparatus for decoding an encoded stream output from an encoder, wherein entropy decoding means for entropy decoding the encoded stream, and intra prediction mode from the encoded data subjected to the entropy decoding process. Read, based on the intra prediction mode, intra prediction processing means for generating an intra prediction signal, and based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, identify the reference block from the neighboring block, Using the decoded value of the intra prediction residual signal in the reference block, a prediction process is performed on the intra prediction residual signal of the decoding target block to generate a prediction value of the intra prediction residual signal, and the quantization index Inverse quantization processing is performed to generate a decoded value of the transform coefficient. Then, inverse transform processing is performed to generate a decoded signal of the residual prediction error signal, and the decoded signal of the residual prediction error signal and the prediction value of the intra prediction residual signal are added to decode the intra prediction residual signal An intra prediction residual signal decoding means for generating a signal, an intra prediction residual signal decoded signal generated by the intra prediction residual signal decoding means, and an intra prediction signal generated by the intra prediction processing means. And generating means for adding and generating a decoded signal.

  The present invention is an image decoding apparatus for decoding an encoded stream output from an encoder, wherein entropy decoding means for entropy decoding the encoded stream, and reading an intra prediction mode from the encoded data that has been entropy decoded Intra prediction processing means for generating an intra prediction signal based on the intra prediction mode, and an intra prediction residual flag indicating the presence / absence of prediction processing for the intra prediction residual is read out from the entropy-decoded encoded data. When it is determined that the prediction residual flag indicates that the prediction process is not applied to the intra prediction residual, the quantized index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the transform coefficient That generates a decoded signal of an intra prediction residual signal by performing an inverse transform process on the decoded value of When it is determined that the first decoding means of the measurement residual signal and the intra prediction residual flag indicate the application of the prediction process to the intra prediction residual, the prediction direction of the intra prediction and the prediction target block and the neighboring block Based on the distance between the reference block, a reference block is identified from neighboring blocks, and a prediction process is performed on the intra prediction residual signal of the decoding target block using a decoded value of the intra prediction residual signal in the reference block, so that intra prediction is performed. Generates a prediction value of the residual signal, performs an inverse quantization process on the quantization index to generate a decoded value of the transform coefficient, performs an inverse transform process on the decoded value of the transform coefficient, and performs a residual prediction Generate a decoded signal of the error signal, add the decoded signal of the residual prediction error signal and the prediction value of the intra prediction residual signal, and generate a decoded signal of the intra prediction residual signal A decoding signal of the intra prediction residual signal generated by the second decoding means of the intra prediction residual signal and the first decoding means of the intra prediction residual signal, or a second of the intra prediction residual signal. And a generating means for adding the decoded signal of the intra prediction residual signal generated by the decoding means and the intra prediction signal generated by the intra prediction processing means to generate a decoded signal.

  The present invention is an image decoding apparatus for decoding an encoded stream output from an encoder, wherein entropy decoding means for entropy decoding the encoded stream, and intra prediction mode from the encoded data subjected to the entropy decoding process. Read, based on the intra prediction mode, intra prediction processing means for generating an intra prediction signal, and based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, identify the reference block from the neighboring block, Using the decoded value of the transform coefficient in the reference block, the prediction process for the transform coefficient of the decoding target block is performed to generate the predicted value of the transform coefficient of the decoding target block, and the inverse quantization process is performed on the quantization index. To generate a decoded value of the residual prediction error signal for the transform coefficient, and to generate the residual prediction error Generated by the transform coefficient decoding means for generating the decoded signal of the transform coefficient of the intra prediction residual signal and the transform coefficient decoding means. Intra prediction residual signal decoding means for generating a decoded signal of an intra prediction residual signal by performing inverse transform processing on the decoded value of the transform coefficient, and intra prediction generated by the decoding means for the intra prediction residual signal And a generation unit configured to add a decoded signal of the residual signal and the intra prediction signal generated by the intra prediction processing unit to generate a decoded signal.

  The present invention is an image decoding apparatus for decoding an encoded stream output from an encoder, wherein entropy decoding means for entropy decoding the encoded stream, and reading an intra prediction mode from the encoded data that has been entropy decoded Intra prediction processing means for generating an intra prediction signal based on the intra prediction mode, and an intra prediction residual flag indicating the presence / absence of prediction processing for the intra prediction residual is read out from the entropy-decoded encoded data. When it is determined that the prediction residual flag indicates that the prediction process is not applied to the intra prediction residual, the quantized index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the transform coefficient That generates a decoded signal of an intra prediction residual signal by performing an inverse transform process on the decoded value of When it is determined that the first decoding means of the measurement residual signal and the intra prediction residual flag indicate the application of the prediction process to the intra prediction residual, the prediction direction of the intra prediction and the prediction target block and the neighboring block Based on the distance between the reference blocks, the reference block is identified from the neighboring blocks, the prediction value for the transform coefficient of the decoding target block is predicted using the decoded value of the transform coefficient in the reference block, and the prediction of the transform coefficient of the decoding target block is performed. A value is generated, an inverse quantization process is performed on the quantization index to generate a decoded value of the residual prediction error signal for the transform coefficient, and the decoded value of the residual prediction error signal and the transform coefficient of the decoding target block Generated by the transform coefficient decoding means for generating a decoded signal of the transform coefficient of the intra prediction residual signal, and the transform coefficient decoding means. Second decoding means for the intra prediction residual signal for generating a decoded signal of the intra prediction residual signal by performing inverse transform processing on the decoded value of the transform coefficient, and first decoding of the intra prediction residual signal A decoded signal of the intra prediction residual signal generated by the means, or a decoded signal of the intra prediction residual signal generated by the second decoding means of the intra prediction residual signal, and the intra prediction processing means. And generating means for generating a decoded signal by adding the intra prediction signal.

  According to this invention, paying attention to the correlation between intra prediction errors, and reducing the intra prediction error, a highly efficient intra coding method can be established.

It is a block diagram which shows the structure of the encoding apparatus using the residual prediction in a pixel area | region by this 1st Embodiment. It is a flowchart for demonstrating operation | movement (when using the residual prediction in a pixel area) of the encoding apparatus by this 1st Embodiment. It is a flowchart for demonstrating operation | movement (when using the residual prediction in a pixel area) of the encoding apparatus by this 1st Embodiment. It is a block diagram which shows the structure of the decoding apparatus using the residual prediction in a pixel area | region by this 1st Embodiment. It is a flowchart for demonstrating operation | movement of the decoding apparatus using the residual prediction in a pixel area | region by this 1st Embodiment. It is a block diagram which shows the structure of the encoding apparatus using the residual prediction with respect to a transform coefficient by this 2nd Embodiment. It is a flowchart for demonstrating operation | movement (when using the residual prediction with respect to a transform coefficient) of the encoding apparatus by this 2nd Embodiment. It is a flowchart for demonstrating operation | movement (when using the residual prediction with respect to a transform coefficient) of the encoding apparatus by this 2nd Embodiment. It is a block diagram which shows the structure of the decoding apparatus using the residual prediction with respect to a transform coefficient by this 2nd Embodiment. It is a flowchart for demonstrating operation | movement of the decoding apparatus using the residual prediction with respect to a transform coefficient by this 2nd Embodiment. It is a block diagram which shows the structure of the encoding apparatus using the conventional intra prediction. It is a flowchart for demonstrating operation | movement of the encoding apparatus using the conventional intra prediction. It is a block diagram which shows the structure of the decoding apparatus using the conventional intra prediction. It is a flowchart for demonstrating operation | movement of the decoding apparatus using the conventional intra prediction.

  Hereinafter, an embodiment of the present invention will be described with reference to the drawings.

  Consider two similar regions in the image. When intra prediction in the same direction is applied to these two regions, it is expected that the intra prediction residuals are similar signals. That is, although a prediction residual occurs, there is a correlation between the generated prediction residuals. Therefore, in the present invention, residual prediction is introduced that uses the correlation between prediction residuals in similar regions, performs prediction between both prediction residuals, and reduces the prediction residual. This is called intra prediction residual prediction (IPRP). The IPRP is configured by setting a reference block used for residual prediction, setting a prediction method for an intra prediction residual signal, and specifying an IPRP. Hereinafter, specific processing of IPRP will be described.

A. Reference Block Setting Method First, a block to be referred to in IPRP is set based on the block size / intra prediction direction of the predicted block.

A-1. Setting method 1
A neighboring block having the same size and the same prediction direction as the block to be predicted is set as a reference block for residual prediction. As the reference block, the nearest block having the shortest distance between the blocks is set. Alternatively, it is possible to limit the designated range of the reference block within a certain range. When the range is limited in this way, residual prediction is not applied if there is no corresponding block within the range. The non-applicability of residual prediction can be shared between the encoding side and the decoding side without any special additional information.

A-2. Setting method 2
A neighboring block having the same size and similar prediction direction to the block to be predicted is set as a reference block for residual prediction. The reference block is the same as described above. Further, which prediction direction is considered to be similar is determined separately. For example, a specific prediction direction is defined as a representative direction, and the other prediction directions are aggregated into a representative direction that minimizes the angle divergence of the prediction directions.

B. Intra prediction residual signal prediction method B-1. Prediction method 1
The decoded value of the intra prediction residual signal in the reference block is regarded as the prediction signal of the intra prediction residual in the predicted block, and the difference from the intra prediction residual signal in the predicted block is calculated.

B-2. Prediction method 2 (weighted difference prediction)
The decoded value of the intra prediction residual signal in the reference block multiplied by the weighting factor is regarded as the prediction signal of the intra prediction residual in the predicted block, and the difference from the intra prediction residual signal in the predicted block is calculated. Is. A method for calculating the weighting coefficient is given separately. As an example of the weighting factor, an intra prediction reference pixel group in a reference block and an intra prediction reference pixel group in a block to be predicted are regarded as two vector data, and an inner product value between both vectors can be raised. Alternatively, a predetermined value can be used.

B-3. Prediction method 3 (spatial adaptive prediction)
Only the residual signal at a specific position in the block is limited, and the decoded value of the intra prediction residual signal in the reference block is regarded as the prediction signal of the intra prediction residual in the predicted block. The difference with the intra prediction residual signal is calculated. It is assumed that which position is the target of residual prediction is switched according to the intra prediction direction.

B-4. Prediction method 4 (weighted space adaptive prediction)
Only the residual signal at a specific position in the block is limited, and the decoded value of the intra prediction residual signal in the reference block is regarded as the prediction signal of the intra prediction residual in the predicted block. The difference with the intra prediction residual signal is calculated.

B-5. Prediction method 5 (adaptive transform coefficient prediction)
The decoded value of the transform coefficient obtained by performing the quantization process and the inverse quantization process on the transform coefficient obtained by performing transform coding on the intra prediction residual in the reference block is regarded as a prediction signal. The difference between the conversion coefficient for the prediction residual of the block to be predicted and the prediction signal is calculated by limiting to only the conversion coefficient of the specific component. Which element is the target of prediction is switched according to the intra prediction direction.

B-6. Prediction method 6 (weighted adaptive transform coefficient prediction)
Generate decoded values of transform coefficients obtained by performing quantization processing and inverse quantization processing on transform coefficients obtained by performing transform coding on the intra prediction residual in the reference block, and generate the transform A weighted transformation coefficient obtained by multiplying the decoded value of the coefficient by a weighting factor is obtained, and the weighted transformation coefficient is limited to a transformation coefficient of a specific component as a prediction signal of the transformation coefficient for the prediction residual of the predicted block. Thus, the difference signal between the transform coefficient for the prediction residual of the block to be predicted and the prediction signal (the weighted transform coefficient) is calculated.

  In the above prediction method, it is also possible to transmit the prediction residual code information without using the prediction processing target as the prediction residual as the absolute value of the prediction residual.

C. Residual prediction designation method Whether or not residual prediction is applied is determined in a predetermined unit. Applicability is transmitted using a 1-bit flag (residual prediction flag). The unit to be switched can be a sequence, a frame, or an area within the frame, and this switching unit is determined separately.

  Calculate the correlation between the intra prediction reference pixel group in the reference block and the intra prediction reference pixel group in the block to be predicted, and limit the application of residual prediction when the correlation value is equal to or greater than a certain threshold It is also possible to do. In this case, if the correlation value is less than the threshold value, it is not necessary to transmit the residual prediction flag.

D. First Embodiment First, a first embodiment of the present invention will be described.
FIG. 1 is a block diagram illustrating a configuration of an encoding device using residual prediction in a pixel region according to the first embodiment. The encoding apparatus of the first embodiment is different from the configuration of the related art in that a residual prediction processing unit 110, a prediction residual signal storage unit 111, and an adder A3 are added before the conversion processing unit 102. Different. The residual prediction processing unit 110 performs prediction processing on the intra prediction residual signal, generates a prediction signal of the intra prediction residual signal, and generates a difference signal (residual prediction error signal) between the prediction signal and the intra prediction residual signal. ) Is generated. The prediction residual signal storage unit 111 stores a residual prediction error signal. The adder A3 inputs the decoded signal of the residual prediction error signal and the prediction value of the intra prediction residual signal (from the prediction residual signal storage unit 111), adds them, and decodes the intra prediction residual signal. Generate a value.

  2 and 3 are flowcharts for explaining the operation of the encoding apparatus according to the first embodiment (when residual prediction in the pixel region is used). The following steps S301 to S310 are executed for each block (referred to as a coding unit unit) to be subjected to intra prediction. First, an intra coding scheme (intra coding mode) is set, an intra prediction signal is generated based on the set intra prediction mode, and a difference signal (intra prediction residual signal) between the intra prediction signal and the input signal is generated. ) Is generated (step S302). Next, the process of step S103 of FIG. 12 described in the prior art is executed (step S303).

  Next, it is determined whether to perform residual prediction based on application conditions of residual prediction (step S304). Residual prediction refers to prediction processing for an intra prediction residual signal. When the application condition is satisfied, the process proceeds to step S305. On the other hand, if the application condition is not satisfied, the process proceeds to step S310 described later. As an example of the application condition, as described above, threshold processing for the correlation value between the intra prediction reference pixel group in the reference block and the intra prediction reference pixel group in the predicted block can be given.

  In step S305, the following steps S305-1 to S305-8 are executed. First, the residual prediction processing unit 110 performs a prediction process on the intra prediction residual signal, generates a prediction signal of the intra prediction residual signal, and generates a difference signal (residual prediction) between the prediction signal and the intra prediction residual signal. Error signal) is generated (step S305-1). As a method for generating a prediction signal of an intra prediction residual signal, any one of the prediction methods B-1, B-2, B-3, and B-4 can be used. Next, the conversion processing unit 102 performs conversion processing using the residual prediction error signal from the residual prediction processing unit 110 as input, and generates a conversion coefficient (step S305-2). The conversion process is determined in advance, and for example, discrete cosine conversion is used.

  Next, the quantization processing unit 103 performs quantization processing using the transform coefficient as an input, generates a quantized transform coefficient as a quantization index (step S305-3), and the entropy coding processing unit 109 performs quantization. The entropy encoding process is performed using the conversion index as an input, and conversion into a binary string is performed (step S305-4). The entropy coding method is given separately.

  Next, the inverse quantization processing unit 104 performs an inverse quantization process using the quantization index as an input, and generates a decoded value of the transform coefficient (step S305-6). Next, the inverse transform processing unit 105 inputs the decoded value of the transform coefficient, performs inverse transform processing, generates a decoded signal of the residual prediction error signal (step S305-7), and the adder A3 performs the residual. The decoded signal of the prediction error signal and the prediction value of the intra prediction residual signal are input, and both are added to generate a decoded value of the intra prediction residual signal (step S305-8).

  Next, the RD (Rate-Distortion) cost when the intra prediction residual output at step S305 is used is compared with the RD cost when the intra prediction residual output at step S303 is not used (step S306). ) If the former is smaller, the process proceeds to step S307. Otherwise, the process proceeds to step S308. The RD cost is a value determined as a weighted sum of the coding distortion and the generated code amount of the coded data, with the square sum of the difference signal between the input signal and the decoded signal as coding distortion.

  In step S307, a flag indicating that the intra prediction residual is used is set to ON (binary number 1). On the other hand, in step S308, a flag indicating the use of the intra prediction residual is set to OFF (binary number 0). Then, the adder A2 inputs the decoded value of the intra prediction residual signal and the intra prediction signal, adds them together, and generates a decoded signal (step S309).

  FIG. 4 is a block diagram showing a configuration of a decoding apparatus using residual prediction in the pixel area according to the first embodiment. The decoding apparatus according to the first embodiment is different from the configuration of the prior art in that a residual prediction information storage unit 210, a prediction residual signal storage unit 211, and a residual prediction processing unit 212 are added. The residual prediction information storage unit 210 stores a residual prediction error signal. The prediction residual signal storage unit 211 stores an intra prediction residual signal. The residual prediction processing unit 212 performs prediction processing on the intra prediction residual signal, and generates a prediction value of the intra prediction residual signal.

  FIG. 5 is a flowchart for explaining the operation of the decoding apparatus using the residual prediction in the pixel area according to the first embodiment. The processing of steps S401 to S409 is executed for each encoding unit. First, the entropy decoding processing unit 201 receives the encoded data of the encoding unit as input, performs entropy decoding processing (step S402), and the intra prediction processing unit 205 uses the intra-coded method (from the decoded encoded data ( Intra coding mode) is read out, and an intra prediction signal is generated based on the intra prediction mode (step S403). For example, H.M. In H.264 / AVC, nine prediction modes having different prediction directions are prepared for a 4 × 4 pixel block.

  Next, it is determined whether to perform residual prediction based on application conditions of residual prediction (step S410). When the application condition is satisfied, the process proceeds to step S405. On the other hand, if the application condition is not satisfied, the process proceeds to step S408. Next, a flag (intra prediction residual flag) indicating whether or not the prediction process is applied to the intra prediction residual signal is read from the decoded encoded data (step S404), and whether the intra prediction residual flag is ON. It is determined whether or not (step S405). If the intra prediction residual flag is ON, the process proceeds to step S406. Otherwise, the process of step S204 in FIG. 14 is executed (step S407).

  In step S406, the following steps S406-1 to S406-4 are executed. First, the residual prediction processing unit 212 performs prediction processing on the intra prediction residual signal, generates a prediction value of the intra prediction residual signal (step S406-1), and the inverse quantization processing unit 202 performs quantum processing. Using the quantization index as input, inverse quantization processing is performed to generate a decoded value of the transform coefficient (step S406-2). As a method for generating a prediction signal of an intra prediction residual signal, any one of the prediction methods B-1, B-2, B-3, and B-4 can be used.

  Next, the inverse transform processing unit 203 performs the inverse transform process by using the decoded value of the transform coefficient as an input to generate a decoded signal of the residual prediction error signal (step S406-3), and the adder B2 performs the residual. The decoded signal of the prediction error signal and the prediction value of the intra prediction residual signal (generated in step S406-1) are input, and both are added to generate a decoded value of the intra prediction residual signal (step S406-4). .

  Then, the adder B1 inputs the decoded value of the intra prediction residual signal generated in step S407 or step S406 and the intra predicted signal (generated in step S403), and adds both to generate a decoded signal. (Step S408).

  According to the first embodiment described above, the intra prediction residual can be reduced, and the code amount can be reduced.

E. Second Embodiment Next, a second embodiment will be described.
FIG. 6 is a block diagram showing a configuration of an encoding apparatus using residual prediction for transform coefficients according to the second embodiment. The encoding apparatus according to the second embodiment is different from the configuration of the conventional technique in that a residual prediction processing unit 110 and a prediction residual coefficient storage unit 111 are added after the conversion processing unit 102. The residual prediction processing unit 110 performs prediction processing on the transform coefficient and generates a prediction error signal of the transform coefficient.

  FIGS. 7 and 8 are flowcharts for explaining the operation of the encoding apparatus according to the second embodiment (when residual prediction for transform coefficients is used). The following steps S501 to S511 are executed for each encoding unit. First, an intra coding method (intra coding mode) is set by the intra prediction processing unit 101, an intra prediction signal is generated based on the set intra prediction mode, and an intra prediction mode is further generated by the adder A1. And an input signal difference signal (intra prediction residual signal) is generated (step S502). Next, the conversion processing unit 102 performs a conversion process on the intra prediction residual signal to generate a conversion coefficient (step S503).

  Next, the quantization processing unit 103 performs quantization processing on the transform coefficient to generate a quantization index (step S504-1), and the inverse quantization processing unit 104 performs inverse quantization processing on the quantization index. Then, a decoded value of the transform coefficient is generated (step S504-2), and the inverse transform processing unit 105 performs an inverse transform process on the decoded value of the change coefficient to generate a decoded signal of the intra prediction residual signal ( Step S504-3).

  Next, it is determined whether to perform residual prediction based on application conditions of residual prediction (step S505). If the application condition is satisfied, the process proceeds to step S506. On the other hand, if the application condition is not satisfied, the process proceeds to step S511 described later. As an example of the application condition, as described above, threshold processing for the correlation value between the intra prediction reference pixel group in the reference block and the intra prediction reference pixel group in the predicted block can be given.

  In step S506, the following processes of steps S506-1 to S506-6 are executed. First, prediction processing is performed on the transform coefficient by the residual prediction processing unit 110 to generate a prediction error value of the transform coefficient (step S506-1), and the quantization processing unit 103 performs processing on the prediction error value of the transform coefficient. Quantization processing is performed to generate a quantization index (step S506-2). As the conversion coefficient prediction method, the prediction method of B-5 or the previous term B-6 is used. Subsequently, the entropy encoding processing unit 109 receives the quantization index as input and performs entropy encoding processing to convert it into a binary sequence (step S506-3). The entropy coding method is given separately.

  Next, the inverse quantization processing unit 104 performs an inverse quantization process using the quantization index as an input, generates a decoded error signal prediction value (step S506-4), and uses the adder A4 to perform the conversion. The predicted value of the transform coefficient (generated in step S506-1) is added to the decoded value of the prediction error signal of the coefficient to generate a decoded value of the transform coefficient (step S506-5), and the inverse transform processing unit 105 performs transform. An inverse transform process is performed on the decoded value of the coefficient to generate a decoded value of the prediction error signal (step S506-6).

  Next, the RD cost when the intra prediction residual output at step S506 is used is compared with the RD cost when the intra prediction residual output at step S504 is not used (step S507). If is smaller, the process proceeds to step S508. Otherwise, the process proceeds to step S509.

  In step S508, a flag indicating use of the intra prediction residual is set to ON (binary number 1). On the other hand, in step S509, a flag indicating use of the intra prediction residual is set to OFF (binary number 0). Then, the adder A2 receives the decoded value of the prediction residual signal and the prediction signal, adds them together, and generates a decoded signal (step S510).

  FIG. 9 is a block diagram showing a configuration of a decoding apparatus using residual prediction for transform coefficients according to the second embodiment. The decoding apparatus of the second embodiment is different from the configuration of the prior art in that a residual prediction information storage unit 210, a prediction residual coefficient storage unit 211, and a residual prediction processing unit 212 are added. The residual prediction processing unit 212 receives the decoded value of the transform coefficient as input, performs a prediction process, and generates a predicted value of the transform coefficient.

  FIG. 10 is a flowchart for explaining the operation of the decoding apparatus using residual prediction for transform coefficients according to the second embodiment. The processes in steps S601 to S609 are executed for each encoding unit. First, the entropy decoding processing unit 201 receives the encoded data of the encoding unit as input, performs entropy decoding processing (step S602), and the intra prediction processing unit 205 uses the intra-coded method ( Intra coding mode) is read out, and an intra prediction signal is generated based on the intra prediction mode (step S603).

Next, it is determined whether to perform residual prediction based on application conditions of residual prediction (step S610). When the application condition is satisfied, the process proceeds to step S605. On the other hand, if the application condition is not satisfied, the process proceeds to step S608. Subsequently, a flag (intra prediction residual flag) indicating whether or not residual prediction is applied is read from the decoded encoded data (step S604), and it is determined whether or not the intra prediction residual flag is ON. (Step S605). If the intra prediction residual flag is ON, the process proceeds to step S606. Otherwise, the process proceeds to step S607 described later. In step S607, the process of step S204 of FIG. 14 is executed.

  In step S606, the following steps S606-1 to S606-4 are executed. First, the residual prediction processing unit 212 receives the decoded value of the transform coefficient as input, performs a prediction process, generates a predicted value of the transform coefficient (step S606-1), and the inverse quantization processing unit 202 performs the quantization index. Is input and an inverse quantization process is performed to generate a decoded error prediction value (step S606-2). As the conversion coefficient prediction method, the prediction method of B-5 or the previous term B-6 is used.

  Next, the adder B2 receives the decoded signal of the prediction error signal of the transform coefficient and the predicted value of the transform coefficient (generated in step S606-1) and adds both to generate a decoded value of the transform coefficient ( Step S606-3). Next, the inverse transformation processing unit 203 performs an inverse transformation process on the decoded coefficient decoding value to generate a decoded value of the intra prediction residual signal (step S606-4).

  Then, the decoded value of the intra prediction residual signal and the intra prediction signal (generated in step S603) are input, and both are added to generate a decoded signal (step S608).

  According to the second embodiment described above, intra prediction errors can be reduced, and the amount of codes can be reduced.

  DESCRIPTION OF SYMBOLS 101 ... Intra prediction process part, 102 ... Transformation process part, 103 ... Quantization process part, 104 ... Dequantization process part, 105 ... Inverse transform process part, 106 ... In-loop filter process part, 107 ... Frame memory, 108 ... Intra prediction information encoding unit 109 ... entropy encoding processing unit 110 ... residual prediction processing unit 111 ... prediction residual signal storage unit A1, A2, A3 ... adder, 201 ... entropy decoding processing unit 202 ... Inverse quantization processing unit 203 ... Inverse transformation processing unit 204 ... Intra prediction information storage unit 205 ... Intra prediction processing unit 206 ... In-loop filter processing unit 207 ... Frame memory 210 ... Residual prediction information storage unit 211 ... Prediction residual signal storage unit, 212 ... Residual prediction processing unit

Claims (13)

An image encoding method for generating a prediction signal using spatial inter-pixel prediction and generating a residual signal between the prediction signal and an original signal,
An intra prediction processing step for generating an intra prediction residual signal;
Performing an inter-transform process on the intra prediction residual signal to generate a transform coefficient, performing a quantization process and an inverse quantization process on the transform coefficient, and generating a decoded value of the transform coefficient; A generation step of performing an inverse transformation process on the decoded value to generate a decoded value of the intra prediction residual signal;
An identification step for identifying a reference block from neighboring blocks based on a prediction direction of intra prediction and a distance between the prediction target block and the neighboring block;
A residual prediction error signal generating step of generating a residual prediction error signal by performing a prediction process on the intra prediction residual signal of the prediction target block using a decoded value of the intra prediction residual signal in the reference block. An image encoding method characterized by the above. An image encoding method for generating a prediction signal using spatial inter-pixel prediction and generating a residual signal between the prediction signal and an original signal,
An intra prediction processing step for generating an intra prediction residual signal;
Generating a transform coefficient by performing an inter-transform process on the intra prediction residual signal, generating a decoded value of the transform coefficient by performing a quantization process and an inverse quantization process on the transform coefficient Steps,
An identification step for identifying a reference block from neighboring blocks based on a prediction direction of intra prediction and a distance between the prediction target block and the neighboring block;
Using the decoded value of the transform coefficient of the intra prediction residual signal in the reference block, a prediction process is performed on the transform coefficient of the intra prediction residual signal of the prediction target block, and a residual prediction error signal for the transform coefficient is generated. And a difference prediction error signal generation step. A first quantization obtained by performing transform processing on the residual prediction error signal generated in the residual prediction error signal generating step to generate a transform coefficient, and performing quantization processing on the transform coefficient A first cost value generation step for obtaining a weighted sum of a quantization error due to the quantization process and a code amount for expressing the first quantization transform coefficient as a cost value for the transform coefficient;
A transform process is performed on the intra prediction residual signal generated in the intra prediction process step to generate a transform coefficient, and a second quantized transform coefficient obtained by performing a quantization process on the transform coefficient On the other hand, a second cost value generation step for obtaining a weighted sum of a code amount for expressing the quantization error due to the quantization processing and the second quantization transform coefficient as a cost value;
Compare the cost value obtained in the first cost value generation step with the cost value obtained in the second cost value generation step, the cost value obtained by the first cost value generation step is more If it is smaller, an intra prediction residual flag indicating that the prediction process for the intra prediction residual is applied is output as encoded data. In other cases, the intra process indicating that the prediction process for the intra prediction residual is not applied. The image encoding method according to claim 1, further comprising: an output step of outputting a prediction residual flag as encoded data. A first quantization obtained by performing transform processing on the residual prediction error signal generated in the residual prediction error signal generating step to generate a transform coefficient, and performing quantization processing on the transform coefficient A first cost value generation step for obtaining a weighted sum of a quantization error due to the quantization process and a code amount for expressing the first quantization transform coefficient as a cost value for the transform coefficient;
For the second quantized transform coefficient obtained by performing the quantization process on the intra prediction residual signal for the transform coefficient generated in the intra prediction processing step, the quantization error due to the quantization process and the first A second cost value generation step for obtaining a weighted sum of code amounts for expressing the second quantized transform coefficient as a cost value;
Compare the cost value obtained in the first cost value generation step with the cost value obtained in the second cost value generation step, the cost value obtained by the first cost value generation step is more If it is smaller, an intra prediction residual flag indicating that the prediction process for the intra prediction residual is applied is output as encoded data. In other cases, the intra process indicating that the prediction process for the intra prediction residual is not applied. The image encoding method according to claim 2, further comprising: an output step of outputting a prediction residual flag as encoded data. An image decoding method for decoding an encoded stream output from an encoder,
Entropy decoding for entropy decoding the encoded stream;
An intra prediction processing step of reading an intra prediction mode from the encoded data subjected to the entropy decoding process, and generating an intra prediction signal based on the intra prediction mode;
Based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is identified from the neighboring block, and the decoded block value of the intra prediction residual signal in the reference block is used to determine the intra of the decoding target block. Perform prediction processing on the prediction residual signal to generate a prediction value of the intra prediction residual signal, perform inverse quantization processing on the quantization index to generate a decoded value of the transform coefficient, and decode the transform coefficient Inverse transformation processing is performed on the value to generate a decoded signal of the residual prediction error signal, the decoded signal of the residual prediction error signal and the prediction value of the intra prediction residual signal are added, and an intra prediction residual signal An intra prediction residual signal decoding step for generating a decoded signal of
A generation step of adding a decoded signal of the intra prediction residual signal generated in the decoding step of the intra prediction residual signal and the intra prediction signal generated in the intra prediction processing step to generate a decoded signal. An image decoding method characterized by the above. An image decoding method for decoding an encoded stream output from an encoder,
Entropy decoding for entropy decoding the encoded stream;
An intra prediction processing step of reading an intra prediction mode from the encoded data subjected to entropy decoding, and generating an intra prediction signal based on the intra prediction mode;
When an intra prediction residual flag indicating the presence / absence of prediction processing for an intra prediction residual is read from the encoded data subjected to entropy decoding, and the intra prediction residual flag indicates that the prediction processing is not applied to the intra prediction residual If it is determined, the quantization index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the decoded value of the transform coefficient is subjected to inverse transform processing to obtain a decoded signal of the intra prediction residual signal. A first decoding step of the intra prediction residual signal to generate
When it is determined that the intra prediction residual flag indicates that the prediction process is applied to the intra prediction residual, based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is changed from the neighboring block to the reference block. And performing a prediction process on the intra prediction residual signal of the decoding target block using the decoded value of the intra prediction residual signal in the reference block, generating a prediction value of the intra prediction residual signal, and quantizing An inverse quantization process is performed on the index to generate a decoded value of the transform coefficient, an inverse transform process is performed on the decoded value of the transform coefficient to generate a decoded signal of a residual prediction error signal, and the residual A decoded signal of the prediction error signal and a prediction value of the intra prediction residual signal are added to generate a decoded signal of the intra prediction residual signal. No. and step,
Decoding of the intra prediction residual signal generated in the first decoding step of the intra prediction residual signal or decoding of the intra prediction residual signal generated in the second decoding step of the intra prediction residual signal An image decoding method comprising: a generation step of adding a signal and the intra prediction signal generated in the intra prediction processing step to generate a decoded signal. An image decoding method for decoding an encoded stream output from an encoder,
Entropy decoding for entropy decoding the encoded stream;
An intra prediction processing step of reading an intra prediction mode from the encoded data subjected to the entropy decoding process, and generating an intra prediction signal based on the intra prediction mode;
Based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is identified from the neighboring block, and the prediction of the transform coefficient of the decoding target block is performed using the decoded value of the transform coefficient in the reference block. To generate a prediction value of the transform coefficient of the decoding target block, to perform a dequantization process on the quantization index to generate a decoded value of a residual prediction error signal for the transform coefficient, and to perform the residual prediction A transform coefficient decoding step of adding a decoded value of the error signal and a predicted value of the transform coefficient of the decoding target block to generate a decoded signal of the transform coefficient of the intra prediction residual signal;
An intra prediction residual signal decoding step of performing an inverse transform process on a decoded value of the transform coefficient generated in the transform coefficient decoding step to generate a decoded signal of the intra prediction residual signal;
A generation step of adding a decoded signal of the intra prediction residual signal generated in the decoding step of the intra prediction residual signal and the intra prediction signal generated in the intra prediction processing step to generate a decoded signal. An image decoding method characterized by the above. An image decoding method for decoding an encoded stream output from an encoder,
Entropy decoding for entropy decoding the encoded stream;
An intra prediction processing step of reading an intra prediction mode from the encoded data subjected to entropy decoding, and generating an intra prediction signal based on the intra prediction mode;
When an intra prediction residual flag indicating the presence / absence of prediction processing for an intra prediction residual is read from the encoded data subjected to entropy decoding, and the intra prediction residual flag indicates that the prediction processing is not applied to the intra prediction residual If it is determined, the quantization index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the decoded value of the transform coefficient is subjected to inverse transform processing to obtain a decoded signal of the intra prediction residual signal. A first decoding step of the intra prediction residual signal to generate
When it is determined that the intra prediction residual flag indicates that the prediction process is applied to the intra prediction residual, based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is changed from the neighboring block to the reference block. And predicting the transform coefficient of the decoding target block using the decoded value of the transform coefficient in the reference block to generate a predicted value of the transform coefficient of the decoding target block, and performing inverse processing on the quantization index. Quantization processing is performed to generate a decoded value of the residual prediction error signal for the transform coefficient, the decoded value of the residual prediction error signal and the predicted value of the transform coefficient of the decoding target block are added, and an intra prediction residual signal A transform coefficient decoding step for generating a decoded signal of the transform coefficient;
A second decoding step of the intra prediction residual signal that performs an inverse transform process on the decoded value of the transform coefficient generated in the decoding step of the transform coefficient to generate a decoded signal of the intra prediction residual signal;
Decoding of the intra prediction residual signal generated in the first decoding step of the intra prediction residual signal or decoding of the intra prediction residual signal generated in the second decoding step of the intra prediction residual signal An image decoding method comprising: a generation step of adding a signal and the intra prediction signal generated in the intra prediction processing step to generate a decoded signal. An image encoding device that generates a prediction signal using spatial inter-pixel prediction and generates a residual signal between the prediction signal and an original signal,
Intra prediction processing means for generating an intra prediction residual signal;
Generating a transform coefficient by performing an inter-transform process on the intra prediction residual signal, generating a decoded value of the transform coefficient by performing a quantization process and an inverse quantization process on the transform coefficient Means,
Identification means for identifying a reference block from neighboring blocks based on a prediction direction of intra prediction and a distance between the prediction target block and the neighboring block;
Using the decoded value of the transform coefficient of the intra prediction residual signal in the reference block, a prediction process is performed on the transform coefficient of the intra prediction residual signal of the prediction target block, and a residual prediction error signal for the transform coefficient is generated. An image encoding apparatus comprising: a difference prediction error signal generation unit. An image decoding apparatus for decoding an encoded stream output from an encoder,
Entropy decoding means for entropy decoding the encoded stream;
Intra prediction processing means for reading an intra prediction mode from the encoded data subjected to the entropy decoding process and generating an intra prediction signal based on the intra prediction mode;
Based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is identified from the neighboring block, and the decoded block value of the intra prediction residual signal in the reference block is used to determine the intra of the decoding target block. Perform prediction processing on the prediction residual signal to generate a prediction value of the intra prediction residual signal, perform inverse quantization processing on the quantization index to generate a decoded value of the transform coefficient, and decode the transform coefficient Inverse transformation processing is performed on the value to generate a decoded signal of the residual prediction error signal, the decoded signal of the residual prediction error signal and the prediction value of the intra prediction residual signal are added, and an intra prediction residual signal Means for decoding an intra prediction residual signal for generating a decoded signal of
Generating means for adding a decoded signal of the intra prediction residual signal generated by the decoding means for the intra prediction residual signal and the intra prediction signal generated by the intra prediction processing means to generate a decoded signal; An image decoding apparatus characterized by the above. An image decoding apparatus for decoding an encoded stream output from an encoder,
Entropy decoding means for entropy decoding the encoded stream;
Intra prediction processing means for reading an intra prediction mode from the entropy-decoded encoded data and generating an intra prediction signal based on the intra prediction mode;
When an intra prediction residual flag indicating the presence / absence of prediction processing for an intra prediction residual is read from the encoded data subjected to entropy decoding, and the intra prediction residual flag indicates that the prediction processing is not applied to the intra prediction residual If it is determined, the quantization index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the decoded value of the transform coefficient is subjected to inverse transform processing to obtain a decoded signal of the intra prediction residual signal. First decoding means of the intra prediction residual signal to generate
When it is determined that the intra prediction residual flag indicates that the prediction process is applied to the intra prediction residual, based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is changed from the neighboring block to the reference block. And performing a prediction process on the intra prediction residual signal of the decoding target block using the decoded value of the intra prediction residual signal in the reference block, generating a prediction value of the intra prediction residual signal, and quantizing An inverse quantization process is performed on the index to generate a decoded value of the transform coefficient, an inverse transform process is performed on the decoded value of the transform coefficient to generate a decoded signal of a residual prediction error signal, and the residual A decoded signal of the prediction error signal and a prediction value of the intra prediction residual signal are added to generate a decoded signal of the intra prediction residual signal. And No. means,
Decoding of the intra prediction residual signal generated by the first decoding means of the intra prediction residual signal or decoding of the intra prediction residual signal generated by the second decoding means of the intra prediction residual signal An image decoding apparatus comprising: a generation unit that adds a signal and an intra prediction signal generated by the intra prediction processing unit to generate a decoded signal. An image decoding apparatus for decoding an encoded stream output from an encoder,
Entropy decoding means for entropy decoding the encoded stream;
Intra prediction processing means for reading an intra prediction mode from the encoded data subjected to the entropy decoding process and generating an intra prediction signal based on the intra prediction mode;
Based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is identified from the neighboring block, and the prediction of the transform coefficient of the decoding target block is performed using the decoded value of the transform coefficient in the reference block. To generate a prediction value of the transform coefficient of the decoding target block, to perform a dequantization process on the quantization index to generate a decoded value of a residual prediction error signal for the transform coefficient, and to perform the residual prediction A transform coefficient decoding means for adding a decoded value of the error signal and a predicted value of the transform coefficient of the decoding target block to generate a decoded signal of the transform coefficient of the intra prediction residual signal;
Intra prediction residual signal decoding means for generating a decoded signal of an intra prediction residual signal by performing inverse transform processing on a decoded value of the transform coefficient generated in the transform coefficient decoding means;
Generating means for adding a decoded signal of the intra prediction residual signal generated by the decoding means for the intra prediction residual signal and the intra prediction signal generated by the intra prediction processing means to generate a decoded signal; An image decoding apparatus characterized by the above. An image decoding apparatus for decoding an encoded stream output from an encoder,
Entropy decoding means for entropy decoding the encoded stream;
Intra prediction processing means for reading an intra prediction mode from the entropy-decoded encoded data and generating an intra prediction signal based on the intra prediction mode;
When an intra prediction residual flag indicating the presence / absence of prediction processing for an intra prediction residual is read from the encoded data subjected to entropy decoding, and the intra prediction residual flag indicates that the prediction processing is not applied to the intra prediction residual If it is determined, the quantization index is subjected to inverse quantization processing to generate a decoded value of the transform coefficient, and the decoded value of the transform coefficient is subjected to inverse transform processing to obtain a decoded signal of the intra prediction residual signal. First decoding means of the intra prediction residual signal to generate
When it is determined that the intra prediction residual flag indicates that the prediction process is applied to the intra prediction residual, based on the prediction direction of the intra prediction and the distance between the prediction target block and the neighboring block, the reference block is changed from the neighboring block to the reference block. And predicting the transform coefficient of the decoding target block using the decoded value of the transform coefficient in the reference block to generate a predicted value of the transform coefficient of the decoding target block, and performing inverse processing on the quantization index. Quantization processing is performed to generate a decoded value of the residual prediction error signal for the transform coefficient, the decoded value of the residual prediction error signal and the predicted value of the transform coefficient of the decoding target block are added, and an intra prediction residual signal Transform coefficient decoding means for generating a decoded signal of the transform coefficient;
Second intra prediction residual signal decoding means for performing an inverse transform process on a decoded value of the transform coefficient generated in the transform coefficient decoding means to generate a decoded signal of an intra prediction residual signal;
Decoding of the intra prediction residual signal generated by the first decoding means of the intra prediction residual signal or decoding of the intra prediction residual signal generated by the second decoding means of the intra prediction residual signal An image decoding apparatus comprising: a generation unit that adds a signal and an intra prediction signal generated by the intra prediction processing unit to generate a decoded signal.

Images