JP4831114B2 - Image encoding device, image encoding program, image decoding device, and image decoding program - Google Patents

Description

  The present invention relates to an image encoding device, an image encoding program, an image decoding device, and an image decoding program, and in particular, exists around a rectangular area to be encoded when performing in-plane encoding on an image. A decoded signal is generated by performing local decoding on the encoded rectangular area, and a prediction signal is generated by performing in-plane prediction using the decoded signal, and the result is encoded and decoded. The present invention relates to an applied image encoding device, an image encoding program, an image decoding device, and an image decoding program.

  Various moving image distribution services via digital broadcasting and networks have become common, and in order to record more high-quality and high-quality moving images, further improvement in the coding efficiency of moving images is desired. .

  In recent years, there has been introduced a new encoding method such as MPEG-4 AVC (hereinafter referred to as AVC or H.264) which greatly improves the encoding efficiency of the conventional encoding technology, and these are introduced. Thus, more effective information encoding can be performed.

  In the encoding method as described above, an encoded block existing around a rectangular area (hereinafter referred to as a block) to be encoded when performing in-plane encoding (hereinafter referred to as intra encoding). A decoded signal is generated by performing local decoding on the image, and a prediction signal is generated by performing in-plane prediction (hereinafter referred to as intra prediction) using the decoded signal. After that, by performing predetermined orthogonal transform and quantization on the residual signal obtained by the difference between the signal component in the block to be encoded and the generated prediction signal, the encoding efficiency is improved. (For example, refer to Patent Document 1).

Japanese Patent No. 3734492

  In AVC, which is a conventional moving picture coding, many block sizes can be used as a coding processing unit. For example, the block size in units of macroblocks is a block having 16 × 16 pixels, 16 × 8 pixels, 8 × 16 pixels, and 8 × 8 pixels, where M pixels in the horizontal direction and N pixels in the vertical direction (M and N are natural numbers) are expressed as M × N pixels. Is available. In addition, a block having 8 × 8 pixels, 8 × 4 pixels, 4 × 8 pixels, and 4 × 4 pixels can be used as the block size of each sub macroblock.

  When there is no particular change in the encoding order in AVC, processing is performed in the raster order of macroblock units as shown in FIG. In addition, blocks that are processing units of encoding when performing intra prediction are 16 × 16 pixels (FIG. 9E) and 8 × 8 pixels (FIG. 9C), depending on the encoding target such as a luminance component and a color difference component. ) A block size of 4 × 4 pixels (FIGS. 9B and 9D) can be used. Also in the processing order of intra prediction in AVC, processing is performed in a raster order as indicated by the order of arrows in FIGS. 9B, 9C, and 9D.

  In AVC intra prediction, a decoded signal is generated by performing local decoding on a block that is adjacent to a prediction target block and has already been encoded, and prediction is performed using this decoded signal. Generate a signal.

  Here, in the conventional intra prediction, when the encoding order is raster order, as shown in FIG. 10A, the left, upper, upper right, and upper left blocks are encoded with reference to the encoding target block. From the blocks other than these blocks, it is impossible to generate a prediction signal of the encoding target block using a decoded signal obtained by local decoding. That is, in the coding order, the decoded signal of the already processed block can be used for the prediction process, but the decoded signal cannot be obtained from the block that has not been processed yet because local decoding cannot be performed, and as a result, the prediction process is performed. It means that it cannot be used. Therefore, the prediction direction in the conventional intra prediction is a limited direction as shown in FIG. 10C, and a sufficient prediction result cannot always be obtained when trying to perform intra prediction from surrounding pixels. There is.

  The present invention has been made in view of the above points. An image code capable of generating a prediction signal with higher accuracy by avoiding the problem that the prediction direction is limited by the encoding order and the accuracy of the prediction signal is reduced. It is an object to provide an encoding device, an image encoding program, an image decoding device, and an image decoding program.

In order to achieve the above object, the first invention divides one screen area of an input encoding target image signal into predetermined rectangular areas, and processes the encoding target area in the divided rectangular areas. A rectangular area that is a target area to be processed and a predetermined surface using a decoded signal of a rectangular area around the rectangular area that is a target area to be processed with respect to an original signal of an image belonging to the encoding target area An image encoding device that performs intra prediction and generates a prediction signal,
The processing position control means for managing the position in one screen area of the encoding target area and outputting a processing position control signal for indicating this position, and the original signal of the image belonging to one predetermined rectangular area among the rectangular areas On the other hand, a predetermined in-plane prediction is performed using a decoded signal of a rectangular area around the rectangular area, and prediction specified based on supplied prediction auxiliary information among a plurality of predetermined prediction methods A prediction unit that generates a prediction signal in a predetermined rectangular area by the method, outputs the generated prediction signal, and information indicating a prediction method used to generate the prediction signal, and an image in the predetermined rectangular area from the original signal, to generate a Rizan difference signal by the subtracting the prediction signal supplied from the prediction unit, based on the information indicating the prediction method supplied from the prediction means, a plurality of prediction methods After that, the unused prediction method is identified, the process of supplying the identified prediction information to the prediction means as the prediction auxiliary information and generating the prediction signal is repeated, and if there are no unused prediction methods, the plurality of prediction methods are supported. A residual signal generating / selecting means for selecting a residual signal from a plurality of residual signals obtained based on a predetermined evaluation method, and orthogonal to a selected residual signal in a predetermined rectangular area A post-quantization information generating means for generating orthogonal transform coefficient information by performing transformation, quantizing the orthogonal transform coefficient information to generate post-quantization information, and after quantization in a predetermined rectangular area Encoding means for performing entropy encoding on information, decoding residual signal generating means for performing inverse quantization and inverse orthogonal transformation on the quantized information to generate a decoding residual signal, and a predetermined one Decoding in a rectangular region Based on the addition means for adding the difference signal and the prediction signal to generate a decoded signal in a predetermined rectangular area, and the processing position control signal from the processing position control means, the prediction signal is already generated and encoded. against rectangular region, anda re-predictive control unit that controls on the basis of whether or not to execute the prediction process again to a predetermined rule, the prediction signal generating means by the re-predictive control means, residual signal generator Control the selection unit, the post-quantization information generation unit, the decoded residual signal generation unit, and the addition unit, set the encoding target region as the first predetermined rectangular region, and predict the signal in the encoding target region , Generating a residual signal, post-quantization information, a decoded residual signal, and a decoded signal, and then encoding one of the rectangular areas around the encoding target area as a predetermined one rectangular area. Set a rectangular area Encoding that generates a prediction signal, a residual signal, post-quantization information, a decoded residual signal, and a decoded signal in a rectangular area, and then generates a decoded signal for an encoded image signal in the encoding target area Using the decoded signal in the rectangular area around the target area, after updating the prediction signal, residual signal, post-quantization information, decoded residual signal and decoded signal in the encoding target area, Re-encoding is performed.

  In order to achieve the above object, the second invention is an image encoding program that causes a computer to execute each component of the first invention.

In order to achieve the above object, the third invention divides one screen area of the image signal to be encoded into predetermined rectangular areas and processes the encoding target area in the divided rectangular areas. A predetermined rectangular in-plane prediction using a decoded signal of a rectangular area around the rectangular area to be processed with respect to an original signal of an image belonging to the encoding target area. An image decoding device that performs a decoding operation on an encoded bitstream obtained by performing encoding and outputs a decoded image signal,
Entropy decoding means for decoding at least post-quantization information obtained by performing orthogonal transform and inverse quantization on the residual signal in each rectangular area from the encoded bit stream, and for the decoded post-quantization information Te manages the decoded residual signal generating means for generating a decoded residual signal of each rectangular region is subjected to inverse quantization and inverse orthogonal transformation process, the position of the rectangular area to be decoded to generate a decoded signal, the position A processing position control means for outputting a processing position control signal for indicating the position of the rectangular area which is the current decoding target based on the processing position control signal from the processing position control means, and becomes the decoding target prediction method using the decoded signal of the surrounding area of the rectangular area performs a predetermined intra prediction among a plurality of prediction methods predetermined specified based on the supplied predicted auxiliary information are Therefore to generate a prediction signal in the rectangular region, and the generated predicted signal, prediction means for outputting the information indicating the prediction method used for generating the prediction signal, decoding by adding the prediction signal and the decoded residual signal A re-prediction control unit that includes an adding unit that generates a signal and a re-prediction control unit that controls whether or not the prediction process is performed again on a rectangular region that has already generated a prediction signal based on a predetermined rule. The prediction means, the decoded residual signal generating means, and the adding means are controlled by the means to generate a prediction signal, a decoded residual signal, and a decoded signal in the rectangular area to be decoded, and subsequently the decoding target. One rectangular area that has not been decoded is set out of the rectangular areas around the rectangular area, and a prediction signal, a decoded residual signal, and a decoded signal are generated in the rectangular area, and then the decoding target Rectangular Update the prediction signal, decoding residual signal, and decoding signal in the rectangular area to be decoded using the decoded signal in the surrounding rectangular area that generated the decoded signal for the image signal in the area It is characterized by.

  In order to achieve the above object, a fourth invention is an image decoding program that causes a computer to execute each component of the second invention.

  According to the first invention, it is possible to increase the prediction direction of the in-plane prediction that has been conventionally limited by using the decoded signal of the area around the encoding target for the in-plane prediction of the encoding target area. Thus, it is possible to generate a prediction signal with higher prediction accuracy than a prediction signal obtained by conventional in-plane prediction.

  Further, according to the first invention, the prediction direction of the in-plane prediction is increased, a prediction signal with higher prediction accuracy than the conventional prediction signal is generated, and the previously generated prediction signal is updated. Since the signal amplitude of the new residual signal to be obtained is suppressed, the coding efficiency of the coded bitstream obtained by the image coding method of the present invention can be improved.

  Further, according to the second invention, the encoding operation of the image encoding device according to the first invention is controlled by the computer by controlling the operation of the computer by the image encoding program describing the similar encoding operation. An encoding operation similar to that of the image encoding device of the first invention can be realized.

  Further, according to the third aspect, it is possible to correctly decode the encoding result from the encoded bit stream generated according to the first aspect or the second aspect.

  Further, according to the fourth invention, the decoding operation of the image decoding apparatus according to the third invention is controlled by the computer by the image decoding program describing the similar decoding operation, so that the computer can A decoding operation similar to that of the image decoding apparatus can be realized.

  Next, the best mode for carrying out the invention will be described in detail with reference to the drawings.

(Image coding device)
First, an embodiment of an image encoding device of the present invention will be described with reference to the drawings.

  FIG. 1 shows a block diagram of an embodiment of an image encoding apparatus according to the present invention. As shown in FIG. 1, an image coding apparatus 100 according to the present embodiment includes a picture storage memory 102, a re-prediction controller 104, a processing position controller 106, a decoded picture storage memory 107, a predictor 109, and a predicted picture storage memory. 110, subtractor 111, difference picture storage memory 112, residual controller 113, orthogonal transformer 115, quantizer 116, switch 117, inverse quantizer 118, inverse orthogonal transformer 119, adder 120, entropy coding The device 121 is at least provided.

  Next, details of each component of the image encoding device 100 according to the embodiment of the present invention shown in FIG. 1 will be described.

  In order to simplify the explanation, the processing unit for intra prediction is assumed to be a block (rectangular area) of 4 × 4 pixels, but the block size of intra prediction is not particularly limited, and 8 × 8 pixels. It is also possible to adopt a configuration in which intra prediction is performed with various block sizes such as 16 × 16 pixels.

  The picture storage memory 102 has a function of acquiring, as an input signal, one screen area of an input image signal to be encoded or one screen area of a moving image signal as an encoding target picture. The stored encoding target picture is handled in the picture storage memory 102 as being divided into a predetermined rectangular area, that is, a block having a predetermined size. Here, it is treated as a block having a block size of 4 × 4 pixels.

  Also, the picture storage memory 102 is supplied to each means that needs information belonging to each block and a function that manages information belonging to each block in the plane obtained from each external means and the state in the picture storage memory 102. Has the function of Here, the information belonging to the block corresponds to at least the image signal in the block, here the original signal. It is further preferable to include information such as the size of the block, the state of the designated block, the use history of the block, and the state of the blocks around the designated block. Here, the picture storage memory 102 has a function of supplying at least the original signal belonging to the block to be encoded to the subtractor 111.

  Further, the picture storage memory 102 has a function of acquiring a processing position control signal for specifying the position of a block that is currently an encoding target from a processing position controller 106 described later.

  The re-prediction controller 104 has a function of controlling whether to perform intra prediction again on a block that has already been subjected to intra prediction. The re-prediction controller 104 preferably has a function of performing re-prediction control of intra prediction based on a predetermined rule. Further, if necessary, it is better to have a function of acquiring re-prediction control information 103 for specifying a rule for performing re-prediction control of intra prediction. Here, the re-prediction control information 103 has a better configuration if predetermined entropy encoding is separately performed at the time of encoding output and is output as an encoded bit stream together with the output code 122.

Further, re-predictive controller 104 obtains the re-predictive control information 103 for controlling the re-predictive intra prediction, predictor 109 generates a re-predictive control signals based thereon, the residual controller 113 It has a function to supply. In addition, the re-prediction controller 104 has a function of acquiring a processing position control signal for specifying the position of the block currently being encoded from the processing position controller 106. Furthermore, the re-prediction controller 104 has a function of supplying a control signal for switching the switch 117 to the switch 117 in order to control the encoded output in accordance with the necessity of re-prediction of intra prediction.

Based on the processing position control information 105 , the processing position controller 106 determines the order of blocks to be processed, the positions of blocks to be encoded, and the encoding status of blocks around the block to be encoded. And at least a processing position control signal for specifying the position of the block to be encoded is generated, and the picture storage memory 102, the re-prediction controller 104, the decoded picture storage memory 107, the predictor 109, The prediction picture storage memory 110, the difference picture storage memory 112, and the residual controller 113 are provided. Here, the processing position control signal may include an encoding state of blocks around the block to be encoded.

  Further, the processing position controller 106 has a function of acquiring the processing position control information 105 for determining the order of blocks to be processed and determining the order of blocks to be processed based on the acquired processing position control information 105. A better configuration is obtained. Here, the processing position control information 105 has a better configuration if predetermined entropy encoding is separately performed at the time of encoding output and is output as an encoded bit stream together with the output code 122.

  The decoded picture storage memory 107 has a function of acquiring a processing position control signal from the processing position controller 106 and specifying the position of a block to be encoded. The decoded picture storage memory 107 has a function of acquiring a block including a decoded signal from the adder 120 and storing it as a decoding result of the position of the block to be encoded. Here, it is desirable that the decoded picture storage memory 107 has a function capable of associating and storing blocks including a plurality of decoded signals at the same block position. In this case, a decoded signal may be generated from a prediction signal generated based on different prediction methods and a decoded residual signal at the same position in the plane by re-prediction processing and supplied to the decoded picture storage memory 107. Because.

  The decoded picture storage memory 107 has a function of managing information belonging to each block in the plane, which is obtained from each external unit and the state in the decoded picture storage memory 107. Here, information belonging to a block corresponds to at least an image signal in the block, here, a decoded signal. In addition, it is preferable that each block includes information indicating whether encoding is completed. Further, the decoded picture storage memory 107 has a function of supplying the predictor 109 with information belonging to the block at the position in the requested plane.

  The predictor 109 has a function of acquiring a re-prediction control signal from the re-prediction controller 104. Based on the acquired re-prediction control signal, it has a function of specifying whether or not re-prediction is performed on a block that is currently an encoding target. Further, the predictor 109 acquires a processing position control signal necessary for specifying the position of the block to be encoded from the processing position controller 106, and based on the acquired processing position control signal, It has a function of specifying the position in the plane of the block currently being encoded. If necessary, the processing position controller 106 has a function of acquiring the encoding state of the blocks around the block to be encoded as the processing position control signal.

  Further, the predictor 109 obtains a decoded signal necessary for performing intra prediction, so that a decoded signal that can be used in blocks around the block to be encoded from the decoded picture storage memory 107 and can be used. Has a function of acquiring information belonging to the block in which the exists. Further, if necessary, the predictor 109 has a function of acquiring information belonging to a block in which a usable prediction signal exists among the blocks around the block to be encoded from the predicted picture storage memory 110. Then, it becomes a still better structure. Further, the predictor 109 has a function of acquiring prediction auxiliary information from the residual controller 113 and specifying a prediction method when performing intra prediction based on the acquired prediction auxiliary information.

  Further, the predictor 109 generates a prediction signal of the block to be encoded by the identified prediction method based on the decoded signal that can be used in the blocks around the acquired block to be encoded. It has the function to do. Further, the predictor 109 has a function of supplying information belonging to the block to be encoded, including the generated prediction signal, to the predicted picture storage memory 110. Here, by adding the acquired prediction auxiliary information as the information belonging to the block to be supplied, it is possible to specify the prediction method generated by the generated prediction signal, and the configuration is further improved. .

  Further, the predictor 109 has a function of acquiring, from the residual controller 113, a signal indicating that the residual signal has been selected and prediction auxiliary information included in the information belonging to the block including the selected residual signal. Have. When the predictor 109 acquires a signal indicating that a residual signal has been selected, the predictor 109 has a function of acquiring a block including a prediction signal associated with prediction auxiliary information from the prediction picture storage memory 110.

  Furthermore, the predictor 109 has a function of supplying to the adder 120 a block including a prediction signal corresponding to the prediction auxiliary information acquired from the predicted picture storage memory 110. Here, when the block including the prediction signal corresponding to the prediction auxiliary information cannot be acquired from the predicted picture storage memory 110, the prediction method is identified again based on the acquired prediction auxiliary information, and the prediction signal is generated. If the block including the generated prediction signal is supplied to the adder 120, the configuration is further improved.

  The predicted picture storage memory 110 has a function of acquiring a processing position control signal necessary for specifying the position of a block to be encoded from the processing position controller 106. The predicted picture storage memory 110 has a function of acquiring a block including a prediction signal from the predictor 109 and storing it as a prediction result of the position of the block to be encoded. Here, it is desirable that the predicted picture storage memory 110 has a function capable of associating and storing blocks including a plurality of predicted signals at the same block position. This is because prediction signals generated based on different prediction methods may be supplied by the predictor 109 at the same position in the plane.

  The predicted picture storage memory 110 has a function of managing information belonging to each block in the plane, which is obtained from each external unit and the state in the predicted picture storage memory 110. Here, the information belonging to the block corresponds to at least an image signal in the block, here a prediction signal. In addition, it is preferable that each block includes information indicating whether encoding is completed.

  Furthermore, the predicted picture storage memory 110 has a function of supplying information belonging to a block at a requested position in the plane to the predictor 109 and the subtractor 111. Normally, information belonging to the block to be encoded is supplied to the predictor 109. However, when the predictor 109 requests a block with prediction auxiliary information, the prediction auxiliary information is supplied. It is preferable that information belonging to the block associated with is supplied.

  The subtractor 111 acquires the original signal belonging to the block to be encoded from the picture storage memory 102 and the prediction signal belonging to the block to be encoded from the prediction picture storage memory 110. It has the function to do. The subtractor 111 has a function of generating a residual signal by subtracting the acquired prediction signal from the acquired original signal and supplying a block including the generated residual signal to the differential picture storage memory 112.

  The difference picture storage memory 112 has a function of acquiring a processing position control signal from the processing position controller 106 and specifying the position of a block to be encoded. The differential picture storage memory 112 has a function of acquiring a block including a residual signal from the subtractor 111 and storing it as a subtraction result of the position of the block to be encoded.

  Here, it is desirable that the difference picture storage memory 112 has a function capable of associating and storing blocks including a plurality of residual signals at the same block position. This is due to the prediction signals generated based on different prediction methods when the predictor 109 performs intra prediction on the same position in the plane, and the block including the plurality of different prediction signals is subtracted. This is because different residual signals may be generated by being supplied to 111 and supplied to the difference picture storage memory 112.

  The differential picture storage memory 112 has a function of managing information belonging to each block in the plane, which is obtained from each external unit and the state in the differential picture storage memory 112. Here, the information belonging to the block corresponds to at least an image signal in the block, here, a residual signal. In addition, it is preferable that each block includes information indicating whether encoding is completed. In addition, it is preferable that prediction auxiliary information is included as information belonging to each block as information for discriminating a plurality of blocks associated with the position of the same block. The difference picture storage memory 112 has a function of supplying information belonging to a block at a position in the requested plane to the residual controller 113.

The residual controller 113 has a function of acquiring a re-predictive control signal from the re-predictive controller 104. The residual controller 113 has a function of acquiring a processing position control signal from the processing position controller 106 and specifying the position of the block to be encoded. Further, the residual controller 113 has a function of acquiring information belonging to a plurality of blocks associated with the position of the block to be encoded from the differential picture storage memory 112. Further, the residual controller 113 analyzes information belonging to a plurality of blocks associated with the obtained position of the encoding target block, and uses a prediction method that has not yet been performed in the intra prediction of the predictor 109. It has the function of specifying and generating prediction auxiliary information 108 for specifying the specified prediction method and supplying it to the predictor 109. Here, the final prediction auxiliary information 114 , which is the prediction control information 108 that specifies the prediction method with the least amount of information among the residual signals generated from all the intra prediction prediction methods, is output at the time of encoding output. If the predetermined entropy encoding is separately performed and output as an encoded bit stream together with the output code 122, the configuration is further improved.

Further, the residual controller 113 needs to analyze information belonging to a plurality of blocks associated with the obtained position of the encoding target block, and generate a prediction signal by a new prediction method as a result of the analysis. If it is determined that there is no such information, it has a function of comparing the plurality of residual signals acquired and selecting the one with the smallest amount of information among the plurality of residual signals. The residual controller 113 has a function of supplying a block including the selected residual signal to the orthogonal transformer 115. In addition, after selecting the residual signal, the residual controller 113 includes a signal indicating that the residual signal is selected to the predictor 109 and information included in the block including the selected residual signal. The function of supplying the predicted auxiliary information 108 is provided.

  The orthogonal transformer 115 has a function of acquiring a block including the residual signal selected from the residual controller 113, performing a predetermined orthogonal transformation on the acquired residual signal, and generating orthogonal transformation coefficient information. Have. The orthogonal transformer 115 has a function of supplying the generated orthogonal transform coefficient information to the quantizer 116.

  The quantizer 116 acquires orthogonal transform coefficient information from the orthogonal transformer 115, performs quantization on the acquired orthogonal transform coefficient information based on a predetermined quantization parameter, and generates post-quantization information. Have Further, the quantizer 116 supplies the generated post-quantization information to the entropy encoder 121 and the inverse quantizer 118 via the switch 117.

  The switch 117 has a function of acquiring a control signal for switching the switch from the re-predictive controller 104, and connecting and disconnecting the switch according to the control signal. The post-quantization information is supplied from the quantizer 116 to the terminal of the switch 117, and the post-quantization information is supplied to the entropy encoder 121 when the switch is connected.

  The inverse quantizer 118 acquires post-quantization information from the quantizer 116, performs inverse quantization on the acquired post-quantization information based on a predetermined quantization parameter, and decodes orthogonal transform coefficient information. It has the function to generate. Further, the inverse quantizer 118 has a function of supplying the generated decoded orthogonal transform coefficient information to the inverse orthogonal transformer 119.

  The inverse orthogonal transformer 119 acquires the decoded orthogonal transform coefficient information from the inverse quantizer 118 and performs a predetermined inverse orthogonal transform on the acquired decoded orthogonal transform coefficient information to generate a decoded residual signal. It has a function. Further, the inverse orthogonal transformer 119 has a function of supplying a block including the generated decoded residual signal to the adder 120.

  The adder 120 acquires a block including a prediction signal from the predictor 109 and a block including a decoded residual signal from the inverse orthogonal transformer 119, and adds the acquired prediction signal and the decoded residual signal to add a decoded signal. It has the function to generate. The adder 120 has a function of supplying a block including the generated decoded signal to the decoded picture storage memory 107.

  The entropy encoder 121 acquires post-quantization information supplied from the quantizer 116 via the switch 117, performs predetermined entropy encoding on the acquired post-quantization information, and performs encoding. Generate a bitstream. The entropy encoder 121 has a function of outputting the generated encoded bit stream as an output signal.

  Next, details of the intra prediction processing order and the selected intra prediction method in the predictor 109 will be described with reference to FIGS. 7, 8, and 10.

  FIG. 7 is a conceptual diagram showing the order of blocks for performing intra prediction and the state of re-prediction processing according to the present invention. In the present invention, the process proceeds using a predetermined number of blocks as a unit of encoding. Here, for simplicity of explanation, as shown in FIG. 7, four blocks of block 0, block 1A, block 1B, and block 2 are handled as encoding units, but this processing unit is particularly limited. However, it may be an M × N block (M and N are integers) such as 4 horizontal blocks and 4 vertical blocks. Here, it is assumed that block 0 is a block to be encoded, and other blocks 1A, 1B, and 2 are surrounding blocks. In FIG. 7, (0/0), (0 / 1A), and the like indicate (the number of re-prediction / the block on which the current prediction is performed) in the intra prediction of the present invention.

  First, as shown in (0/0) of FIG. 7, among the four blocks to be encoded, block 0 which is the upper left block is set as a block to be encoded, and the other three blocks Let one block be a surrounding block. The decoded signal that can be used when intra prediction is performed on the block 0 to be encoded is to be used because the blocks 1A, 1B, and 2 that are surrounding blocks are not yet encoded. I can't.

  Therefore, a prediction signal is generated using a circle-shaped decoded signal written in the encoded area in (0/0) of FIG. In this case, for example, a prediction signal of block 0 is generated by using a prediction direction and a prediction method used in AVC intra prediction, which is conventional encoding. Here, in intra prediction of AVC, nine prediction modes (prediction methods) from mode 0 to mode 8 exist. As the prediction direction, eight directions as shown in FIG. 10C can be used. In addition, as shown in FIG. 10D to FIG. 10L, each prediction method may generate a prediction signal in a block using a decoded signal adjacent to the block. In this way, the prediction signal of block 0 is obtained, and coding and decoding are performed on block 0, so that the decoded signal of block 0 can be used in surrounding blocks.

  Here, when obtaining a prediction signal, normally, prediction signals in all prediction modes are generated by trying all prediction modes, that is, prediction methods, and residual signals obtained in subsequent stages are generated. It is desirable that the final prediction method is selected by selecting the residual signal generated from all the prediction modes with the least amount of information.

  Next, as in (0 / 1A) of FIG. 7, intra prediction is similarly performed on the block 1A. In order to simplify the description, the order of proceeding intra prediction here proceeds in the order of “block 0, block 1A, block 1B, block 2”, but block 1A and block 1B are in reverse order. It doesn't matter. Similarly for block 1A, a prediction signal is obtained by the normal intra prediction method as shown in FIG. 10, and the block 1A is encoded and decoded so that the decoded signal of block 1A is used in the surrounding blocks. It can be so. Similarly, in (0 / 1B) and (0/2) in FIG. 7, the same intra prediction is performed for the block 1B and the block 2 so that the decoded signal can be used.

  In the processing up to this point, the decoded signals of the block 0 that is the block to be encoded and the blocks 1A, 1B, and 2 that are the surrounding blocks can be used. As shown in 1/0), intra prediction is performed again on block 0 using a decoded signal that can be newly used from surrounding blocks. Here, the prediction signal of block 0 is updated by using the prediction method and the prediction method used in the intra prediction of AVC which is the conventional encoding, for example, as the method of intra prediction to be used.

  Further, as shown in FIG. 8, a new prediction mode may be added. Since the decoded signals of the lower, right, and lower right blocks can be used for prediction, FIG. 8 refers to nine prediction modes from mode 0 to mode 8 existing in AVC intra prediction. Modes based on vertically and horizontally symmetrical prediction directions may be prepared and used as modes 9 to 17 as shown in FIGS. By this re-prediction, the prediction signal of block 0 is updated, and the decoded signal is updated by encoding and decoding. In addition, the prediction direction in the case of a re-prediction process is shown in FIG.8 (C).

  Using the updated decoded signal of block 0, intra prediction is performed again on block 1A as shown in (1 / 1A) of FIG. As described above, by using the updated decoded signal, it is possible to determine the prediction mode and improve the accuracy of the generated prediction signal. Similarly, in (1 / 1B) and (1/2) in FIG. 7, intra prediction is again performed on the block 1B and the block 2, and the decoded signal is updated.

  Also, as shown in (N / 0) to (N / 2) in FIG. 7, intra prediction is repeated N times and re-prediction is repeated to update the prediction signal in the block to be encoded and its surrounding blocks. Thus, it is possible to generate a prediction signal with higher accuracy than the conventional intra prediction.

  Next, the operation of the embodiment of the image coding apparatus according to the present invention shown in the block diagram of FIG. 1 will be described with reference to the flowchart of FIG.

  First, it is a screen area of an image signal to be encoded input as an input signal (in this specification, an image signal includes a moving image signal, a still image signal, a continuous still image signal, and the like). The picture to be encoded is stored in the picture storage memory 102. Further, the re-prediction controller 104 disconnects the switch 117 in order to limit the encoded output until the re-prediction processing is completed (step S101).

  Subsequently, the processing position controller 106 identifies the position of the block to be encoded (step S102), and the picture storage memory 102, the re-prediction controller 104, the decoded picture storage memory 107, the predictor 109, and the predicted picture storage. By supplying the processing position control signal to the memory 110, the difference picture storage memory 112, and the residual controller 113, the position of the block to be encoded by each means is determined.

  Subsequently, the residual controller 113 acquires information belonging to a plurality of blocks associated with the position of the block to be encoded from the differential picture storage memory 112 and further becomes the acquired encoding target. The information belonging to a plurality of blocks associated with the position of the current block is analyzed, and a prediction method that has not yet been performed in the intra prediction of the predictor 109 is specified. Then, the residual controller 113 supplies the predictive auxiliary information 108 for specifying the specified prediction method to the predictor 109. Thereby, the predictor 109 acquires the prediction auxiliary information 108 from the residual controller 113 (step S103).

Then, the prediction method at the time of performing intra prediction based on the acquired prediction auxiliary information is specified. That is, the predictor 109 obtains the re-predictive control signals from the re-predictive controller 104, to identify whether to re-prediction on the block that is currently encoded object. Here, the discussion proceeds assuming that it is determined that re-prediction is necessary. If necessary, the predictor 109 acquires from the processing position controller 106, the state of the coded blocks surrounding the block as a processing position control signal which is an encoding target.

  The predictor 109 obtains a decoded signal necessary for performing intra prediction, and therefore there is a decoded signal that can be used from the decoded picture storage memory 107 in blocks around the block to be encoded. By acquiring information belonging to the block, pixels that can be used for prediction are acquired (step S104). Further, if necessary, the predictor 109 acquires information belonging to a block in which a usable prediction signal exists among the blocks around the block to be encoded from the predicted picture storage memory 110. The predictor 109 generates a prediction signal of the block to be encoded by the specified prediction method based on the decoded signal that can be used in the blocks around the acquired block to be encoded. Then, a prediction process is performed (step S105). Thereafter, the predictor 109 supplies information that belongs to the block to be encoded, including the generated prediction signal, to the predicted picture storage memory 110.

  The prediction picture storage memory 110 acquires a block including a prediction signal from the predictor 109 and stores it as a prediction result of the position of the block to be encoded. Subsequently, the subtractor 111 acquires the original signal belonging to the block to be encoded from the picture storage memory 102. Also, a prediction signal belonging to the block to be encoded is acquired from the prediction picture storage memory 110. The subtractor 111 performs difference processing by subtracting the acquired prediction signal from the acquired original signal and generating a residual signal (step S106). The subtractor 111 supplies a block including the generated residual signal to the difference picture storage memory 112. The difference picture storage memory 112 acquires a block including a residual signal from the subtractor 111 and stores it as a subtraction result of the position of the block to be encoded.

  Thereafter, the residual controller 113 acquires information belonging to a plurality of blocks associated with the position of the block to be encoded from the differential picture storage memory 112. The residual controller 113 analyzes information belonging to a plurality of blocks associated with the acquired position of the block to be encoded, and performs all prediction methods in the intra prediction of the predictor 109, here all predictions. It is determined whether the mode has been tried (step S107).

  If the residual controller 113 determines that not all prediction modes have been tried (NO in step S107), a prediction method that has not yet been performed in the intra prediction of the predictor 109 is specified, and the specified prediction is performed. By supplying the predictive auxiliary information 108 for specifying the method to the predictor 109, the predictive auxiliary information 108 is updated (step S108). Then, it returns to step S103 and performs the prediction process by a different prediction method.

When all the prediction modes are tried in the residual controller 113 (YES in step S107), the plurality of obtained residual signals are compared, and the one having the smallest information amount is selected from the plurality of residual signals. (Step S109). Thereafter, the residual controller 113 supplies a block including the selected residual signal to the orthogonal transformer 115. Also, after selecting the residual signal, the residual controller 113 outputs a signal indicating that the residual signal has been selected and final prediction auxiliary information 114 indicating the prediction mode of the selected residual signal. .

  Thereafter, the re-prediction controller 104 determines whether re-prediction processing is necessary (step S110). Here, if necessary, it is determined based on the re-prediction control information 103 whether re-prediction processing is necessary. When the re-prediction controller 104 determines that the re-prediction process is not necessary (NO in step S110), the re-prediction controller 104 connects the switch 117 so that the encoded output can be performed (step S110). S111).

  Thereafter, the orthogonal transformer 115 obtains a block including the residual signal selected from the residual controller 113, performs predetermined orthogonal transformation on the obtained residual signal (step S112), and obtains orthogonal transformation coefficient information. Is supplied to the quantizer 116. The quantizer 116 acquires the orthogonal transform coefficient information from the orthogonal transformer 115, performs quantization on the acquired orthogonal transform coefficient information based on a predetermined quantization parameter (step S113), and generates the generated post-quantization Information is supplied to the entropy encoder 121 via the switch 117.

  The entropy encoder 121 acquires post-quantization information supplied from the quantizer 116 via the switch 117, and performs predetermined entropy encoding on the acquired post-quantization information (step S114). The generated encoded bit stream is output as an output signal 122. Thereafter, the process proceeds to the inverse quantization process in step S117.

  When the re-prediction controller 104 determines that re-prediction processing is necessary (YES in step S110), the orthogonal transformer 115 acquires a block including the residual signal selected from the residual controller 113. Then, predetermined orthogonal transformation is performed on the obtained residual signal (step S115), and orthogonal transformation coefficient information is generated and supplied to the quantizer 116. The quantizer 116 acquires the orthogonal transform coefficient information from the orthogonal transformer 115, performs quantization on the acquired orthogonal transform coefficient information based on a predetermined quantization parameter (step S116), and generates the generated post-quantization Information is supplied to inverse quantizer 118.

  Thereafter, the inverse quantizer 118 acquires post-quantization information from the quantizer 116, performs inverse quantization on the acquired post-quantization information based on a predetermined quantization parameter (step S117), The generated decoded orthogonal transform coefficient information is supplied to the inverse orthogonal transformer 119. The inverse orthogonal transformer 119 acquires the decoded orthogonal transform coefficient information from the inverse quantizer 118, and performs a predetermined inverse orthogonal transform on the acquired decoded orthogonal transform coefficient information (step S118), thereby decoding the residual signal. And a block including the decoded residual signal is supplied to the adder 120.

  Thereafter, the adder 120 acquires a block including the prediction signal from the predictor 109, acquires a block including the decoded residual signal from the inverse orthogonal transformer 119, and adds the acquired prediction signal and the decoded residual signal. In step S119, a decoded signal is generated.

  Thereafter, the decoded picture storage memory 107 acquires a block including the decoded signal from the adder 120 and stores it as a decoding result of the position of the block to be encoded, thereby preparing for the next intra prediction process. Then, the processing position controller 106 determines whether or not the in-plane encoding process has been completed (step S120).

  If the processing position controller 106 determines that the in-plane processing has not yet been completed (NO in step S120), the process returns to step S101, and the image encoding process of the present embodiment is continued. On the other hand, when the processing position controller 106 determines that the in-plane processing has already been completed (YES in step S120), the image encoding processing of the present embodiment is terminated.

  The output signal 122 output from the entropy encoder 121 in this way is distributed as an encoded bit stream by transmission through a storage medium or a network, and is reproduced by a terminal device. In the case of reproduction by the terminal device, the distributed coded bit stream is reproduced after being decoded by the decoding device of the terminal device.

  As described above, according to the image coding apparatus of the present embodiment, first, the prediction signal obtained by the predetermined in-plane prediction is subtracted from the original signal of the encoding target region using the available surrounding decoded signals. Then, a residual signal is generated, and predetermined encoding and decoding are performed to generate a decoded residual signal. A decoded signal is generated by adding the generated decoded residual signal and the already obtained prediction signal. A prediction signal for the surrounding area to be encoded is generated by predetermined in-plane prediction by newly using the decoded signal for the encoding target area and the decoded signal for the surrounding area where there is a usable decoding signal. To do. Subsequently, a residual signal is generated by subtracting the generated prediction signal from the original signal in a region around the encoding target, and a decoding residual signal is generated by performing predetermined encoding and decoding. Furthermore, a decoded signal in a region around the encoding target is obtained by adding the generated decoded residual signal and the generated prediction signal.

In the present embodiment, the prediction direction of in-plane prediction, which has been conventionally limited, is obtained by using the decoded signal of the surrounding area of the encoding target obtained in this way for the in-plane prediction of the encoding target area. Therefore, it is possible to generate a prediction signal with higher prediction accuracy than a prediction signal obtained by conventional in-plane prediction. In addition, by increasing the prediction direction of in-plane prediction, generating a prediction signal with higher prediction accuracy than the conventional prediction signal, and updating the previously generated prediction signal, a new residual signal obtained in the subsequent processing Since the signal amplitude is suppressed, the encoding efficiency of the encoded bit stream obtained is improved.
By using the decoded signal of the surrounding area of the encoding target for in-plane prediction of the encoding target area, it becomes possible to increase the prediction direction of the in-plane prediction that has been limited in the past, Since it is possible to generate a prediction signal with higher prediction accuracy than the obtained prediction signal, and updating the previously generated prediction signal, the signal amplitude of a new residual signal obtained in subsequent processing is suppressed, The encoding efficiency of the encoded bit stream obtained by this embodiment can be improved.

  In addition, according to the present embodiment, the decoded signal is updated using the updated prediction signal of the encoding target region, and the decoded signal in the surrounding region where the updated decoded signal and the usable decoded signal exist are present. It is possible to generate a new prediction signal with higher prediction accuracy by performing in-plane prediction of the region around the encoding target using.

  Moreover, according to this Embodiment, the precision of a prediction signal can be improved by updating repeatedly the prediction signal of the encoding object area | region and the prediction signal of the area | region of the circumference | surroundings of encoding object sequentially. As a result, the prediction accuracy of the obtained prediction signal is improved as compared with the prediction signal of the encoding target region and the prediction signal of the surrounding region of the encoding target region. Therefore, a new residual signal signal obtained in the subsequent processing Since the amplitude is suppressed, it is possible to further improve the encoding efficiency of the encoded bit stream obtained by the present embodiment.

(Image coding program)
FIG. 3 is a block diagram showing an example of an information processing apparatus that operates according to an embodiment of an image encoding program according to the present invention. In the figure, an information processing apparatus 300 is operated by an input device 301 for inputting various types of information, an output unit 302 for outputting various types of information, and an image encoding program according to an embodiment of the present invention. A central processing control device 303, an external storage device 304, a temporary storage device 305 used for a work area in the arithmetic processing by the central processing control device 303, and a communication device 306 for communicating with the outside. The bus 307 is connected.

  The central processing control device 303 corresponds to the re-prediction controller 104 shown in FIG. 1, in which the image encoding program according to the embodiment is distributed from a recording medium or distributed via a communication network and is captured by the communication device 306. Re-predictive control means 308, processing position control means 309 corresponding to the processing position controller 106, prediction means 310 equivalent to the predictor 109, residual control means 311 equivalent to the residual controller 113, and orthogonal transformer 115 Corresponding orthogonal transformation means 312, quantization means 313 corresponding to the quantizer 116, inverse quantization means 314 equivalent to the inverse quantizer 118, inverse orthogonal transformation means 315 equivalent to the inverse orthogonal transformer 119, entropy coding Entropy encoding means 316 corresponding to the unit 121, subtraction means 317 corresponding to the subtractor 111, and addition corresponding to the adder 120 At least it has the functions of the step 318 is performed by software processing the same operation as the image encoding apparatus shown in FIG.

  Thereby, also in the image coding program of this embodiment, as with the image coding apparatus shown in FIG. 1, the blocks around the block to be coded in the same manner as the intra prediction such as AVC are used. By making the decoded signal available and then re-intra-predicting the block that was originally encoded again, a more accurate prediction signal is generated than the prediction signal generated by the first intra-prediction. To do. As a result, the prediction signal of the block surrounding the block to be encoded is also improved in accuracy by performing intra prediction again on the surrounding block that uses the block to be encoded for prediction. As a result, encoding efficiency can be improved as a result. Furthermore, by repeatedly performing such intra-prediction re-prediction, it is possible to mutually improve the accuracy of the prediction signals of the block to be encoded and its surrounding blocks.

(Image decoding device)
Next, an image decoding apparatus according to an embodiment of the present invention that acquires and decodes an encoded bitstream generated by the image encoding apparatus of the present invention or a computer that operates according to the image encoding program of the present invention. This will be described with reference to FIGS.

  FIG. 4 shows a block diagram of an embodiment of an image decoding apparatus according to the present invention. As shown in FIG. 4, the decoding apparatus 400 of this embodiment includes an entropy decoder 402, a re-prediction controller 404, a processing position controller 406, a switch 407, an inverse quantizer 408, an inverse orthogonal transformer 409, a difference A picture storage memory 410, an adder 411, a predicted picture storage memory 413, a predictor 414, and a decoded picture storage memory 415 are provided.

  The entropy decoder 402 obtains, as an input signal 401, a coded bitstream generated by a computer operating by the image coding apparatus according to one embodiment of the present invention or the image coding program according to one embodiment of the present invention. Has the function of In addition, the entropy decoder 402 performs predetermined entropy decoding on the acquired encoded bitstream to generate decoded quantized information, and the generated decoded quantized information is inversely quantized via the switch 407. And a function of supplying to the generator 408.

  The re-prediction controller 404 has a function of controlling whether to perform intra prediction again on a block that has already been subjected to intra prediction. The re-prediction controller 404 is preferably configured to have a function of performing re-prediction control of intra prediction based on a predetermined rule. Further, if necessary, the re-prediction controller 404 has a function of decoding and acquiring re-prediction control information 403 for specifying a rule for performing re-prediction control of intra prediction from the encoded bitstream. A better configuration is obtained. The re-prediction controller 404 has a function of supplying a re-prediction control signal, which is a signal for controlling re-prediction of intra prediction, to the predictor 414, and the position of the block currently being decoded from the processing position controller 406. A function of acquiring a processing position control signal for specifying the signal and a function of supplying a control signal for switching the switch 407 to the switch 407 in order to control decoding input in accordance with the necessity of re-prediction of intra prediction And have.

  The processing position controller 406 manages the order of blocks to be processed, the position of the block to be decoded, the decoding status of the blocks around the block to be decoded, and at least the block to be decoded The processing position control signal for specifying the position of the video signal is supplied to the re-prediction controller 404, the difference picture storage memory 410, the prediction picture storage memory 413, the predictor 414, and the decoded picture storage memory 415, respectively. Here, it is preferable that the processing position control signal includes a decoding state of blocks around the block to be decoded. In addition, the processing position controller 406 acquires processing position control information 405 for determining the order of blocks to be processed from the encoded bitstream, and processes the block based on the acquired processing position control information 405. If it has the function to determine an order, it will become a still better composition.

  The switch 407 has a function of acquiring a control signal for switching the switch from the re-predictive controller 404, and connecting and disconnecting the switch according to the control signal. The decoded quantized information is supplied from the entropy decoder 402 to the terminal of the switch 404, and when the switch is connected, the supplied decoded quantized information is directly sent to the inverse quantizer 408. Output.

  The inverse quantizer 408 acquires the decoded quantized information from the entropy decoder 402 via the switch 407, and performs inverse quantization on the acquired decoded quantized information based on a predetermined quantization parameter. And a function of generating decoded orthogonal transform coefficient information. The inverse quantizer 408 supplies the generated decoded orthogonal transform coefficient information to the inverse orthogonal transformer 409.

  The inverse orthogonal transformer 409 obtains decoded orthogonal transform coefficient information from the inverse quantizer 408, and performs a predetermined inverse orthogonal transform on the obtained decoded orthogonal transform coefficient information to generate a decoded residual signal. Have The inverse orthogonal transformer 409 supplies the generated block including the decoded residual signal to the difference picture storage memory 410.

  The difference picture storage memory 410 has a function of acquiring a processing position control signal from the processing position controller 406 and specifying the position of a block to be decoded. The differential picture storage memory 410 has a function of acquiring a block including a decoded residual signal from the inverse orthogonal transformer 409 and storing it as a result of the position of the block to be decoded. Further, the differential picture storage memory 410 has a function of managing information belonging to each block in the plane obtained from each external unit and the state in the differential picture storage memory 410. Here, the information belonging to the block corresponds to at least an image signal in the block, here, a residual signal. In addition, each block may include information indicating whether or not the decoding is completed. The difference picture storage memory 410 has a function of supplying the adder 411 with information belonging to the block at the requested position in the plane.

  The adder 411 acquires the block including the prediction signal from the predictor 414 and the block including the decoded residual signal from the difference picture storage memory 410, and decodes the acquired prediction signal and the decoded residual signal by adding them. It has a function to generate a signal. The adder 411 supplies the block including the generated decoded signal to the decoded picture storage memory 415.

  The predicted picture storage memory 413 obtains a processing position control signal necessary for specifying the position of the block to be decoded from the processing position controller 406, and receives a block including the prediction signal from the predictor 414. It has a function of acquiring and storing the prediction result of the position of the block to be decoded. Here, it is desirable that the predicted picture storage memory 413 has a function capable of associating and storing blocks including a plurality of predicted signals at the same block position. This is because prediction signals generated based on different prediction methods may be supplied by the predictor 414 at the same position in the plane.

  The predicted picture storage memory 413 has a function of managing information belonging to each block in the plane, which is obtained from external means and the state in the predicted picture storage memory 413. Here, the information belonging to the block corresponds to at least an image signal in the block, here a prediction signal. In addition, each block may include information indicating whether or not the decoding is completed. The predicted picture storage memory 413 supplies the information belonging to the block at the position in the requested plane to the predictor 414. Normally, information belonging to a block to be decoded is supplied to the predictor 414, but the predicted picture storage memory 413 receives a block request from the predictor 414 with prediction auxiliary information. Is preferably supplied with information belonging to the block associated with the prediction auxiliary information.

  The predictor 414 has a function of acquiring a re-prediction control signal from the re-prediction controller 404 and specifying whether to perform re-prediction on the block currently being decoded based on the acquired re-prediction control signal. Have. Further, the predictor 414 acquires a processing position control signal necessary for specifying the position of the block to be decoded from the processing position controller 406, and based on the acquired processing position control signal, It has a function of specifying the position in the plane of the block to be decoded. Further, if necessary, the predictor 414 has a function of acquiring a decoding state of blocks around the block to be decoded as a processing position control signal from the processing position controller 406. Become.

  Furthermore, since the predictor 414 obtains a decoded signal necessary for performing intra prediction, there are available decoded signals in the blocks around the block to be decoded from the decoded picture storage memory 415 and available. It has a function of acquiring information belonging to an existing block. Further, if necessary, the predictor 414 has a function of acquiring information belonging to a block in which a prediction signal that can be used exists in a block around the block to be decoded from the predicted picture storage memory 413. A better configuration is obtained.

  Furthermore, the predictor 414 has a function of acquiring prediction auxiliary information obtained by decoding the encoded bitstream, and specifying a prediction method when performing intra prediction based on the acquired prediction auxiliary information. Furthermore, the predictor 414 generates a prediction signal of the block to be decoded by the specified prediction method based on the decoded signal that can be used in the blocks around the acquired block to be decoded. Have In addition, it has a function of supplying information belonging to the block to be decoded, including the generated prediction signal, to the prediction picture storage memory 413. Here, by adding the acquired prediction auxiliary information as the information belonging to the block to be supplied, it is possible to specify the prediction method generated by the generated prediction signal, and the configuration is further improved. .

  In addition, the predictor 414 may have a function of acquiring a block including a prediction signal associated with prediction auxiliary information from the prediction picture storage memory 413 as necessary. The predictor 414 may have a function of supplying to the adder 411 a block including a prediction signal corresponding to the prediction auxiliary information acquired from the predicted picture storage memory 413. Here, when the block including the prediction signal corresponding to the prediction auxiliary information cannot be acquired from the predicted picture storage memory 413, the predictor 414 specifies the prediction method again based on the acquired prediction auxiliary information, and performs prediction. If a block including the generated prediction signal is supplied to the adder 411 after the signal is generated, the configuration is further improved.

  The decoded picture storage memory 415 has a function of acquiring the processing position control signal from the processing position controller 406 and specifying the position of the block to be decoded. The decoded picture storage memory 415 has a function of acquiring a block including a decoded signal from the adder 411 and storing it as a decoding result of the position of the block to be decoded. Here, it is desirable that the decoded picture storage memory 415 has a function capable of storing a block including a plurality of decoded signals in association with each other at the same block position. This is because a decoded signal may be generated from a prediction signal and a decoded residual signal generated based on different prediction methods at the same position in the plane by re-prediction processing and supplied to the decoded picture storage memory 415. Because.

  The decoded picture storage memory 415 has a function of managing information belonging to each block in the plane, which is obtained from each external unit and the state in the decoded picture storage memory 415. Here, information belonging to a block corresponds to at least an image signal in the block, here, a decoded signal. In addition, each block may include information indicating whether or not the decoding is completed. Further, the decoded picture storage memory 415 has a function of supplying information that belongs to the block at the position in the requested plane to the predictor 414.

  Next, the operation of the embodiment of the image decoding apparatus of the present invention shown in FIG. 4 will be described with reference to the flowchart of FIG.

  First, the processing position controller 406 obtains processing position control information 405 for determining the order of blocks to be processed by decoding from the encoded bitstream, and processes the blocks to be processed based on the acquired processing position control information 405. Determine the order. The processing position controller 406 generates at least a processing position control signal for specifying the position of a decoding target block, a re-prediction controller 404, a difference picture storage memory 410, a prediction picture storage memory 413, a predictor 414, The decoded picture storage memory 415 is supplied.

  The re-prediction controller 404 decodes and acquires re-prediction control information 403 for specifying a rule for performing re-prediction control of intra prediction, if necessary. Further, the re-prediction controller 404 acquires a processing position control signal for specifying the position of the block currently being decoded from the processing position controller 406. Further, the re-prediction controller 404 supplies a control signal for switching the switch 407 to the switch 407 in order to control the input of decoding according to the necessity of re-prediction of intra prediction. Thereafter, the switch 407 connects the switch according to the control signal notified from the re-prediction controller 404 (step S201).

  The entropy decoder 402 obtains, as an input signal 401, an encoded bit stream generated by a computer operating by the image encoding device of the embodiment of the present invention or the image encoding program of the embodiment of the present invention. To do. The entropy decoder 402 performs predetermined entropy decoding on the acquired encoded bitstream (step S202), and generates decoded quantized information. The entropy decoder 402 supplies the generated decoded quantized information to the inverse quantizer 408 via the switch 407.

  The inverse quantizer 408 acquires the decoded quantized information from the entropy decoder 402 via the switch 407, and performs inverse quantization on the acquired decoded quantized information based on a predetermined quantization parameter. (Step S203), the generated decoded orthogonal transform coefficient information is supplied to the inverse orthogonal transformer 409.

  The inverse orthogonal transformer 409 acquires the decoded orthogonal transform coefficient information from the inverse quantizer 408, performs predetermined inverse orthogonal transform on the acquired decoded orthogonal transform coefficient information (step S204), and generates the generated decoded residual signal. Are supplied to the differential picture storage memory 410.

  The difference picture storage memory 410 acquires the processing position control signal from the processing position controller 406 and specifies the position of the block to be decoded. Further, the difference picture storage memory 410 acquires a block including a decoded residual signal from the inverse orthogonal transformer 409, and stores it as a result of the position of the block to be decoded.

  The predictor 414 specifies whether or not to perform re-prediction on the block currently being decoded based on the re-prediction control signal acquired from the re-prediction controller 404. Further, the predictor 414 acquires a processing position control signal necessary for specifying the position of the block to be decoded from the processing position controller 406, and based on the acquired processing position control signal, The position in the plane of the block to be decoded is specified (step S205).

  Subsequently, the predictor 414 acquires prediction auxiliary information obtained by decoding the encoded bitstream (step S206), and specifies a prediction method for performing intra prediction based on the acquired prediction auxiliary information. Subsequently, the predictor 414 obtains information that belongs to a block around the block to be decoded from the decoded picture storage memory 415 and belongs to a block in which a usable decoded signal exists, thereby performing intra prediction. A decoded signal necessary for performing, that is, a pixel that can be used for prediction is acquired (step S207).

  Further, if necessary, the predictor 414 acquires information belonging to a block in which a usable prediction signal exists among the blocks around the block to be decoded from the predicted picture storage memory 413. The predictor 414 generates a prediction signal of the block to be decoded by the specified prediction method based on the decoded signal that can be used in the blocks around the acquired block to be decoded, A prediction process is performed (step S208). Also, the predictor 414 supplies information belonging to the block to be decoded, including the generated prediction signal, to the predicted picture storage memory 413.

  Further, the predictor 414 acquires a block including a prediction signal associated with the prediction auxiliary information from the predicted picture storage memory 413 as necessary, and adds a block including the prediction signal corresponding to the prediction auxiliary information to the adder. 411 is supplied. Here, when the predictor 414 cannot acquire the block including the prediction signal corresponding to the prediction auxiliary information from the predicted picture storage memory 413, the predictor 414 specifies the prediction method again based on the acquired prediction auxiliary information, and performs prediction. After generating the signal, a block including the generated prediction signal is supplied to the adder 411.

  The adder 411 obtains a block including the prediction signal from the predictor 414 and a block including the decoded residual signal from the difference picture storage memory 410 and adds them (step S209), thereby generating a decoded signal. . The decoded picture storage memory 415 acquires the processing position control signal from the processing position controller 406 and specifies the position of the block to be decoded. Also, the decoded picture storage memory 415 acquires a block including a decoded signal from the adder 411 and stores it as a decoding result of the position of the block to be decoded.

  Thereafter, the re-prediction controller 404 determines whether re-prediction processing is necessary for the block currently being decoded based on the re-prediction control information 403 (step S210). If the re-prediction controller 404 determines that re-prediction processing is necessary (YES in step S210), the re-prediction controller 404 supplies a control signal for disconnecting the switch to the switch 407, and the switch 407 is cut (step S211). Then, it returns to step S205 and continues a prediction process.

  When the re-prediction controller 404 determines that re-prediction processing is not necessary (NO in step S210), the processing position controller 406 confirms the current decoding state and whether or not the in-plane decoding processing is completed. Is determined (step S212). If the processing position controller 406 determines that the in-plane processing has not been completed (NO in step S212), the process returns to step S201, and after the connection processing of the switch 407, the predetermined decoding processing is continued.

On the other hand, if the processing position controller 406 determines that the in-plane processing has been completed (YES in step S212), the decoded picture stored in the decoded picture storage memory 415 is output as the output signal 416 , and the decoding process ends. To do.

  Thus, according to the image decoding apparatus of the present embodiment, when performing intra prediction of a block to be decoded, attention is paid to the block to be decoded and its surrounding blocks, and decoding is performed first. Perform predetermined intra prediction on the target block using the decoded signals of the left, upper left, upper, and upper right blocks for prediction processing, and perform predetermined decoding to obtain a residual signal Thus, a decoded signal of the block to be decoded is generated.

  Subsequently, by using the decoded signal of the block to be decoded, predetermined intra prediction is sequentially performed on blocks around the block to be decoded, and predetermined decoding for obtaining a residual signal is performed. I do. After that, since the decoded signal of the surrounding block can be used in the block to be decoded, the prediction signal with higher accuracy can be generated again by performing the intra prediction again by using the surrounding decoded signal. The decoded signal can be obtained by adding the residual signal. Therefore, even an encoded bitstream encoded by the image encoding device of the present invention can be correctly decoded by the image decoding device of the present embodiment, and the quality is good even if the code amount is small as a whole. A decoded image can be obtained.

(Image decoding program)
FIG. 6 shows a configuration diagram of an example of an information processing apparatus that operates according to an embodiment of an image decoding program according to the present invention. In the figure, an information processing device 600 is operated by an input device 601 for inputting various types of information, an output device 602 for outputting various types of information, and an image decoding program according to an embodiment of the present invention. A central processing control device 603, an external storage device 604, a temporary storage device 605 used as a work area for arithmetic processing by the central processing control device 603, and a communication device 606 for communicating with the outside are bidirectional. It is configured to be connected by a bus 607.

  The central processing control device 603 receives the image decoding program according to the embodiment of the present invention from a recording medium or via a communication network and is captured by the communication device 606, and is stored in the entropy decoder 402 illustrated in FIG. Corresponding entropy decoding means 608, re-prediction control means 609 corresponding to the re-prediction controller 404, processing position control means 610 corresponding to the processing position controller 406, inverse quantization means 611 equivalent to the inverse quantizer 408, inverse 4 has at least the functions of an inverse orthogonal transform means 612 corresponding to the orthogonal transformer 409, a prediction means 613 corresponding to the predictor 414, and an adder means 614 corresponding to the adder 411. The same operation as that of the image decoding apparatus is executed by software processing.

  As a result, the encoded bitstream encoded by the image encoding device of the present invention can be encoded by the image decoding program of the present embodiment, similarly to the image decoding device of the embodiment shown in FIG. It is possible to correctly decode, and it is possible to obtain a high-quality decoded image even if the code amount as a whole is small.

  In the above description of the embodiment, the method such as the prediction mode of intra prediction has been described based on the H.264 / AVC encoding or decoding apparatus, the method, and the program thereof. Not only, but also other intra prediction methods, orthogonal transforms, quantization, inverse quantization, inverse orthogonal transforms, entropy encoding, and other components such as MPEG1, MPEG2, MPEG4, etc. Applicable to any encoding method, device or decoding method, device, etc. that, when performing intra prediction, generates a decoded signal of a surrounding block and then updates the prediction signal of the block to be encoded again by re-prediction It is.

  In addition, the image decoding apparatus of the present invention can be applied to encoded bit streams other than the encoded bit stream obtained by encoding with the image encoding apparatus of FIG. 1 or the computer of FIG. is there.

It is a block diagram of one embodiment of an image coding device of the present invention. It is a flowchart for operation | movement description of one Embodiment of the image coding apparatus of this invention. It is a block diagram of an example of the information processing apparatus which operate | moves by one Embodiment of the image coding program of this invention. It is a block diagram of one embodiment of an image decoding device of the present invention. It is a flowchart for operation | movement description of one Embodiment of the image decoding apparatus of this invention. It is a block diagram of an example of the information processing apparatus which operate | moves by one Embodiment of the image decoding program of this invention. It is the conceptual diagram which showed the order of the block which performs the intra prediction of this invention, and the mode of a re-prediction process. In the intra prediction method of this invention, it is the conceptual diagram which showed an example of the mode and prediction direction of the newly added intra prediction. It is the conceptual diagram which showed the processing order of the block in the case of performing the conventional intra prediction. It is the conceptual diagram which showed the mode and prediction direction of the intra prediction in the conventional intra prediction method.

Explanation of symbols

DESCRIPTION OF SYMBOLS 100 Image coding apparatus 101, 401 Input signal 102 Picture storage memory 103, 403 Re-prediction control information 104, 404 Re-prediction controller 105, 405 Processing position control information 106, 406 Processing position controller 107, 415 Decoded picture storage memory 108 , 114, 412 Predictive auxiliary information 109, 414 Predictor 110, 413 Predictive picture storage memory 111 Subtractor 112, 410 Differential picture storage memory 113 Residual controller 115 Orthogonal transformer 116 Quantizer 117, 407 Switch 118, 408 Inverse Quantizer 119, 409 Inverse orthogonal transformer 120, 411 Adder 121 Entropy encoder 122, 416 Output signal 400 Image decoding device 402 Entropy decoder

Claims (8)

One screen area of the input encoding target image signal is divided into predetermined rectangular areas, and the encoding target area is set as a rectangular area to be processed in the divided rectangular area. An image that performs a predetermined in-plane prediction and generates a prediction signal for an original signal of an image belonging to a conversion target region using a decoded signal of a rectangular region around the rectangular region to be processed An encoding device comprising:
Processing position control means for managing the position of the encoding target area in the one-screen area and outputting a processing position control signal for indicating the position ;
To the original signal of an image belonging to a predetermined one of the rectangular area of the rectangular area, it performs the predetermined intra prediction using the decoded signal of the rectangular area surrounding the rectangular region, a plurality of predetermined Among the prediction methods, a prediction signal in the predetermined rectangular area is generated by a prediction method specified based on the supplied prediction auxiliary information, and the prediction signal generated and the prediction used for generating the prediction signal Prediction means for outputting information indicating a method ;
From the original signal of the image in the predetermined one of the rectangular regions, and generates a Rizan difference signal by the subtracting the prediction signal supplied from the prediction unit, the prediction method supplied from the prediction means Based on the information indicated, an unused prediction method is identified from among the plurality of prediction methods, and the process of supplying the identified prediction information as prediction auxiliary information to the prediction means to generate a prediction signal is repeated. When there is no prediction method, residual signal generation / selection means for selecting a residual signal based on a predetermined evaluation method from a plurality of residual signals obtained corresponding to the plurality of prediction methods ;
Orthogonal transform is performed on the selected residual signal in the predetermined rectangular area to generate orthogonal transform coefficient information, and the orthogonal transform coefficient information is quantized to generate post-quantization information. Post-quantization information generation means;
Encoding means for performing entropy encoding on the post-quantization information in the predetermined one rectangular area;
Decoding residual signal generation means for generating a decoding residual signal by performing inverse quantization and inverse orthogonal transformation on the post-quantization information;
Adding means for adding the decoded residual signal and the prediction signal in the predetermined rectangular area to generate a decoded signal in the predetermined rectangular area;
Based on the processing position control signal from the processing position control means , based on a predetermined rule whether or not to perform the prediction process again on the rectangular area where the prediction signal has already been generated and encoded. and the re-predictive control means that controls Te,
Have
The reprediction control means controls the prediction signal generation means, the residual signal generation / selection means, the post-quantization information generation means, the decoded residual signal generation means, and the addition means, Setting the encoding target area as one rectangular area and generating the prediction signal, the residual signal, the post-quantization information, the decoded residual signal, and the decoded signal in the encoding target area; Set one rectangular area in which encoding is not completed among rectangular areas around the encoding target area as the predetermined one rectangular area, and the prediction signal, the residual signal, After generating the post-quantization information, the decoded residual signal, and the decoded signal, the encoded image signal in the encoding target area is generated around the encoding target area around which the decoded signal is generated. Rectangle By using the decoded signal of the region, updating the prediction signal, the residual signal, the post-quantization information, the decoded residual signal, and the decoded signal in the encoding target region, respectively, by the encoding unit An image encoding apparatus that performs re-encoding. The re-prediction control means includes
By sequentially setting all surrounding rectangular areas that have not been encoded among the rectangular areas around the encoding target area, the prediction signal, the residual signal, and the post-quantization information in each rectangular area Then, after generating the decoded residual signal and the decoded signal, respectively, the decoded signal of each rectangular area around the encoding target area is used for the encoded image signal of the encoding target area. And updating the prediction signal, the residual signal, the post-quantization information, the decoded residual signal, and the decoded signal in the encoding target region, and then performing re-encoding by the encoding unit. The image encoding device according to claim 1. The re-prediction control means includes
Updating of the prediction signal, the residual signal, the post-quantization information, the decoded residual signal, and the decoded signal in the encoding target region, and encoding of the rectangular region around the encoding target region is completed. The update of the prediction signal, the residual signal, the post-quantization information, the decoded residual signal, and the decoded signal in a surrounding rectangular area that is not performed is performed a set number of times. Item 3. The image encoding device according to Item 2. One screen area of the input encoding target image signal is divided into predetermined rectangular areas, and the encoding target area is set as a rectangular area to be processed in the divided rectangular area. An operation for performing a predetermined in-plane prediction and generating a prediction signal for an original signal of an image belonging to a conversion target region using a decoded signal of a rectangular region around the rectangular region to be processed An image encoding program for causing a computer to execute
The computer,
Processing position control means for managing the position of the encoding target area in the one-screen area and outputting a processing position control signal for indicating the position ;
To the original signal of an image belonging to a predetermined one of the rectangular area of the rectangular area, it performs the predetermined intra prediction using the decoded signal of the rectangular area surrounding the rectangular region, a plurality of predetermined Among the prediction methods, a prediction signal in the predetermined rectangular area is generated by a prediction method specified based on the supplied prediction auxiliary information, and the prediction signal generated and the prediction used for generating the prediction signal Prediction means for outputting information indicating a method ;
From the original signal of the image in the predetermined one of the rectangular regions, and generates a Rizan difference signal by the subtracting the prediction signal supplied from the prediction unit, the prediction method supplied from the prediction means Based on the information indicated, an unused prediction method is identified from among the plurality of prediction methods, and the process of supplying the identified prediction information as prediction auxiliary information to the prediction means to generate a prediction signal is repeated. When there is no prediction method, residual signal generation / selection means for selecting a residual signal based on a predetermined evaluation method from a plurality of residual signals obtained corresponding to the plurality of prediction methods ;
Orthogonal transform is performed on the selected residual signal in the predetermined rectangular area to generate orthogonal transform coefficient information, and the orthogonal transform coefficient information is quantized to generate post-quantization information. Post-quantization information generation means;
Encoding means for performing entropy encoding on the post-quantization information in the predetermined one rectangular area;
Decoding residual signal generation means for generating a decoding residual signal by performing inverse quantization and inverse orthogonal transformation on the post-quantization information;
Adding means for adding the decoded residual signal and the prediction signal in the predetermined rectangular area to generate a decoded signal in the predetermined rectangular area;
Based on the processing position control signal from the processing position control means , based on a predetermined rule whether or not to perform the prediction process again on the rectangular area where the prediction signal has already been generated and encoded. to function as a re-predictive control means that controls Te, wherein the re-predictive control unit, the prediction signal at the encoding target area by setting the encoding target area as the first predetermined one of the rectangular region, The residual signal, the post-quantization information, the decoded residual signal, and the decoded signal are generated, and then encoding is completed among the rectangular regions around the encoding target region as the predetermined one rectangular region After setting one rectangular area that has not been generated and generating the prediction signal, the residual signal, the post-quantization information, the decoded residual signal, and the decoded signal in the rectangular area, the encoding target area Sign of For the encoded image signal, using the decoded signal in the rectangular area around the encoding target area that generated the decoded signal, the prediction signal in the encoding target area, the residual signal, An image encoding program, wherein after the quantization, the decoded residual signal, and the decoded signal are updated, re-encoding is performed by the encoding unit. One screen area of the image signal to be encoded is divided into predetermined rectangular areas, and the encoding target area is a rectangular area to be processed in the divided rectangular area, and the encoding target area A coded bit obtained by performing a predetermined in-plane prediction using a decoded signal of a rectangular area around the rectangular area to be processed with respect to an original signal of an image belonging to An image decoding apparatus that performs a decoding operation on a stream and outputs a decoded image signal,
Entropy decoding means for decoding at least post-quantization information obtained by performing orthogonal transform and inverse quantization on the residual signal in each rectangular region from the encoded bitstream;
Decoding residual signal generating means for generating a decoding residual signal of each rectangular region by performing inverse quantization and inverse orthogonal transform processing on the decoded post-quantization information;
Processing position control means for managing the position of a rectangular area to be decoded for generating a decoded signal and outputting a processing position control signal for indicating the position ;
Based on the processing position control signal from the processing position control means, the position of the rectangular area that is currently the decoding target is specified, and a predetermined surface is used by using the decoding signal of the area around the rectangular area that is the decoding target. The prediction signal in the rectangular area is generated by a prediction method specified based on the supplied prediction auxiliary information among a plurality of predetermined prediction methods, and the generated prediction signal and the prediction Prediction means for outputting information indicating a prediction method used for signal generation ;
Adding means for adding the prediction signal and the decoded residual signal to generate the decoded signal;
Re-prediction control means for controlling whether or not to execute the prediction process again on the rectangular area that has already generated the prediction signal based on a predetermined rule, and the prediction means by the re-prediction control means, The decoding residual signal generating means and the adding means are controlled to generate the prediction signal, the decoding residual signal, and the decoding signal in the rectangular area to be decoded, and subsequently become the decoding target. Setting one rectangular area that has not been decoded among rectangular areas around the rectangular area, and generating the prediction signal, the decoded residual signal, and the decoded signal in the rectangular area, For the image signal of the rectangular area to be decoded, the prediction in the rectangular area to be decoded is performed using the decoded signal of the surrounding rectangular area that has generated the decoded signal. An image decoding apparatus that updates a signal, the decoded residual signal, and the decoded signal, respectively. The re-prediction control means includes
By sequentially setting all surrounding rectangular areas that have not been decoded among the rectangular areas around the rectangular area to be decoded, the prediction signal, the decoding residual signal, and the After generating each decoded signal, for the decoded image signal of the rectangular area to be decoded, using the decoded signal of each rectangular area around the rectangular area to be decoded, The image decoding apparatus according to claim 5, wherein the prediction signal, the decoded residual signal, and the decoded signal in the rectangular area to be decoded are updated. The re-prediction control means includes
The prediction signal, the decoding residual signal, and the decoding signal in the rectangular area to be decoded are updated, and the surroundings of the rectangular area around the rectangular area to be decoded that have not been decoded The image decoding apparatus according to claim 6, wherein the prediction signal, the decoded residual signal, and the update of the decoded signal in the rectangular area are repeatedly performed a set number of times. An image decoding process for performing a decoding operation on an encoded bitstream generated by a computer operating with the image encoding apparatus according to claim 1 or the image encoding program according to claim 4 and outputting a decoded image signal An image decoding program for causing a computer to execute
The computer,
Entropy decoding means for decoding at least post-quantization information obtained by performing orthogonal transform and inverse quantization on the residual signal in each rectangular region from the encoded bitstream;
Decoding residual signal generating means for generating a decoding residual signal of each rectangular region by performing inverse quantization and inverse orthogonal transform processing on the decoded post-quantization information;
Processing position control means for managing the position of a rectangular area to be decoded for generating a decoded signal and outputting a processing position control signal for indicating the position ;
Based on the processing position control signal from the processing position control means, the position of the rectangular area that is currently the decoding target is specified, and a predetermined surface is used by using the decoding signal of the area around the rectangular area that is the decoding target. The prediction signal in the rectangular area is generated by a prediction method specified based on the supplied prediction auxiliary information among a plurality of predetermined prediction methods, and the generated prediction signal and the prediction Prediction means for outputting information indicating a prediction method used for signal generation ;
Adding means for adding the prediction signal and the decoded residual signal to generate the decoded signal;
It functions as a re-prediction control means for controlling whether or not to execute the prediction process again on the rectangular area that has already generated the prediction signal , based on a predetermined rule, and the prediction means by the re-prediction control means Controlling the decoded residual signal generating means and the adding means to generate the prediction signal, the decoded residual signal, and the decoded signal in the rectangular area that is the decoding target; After setting the one rectangular area that has not been decoded among the rectangular areas around the rectangular area that is, and generating the prediction signal, the decoded residual signal, and the decoded signal in the rectangular area, For the image signal in the rectangular area that is the decoding target, the decoded signal in the surrounding rectangular area that generated the decoded signal is used to An image decoding program that updates the prediction signal, the decoded residual signal, and the decoded signal, respectively.