JP2006304107A - Encoding device and program applied to the encoding device - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an encoding device capable of selecting an optimal prediction mode in intra-frame prediction encoding, and a program applied to the encoding device.
An intra-frame prediction predicted image generation unit that generates a prediction block based on pixels included in an encoding target image according to a plurality of types of prediction modes, and adjacent to the encoding target block in the encoding target image A first prediction mode identified by a prediction mode applied when an adjacent block that is a block to be encoded is encoded, and a prediction error that is a difference between the intra-frame prediction prediction block and the encoding target block, and orthogonally transform, The rate-distortion optimization method is applied to the second prediction mode, which is the prediction mode that minimizes the sum of the absolute values of the values obtained by the orthogonal transform, which is the sum of the absolute values of the orthogonal transform values (SATD). The encoding apparatus includes an encoding control unit 101 that selects a prediction mode to be applied when encoding a block.
[Selection] Figure 1

Description

  The present invention relates to an encoding device that encodes a block by intra-frame prediction, and a program applied to the encoding device.

  Conventionally, H.M. A technique for dividing a frame to be encoded (hereinafter referred to as an encoding target frame) into a plurality of blocks, such as H.264 / AVC, and encoding the encoding target frame for each block is generally widely known. ing. In intra-frame prediction encoding (intra prediction encoding), prediction is performed using a pixel value of a block to be encoded (hereinafter referred to as an encoding target block) and a pixel located in the vicinity of the encoding target block. The difference from the pixel value of the predicted block is encoded.

  Specifically, in intra-frame predictive coding, there are nine types of prediction modes (prediction mode 0 to prediction mode 8) applicable when generating a prediction block. In addition, when the prediction mode selected according to the encoding target block matches the prediction mode of a block adjacent to the encoding target block (an already encoded block), a flag indicating that the two match By generating only this, the bit amount (generated bit amount) generated when encoding the block to be encoded is reduced, and the encoding efficiency (ratio between the SNR of luminance and the generated bit amount) is increased. .

  As described above, the selection of the prediction mode is important for improving the coding efficiency, and an optimal prediction mode selection method is required.

  As a prediction mode selection method, a reference software called “Joint Model” is used to select a prediction mode giving priority to image quality (High Complex Mode), and a method to select a prediction mode giving priority to processing speed (Low Complex Mode) is known.

  In a method (High Complex Mode) in which image quality is prioritized, an encoding target block is encoded according to all prediction modes, and a generated bit amount (Rate), an encoding target block before encoding, and encoding after encoding are performed. A difference from the target block is calculated. In addition, a prediction mode is selected using a cost value calculated based on the generated bit amount (Rate) and the difference (Distortion) as a criterion (Rate-Distortion optimization method).

  On the other hand, in the method (Low Complex Mode) in which the processing speed is prioritized, the difference (prediction error) between the prediction block generated according to all prediction modes and the encoding target block is calculated. Next, a sum of absolute values of prediction errors (SAD; Sum of Absolute Difference) or a sum of values obtained by orthogonal transform of prediction errors (SATD: Sum of Absolute Transformed Difference) is calculated. Furthermore, the prediction mode is selected using a cost value calculated based on SAD (or SATD) and the bit amount for specifying the prediction mode as a criterion.

  Here, in the method (High Complex Mode) in which image quality is prioritized, it is necessary to encode the block to be encoded in accordance with all prediction modes, and the amount of calculation is enormous, so the amount of calculation is limited. Then, this method cannot be used. In addition, the circuit scale increases.

  On the other hand, since the rate-distortion optimization method is not used in the method in which the processing speed is given priority (Low Complex Mode), the amount of calculation and the circuit are reduced compared to the method in which the image quality is given priority (High Complex Mode). Although the scale can be reduced, the encoding efficiency is lowered.

  As described above, the method in which image quality is given priority (High Complex Mode) and the method in which processing speed is given priority (Low Complex Mode) have a trade-off relationship in terms of the amount of computation (processing speed) and the like and encoding efficiency. Have

  Furthermore, as a method for balancing the amount of computation and the coding efficiency, a method for reducing the number of prediction modes to which the Rate-Distortion optimization method is applied has been proposed (for example, Non-Patent Document 1).

  Specifically, in this method, a prediction mode with the smallest sum of absolute values of prediction errors (SAD) and a prediction mode of a block adjacent to the upper side or the left side of the encoding target block (an already encoded block) are selected. Among these, the prediction mode is selected by applying the rate-distortion optimization method to the prediction mode with the smallest prediction mode number.

As described above, the above-described method has the Rate mode in which the sum of absolute values of prediction errors (SAD) is the smallest in terms of the generation amount of bits if the sum of absolute values of prediction errors (SAD) is small. -Distortion optimization method is applied.
Koichi Takagi and two others, “H.264 | A Study on Faster Intra Prediction in AVC”, Information Processing Society of Japan, December 2003, Vol. 2003 No. 125 (2003-AVM-43)

  However, in the above-described method, even when the sum of absolute values of prediction errors (SAD) is small, if the prediction error is not uniform, the high-frequency component obtained when orthogonally transforming the prediction error increases. The amount of generated bits also increases. That is, in the method described above, the optimal prediction mode may not be selected.

  Therefore, the present invention has been made to solve the above-described problem, and is applied to an encoding device capable of selecting an optimal prediction mode in intra-frame prediction encoding, and the encoding device. The purpose is to provide a program.

  According to a first aspect of the present invention, there is provided an encoding device that encodes an encoding target block that is a block included in an encoding target image using pixels included in the encoding target image to be encoded. A prediction block generation unit (intraframe prediction prediction image generation unit 108) that generates a prediction block (intraframe prediction prediction block) based on pixels included in the encoding target image according to a plurality of types of prediction modes; A prediction error that is a difference between the prediction block generated by the generation unit and the encoding target block is orthogonally transformed (Hadamard transformed), and an orthogonal transformation value summation that is a summation of absolute values of values obtained by the orthogonal transformation ( A calculation unit (encoding control unit 101) that calculates (SATD) for each prediction mode, and an adjacent block that is a block adjacent to the encoding target block in the encoding target image. Rate-Distortion optimization method into the first prediction mode specified by the prediction mode applied at the time of encoding and the second prediction mode that is the prediction mode in which the sum of the orthogonal transformation values calculated by the calculation unit is minimized And a prediction mode selection unit (encoding control unit 101) that selects a prediction mode to be applied when encoding the encoding target block.

  According to such a feature, the prediction mode selection unit applies the rate-distortion optimization method to the first prediction mode specified by the prediction mode of the adjacent block and the second prediction mode that minimizes the SATD, By selecting a prediction mode to be used when encoding the encoding target block, the encoding device can select an optimal prediction mode in intra-frame prediction encoding. That is, the encoding apparatus can select an optimal prediction mode even when the prediction error is not uniform.

  A second feature of the present invention is that, in the first feature of the present invention, the calculation unit calculates a prediction error sum (SAD), which is a sum of absolute values of prediction errors, for each prediction mode, and the prediction mode selection unit In addition to the first prediction mode and the second prediction mode, the rate-distortion optimization method is applied to the third prediction mode that is the prediction mode in which the total prediction error calculated by the calculation unit is minimized, The gist is to select a prediction mode to be applied when a block is encoded.

  The third feature of the present invention is applied to an encoding device that encodes an encoding target block, which is a block included in the encoding target image, using pixels included in the encoding target image to be encoded. Is a difference between the prediction block generated in step A and the encoding target block, which is generated in step A based on the pixels included in the encoding target image according to a plurality of types of prediction modes. Step B in which the prediction error is orthogonally transformed and the orthogonal transformation value sum, which is the sum of the absolute values of the values obtained by the orthogonal transformation, is calculated for each prediction mode, and adjacent to the coding target block in the coding target image The first prediction mode specified by the prediction mode applied when the adjacent block which is a block is encoded, and the second prediction mode in which the sum of the orthogonal transformation values calculated in the process B is minimized. In the measurement mode by applying the Rate-Distortion optimization method, and summarized in that to execute a step C of selecting a prediction mode to be applied when coding the coding target block on the computer.

  An object of the present invention is to provide an encoding device capable of selecting an optimal prediction mode in intra-frame prediction encoding, and a program applied to the encoding device.

[First Embodiment]
(Configuration of Encoding Device According to First Embodiment of the Present Invention)
Hereinafter, the configuration of the encoding device according to the first embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing a configuration of an encoding apparatus 100 according to the first embodiment of the present invention.

  As illustrated in FIG. 1, the encoding device 100 includes an encoding control unit 101, a difference calculation unit 102, a motion vector detection unit 103, an orthogonal transformation / quantization unit 104, an entropy encoding unit 105, and an inverse unit. A quantization unit / inverse orthogonal transform 106, an addition unit 107, an intra-frame prediction predicted image generation unit 108, a deblocking filter 109, an inter-frame prediction predicted image generation unit 110, and a switch 111 are included.

  An input signal that is an image signal is input to the encoding control unit 101 and the motion vector detection unit 103. Specifically, an input signal corresponding to an image to be encoded (hereinafter referred to as an encoding target image) is input to the encoding control unit 101 and the motion vector detection unit 103.

  The encoding control unit 101 determines the encoding method of the encoding target image according to the input signal corresponding to the encoding target image. Specifically, when the encoding target image is an image encoded by inter-frame prediction (bidirectional prediction or forward prediction), the encoding control unit 101 switches the switch 111 to thereby calculate a difference calculating unit. 102 and the inter-frame prediction predicted image generation unit 110 are connected.

  On the other hand, when the encoding target image is an image encoded by intra-frame prediction, the encoding control unit 101 switches the switch 111 to change the difference calculation unit 102 and the intra-frame prediction predicted image generation unit 108. And connect.

  In addition, when the encoding target image is an image that is encoded by intra-frame prediction, the encoding control unit 101 uses prediction modes corresponding to a plurality of types (in this embodiment, nine types) of prediction directions. Select one of the prediction modes from among the blocks.

  Specifically, the encoding control unit 101 uses a prediction mode applied to a block (hereinafter referred to as an encoding target block) adjacent to an encoding target block (hereinafter referred to as an encoding target block) in the encoding target image. Based on this, the first prediction mode is specified. For example, the encoding control unit 101 first selects a prediction mode having a small mode number from the prediction mode of the adjacent block adjacent to the left side of the encoding target block and the prediction mode of the adjacent block adjacent to the upper side of the encoding target block. Specified as prediction mode.

  In addition, the encoding control unit 101 generates an intra-frame prediction predicted image (hereinafter, intra-frame prediction prediction block) to be used in intra-frame prediction according to all prediction modes (9 types of prediction modes). The difference calculation unit 102 is instructed to generate a difference (hereinafter, prediction error) between the prediction block for intra-frame prediction and the encoding target block according to all prediction modes (9 types of prediction modes). Instruct.

  Furthermore, the encoding control unit 101 calculates a sum of absolute values (SATD: Sum of Absolute Transformed Difference) of values obtained by performing orthogonal transform (Hadamard transform) on the prediction error calculated by the difference calculation unit 102. In addition, the encoding control unit 101 specifies the prediction mode in which the calculated SATD is minimized as the second prediction mode.

  At this time, the encoding control unit 101 instructs the difference calculation unit 102 to input the prediction error calculated according to the first prediction mode and the second prediction mode to the orthogonal transform / quantization unit 104.

  Also, the encoding control unit 101 applies the rate-distortion optimization method to the identified first prediction mode and second prediction mode, and calculates cost values corresponding to the first prediction mode and the second prediction mode. Also, the encoding control unit 101 selects a prediction mode to be applied when encoding the block to be encoded based on the calculated cost value. Note that the encoding control unit 101 instructs the entropy encoding unit 105 to output a prediction error calculated according to the selected prediction mode as an output signal.

  When the encoding target block is encoded by inter-frame prediction, the difference calculation unit 102 predicts the prediction error between the inter-frame prediction prediction block acquired from the inter-frame prediction predicted image generation unit 110 and the encoding target block. Is calculated.

  Further, when the encoding target block is encoded by intra-frame prediction, the difference calculation unit 102 obtains from the intra-frame prediction predicted image generation unit 108 according to all prediction modes (9 types of prediction modes). Prediction errors between the intra-frame prediction block and the encoding target block are calculated. The difference calculation unit 102 inputs the prediction error calculated according to each prediction mode to the encoding control unit 101.

  Further, the difference calculation unit 102 inputs the prediction error calculated according to the prediction mode (first prediction mode and second prediction mode) instructed by the encoding control unit 101 to the orthogonal transform / quantization unit 104.

  The motion vector detection unit 103 calculates a motion vector used for encoding the encoding target block based on the encoded block and the encoding target block acquired from the deblock filter 109. In addition, the motion vector detection unit 103 inputs the calculated motion vector to the inter-frame prediction predicted image generation unit 110 and the entropy encoding unit 105.

  The orthogonal transform / quantization unit 104 performs orthogonal transform on the prediction error acquired from the difference calculation unit 102, and quantizes a value obtained by orthogonal transform of the prediction error. Further, the orthogonal transform / quantization unit 104 inputs the quantized prediction error (quantized prediction error) to the entropy coding unit 105 and the inverse quantization unit / inverse orthogonal transform 106.

  The entropy encoding unit 105 encodes the quantization prediction error acquired from the orthogonal transform / quantization unit 104 and the motion vector acquired from the motion vector detection unit 103. Specifically, the entropy encoding unit 105 encodes the quantization prediction error and the motion vector according to CAVLC (Context-Adaptive Variable Length Coding) or CABAC (Context-Adaptive Binary Arithmetic Coding).

  The inverse quantization unit / inverse orthogonal transform 106 inversely quantizes the quantization prediction error acquired from the orthogonal transform / quantization unit 104, and performs inverse orthogonal transform on a value obtained by inverse quantization of the quantization prediction error. Further, the inverse quantization unit / inverse orthogonal transform 106 inputs a value (ie, prediction error) obtained by the inverse orthogonal transform to the adding unit 107.

  When the encoding target block is encoded by inter-frame prediction, the addition unit 107 converts the prediction block for inter-frame prediction acquired from the inter-frame prediction predicted image generation unit 110 into an inverse quantization unit / inverse orthogonal transform. The prediction error acquired from 106 is added to generate an encoded block. In addition, the addition unit 107 inputs the generated encoded block to the deblock filter 109.

  On the other hand, when the encoding target block is encoded by intra-frame prediction, the addition unit 107 converts the inverse quantization unit / inverse to the intra-frame prediction prediction block acquired from the intra-frame prediction predicted image generation unit 108. A prediction error acquired from the orthogonal transform 106 is added to generate an encoded block. In addition, the addition unit 107 inputs the generated encoded block to the intra-frame prediction predicted image generation unit 108 and the deblock filter 109.

  The intra-frame prediction predicted image generation unit 108 generates an intra-frame prediction prediction block used when encoding the encoding target block according to all prediction modes (9 types of prediction modes). Specifically, the intra-frame prediction predicted image generation unit 108 includes an encoded adjacent block located on the left side of the encoding target block, an encoded adjacent block located above, and an encoded position located on the upper left. An intra-frame prediction block is generated with reference to the adjacent block or the encoded adjacent block located in the upper right.

  The deblocking filter 109 removes block distortion that occurs in an encoded frame constituted by the encoded blocks input from the adder 107. Specifically, the deblocking filter 109 determines whether each encoded block is encoded by intra-frame prediction, and a pixel located at the boundary between the encoded block and the adjacent encoded adjacent block. Based on whether or not, the block boundary strength (Bs value; Boundary Strength) of each pixel is calculated, and block distortion is removed according to the calculated Bs value.

  The deblocking filter 109 also inputs the encoded block from which block distortion has been removed to the motion vector detection unit 103 and the inter-frame prediction predicted image generation unit 110.

  The inter-frame prediction predicted image generation unit 110 is a frame that is a prediction block used in inter-frame prediction based on the motion vector acquired from the motion vector detection unit 103 and the encoded frame acquired from the deblock filter 109. A prediction block for inter prediction is generated.

(Types of prediction modes according to the first embodiment of the present invention)
Hereinafter, types of prediction modes according to the first embodiment of the present invention will be described with reference to the drawings. FIG. 2 is a diagram illustrating types of prediction modes according to the first embodiment of the present invention. In the present embodiment, a case where the block size is 4 × 4 pixels will be described as an example. Each prediction mode is assigned a mode number (0 to 9), and the appearance frequency of a prediction mode with a small number of mode numbers is higher than the appearance frequency of a prediction mode with a large number of mode numbers.

  As shown in FIGS. 2A to 2I, there are nine types of prediction modes used when generating an intra-frame prediction block from an encoded block. Specifically, as shown in FIG. 2 (a), in prediction mode 0, adjacent blocks (four pixels at the bottom of the adjacent block) adjacent to the current block are referred to, and each pixel is vertically aligned. Prediction is performed to generate an intra-frame prediction block. As shown in FIG. 2 (b), in prediction mode 1, an adjacent block (4 pixels located at the right end of the adjacent block) adjacent to the left of the block to be encoded is referred to, and each pixel is predicted in the vertical direction. Thus, an intra-frame prediction block is generated. As shown in FIG. 2 (c), in prediction mode 2, an adjacent block (four pixels at the bottom of the adjacent block) adjacent on the encoding target block and an adjacent block (adjacent block of the adjacent block) adjacent to the left of the encoding target block are displayed. 4 pixels located at the right end) are referred to, and a prediction block for intra-frame prediction is generated based on the average value of each pixel.

  Similarly, as shown in FIGS. 2D to 2I, in prediction mode 3 to prediction mode 8, a prediction block for intra-frame prediction is generated using pixels of an adjacent block adjacent to the encoding target block. Is done.

(Prediction mode selection method according to the first embodiment of the present invention)
The prediction mode selection method according to the first embodiment of the present invention will be described below with reference to the drawings. FIG. 3 is a flowchart showing a prediction mode selection method according to the first embodiment of the present invention.

  As shown in FIG. 3, in step 110, the encoding apparatus 100 specifies the first prediction mode based on the prediction mode applied to the adjacent block adjacent to the encoding target block. For example, the encoding apparatus 100 performs the first prediction on the prediction mode with the smaller mode number among the prediction mode of the adjacent block adjacent to the left side of the encoding target block and the prediction mode of the adjacent block adjacent above the encoding target block. Specify as a mode.

  In step 120, the encoding apparatus 100 calculates prediction errors between the intra-frame prediction prediction blocks generated according to all the prediction modes (9 types of prediction modes) and the encoding target block.

  In step 130, the encoding apparatus 100 performs orthogonal transform (Hadamard transform) on the prediction error calculated in step 120 for each prediction mode.

  In step 140, the encoding apparatus 100 calculates the sum of absolute values (SATD) of the values obtained by orthogonal transform of the prediction error in step 130 for each prediction mode.

  In step 150, the encoding apparatus 100 specifies the prediction mode in which the SATD calculated in step 140 is minimum as the second prediction mode.

  In step 160, the encoding apparatus 100 encodes the encoding target block according to the first prediction mode specified in step 110, and encodes the encoding target block in accordance with the second prediction mode specified in step 150. Turn into.

  In step 170, the encoding apparatus 100 calculates the cost values of the first prediction mode and the second prediction mode by the rate-distortion optimization method (RDO method).

  In step 180, the encoding apparatus 100 selects a prediction mode to be used when encoding the current block based on the cost values of the first prediction mode and the second prediction mode. Specifically, the encoding apparatus 100 selects the prediction mode that minimizes the cost value calculated in step 170 as the prediction mode used when encoding the current block.

(Operation and effect of the encoding device according to the first embodiment of the present invention)
According to the encoding device 100 according to the first embodiment of the present invention, the encoding control unit 101 performs the first prediction mode specified by the prediction mode of the adjacent block and the second prediction mode in which the SATD is minimized. By applying a rate-distortion optimization method and selecting a prediction mode to be used when encoding a block to be encoded, the encoding apparatus 100 selects an optimal prediction mode in intra-frame prediction encoding. Can do. That is, the encoding apparatus 100 can select an optimal prediction mode even when the prediction error is not uniform.

[Second Embodiment]
(Prediction mode selection method according to the second embodiment of the present invention)
The prediction mode selection method according to the second embodiment of the present invention will be described below with reference to the drawings. In the following, differences from the above-described first embodiment will be mainly described.

  Specifically, in the first embodiment, the Rate-Distortion optimization method is applied to the first prediction mode specified by the prediction mode of the adjacent block and the second prediction mode that minimizes the SATD. In the second embodiment, in addition to the first prediction mode and the second prediction mode, Rate-Distration is optimal for the third prediction mode, which is a prediction mode in which the sum of absolute values of prediction errors (SAD) is minimized. Chemical method is applied.

  FIG. 4 is a flowchart showing a prediction mode selection method according to the second embodiment of the present invention. Steps 110 to 150 are the same as those in the first embodiment described above, and a description thereof will be omitted.

  As illustrated in FIG. 4, in step 122, the encoding apparatus 100 calculates the sum of absolute values (SAD) of the prediction errors calculated in step 120 for each prediction mode.

  In step 124, the encoding apparatus 100 specifies the prediction mode in which the SAD calculated in step 122 is minimized as the third prediction mode.

  In step 160a, the encoding apparatus 100 encodes the encoding target block according to the first prediction mode specified in step 110, and encodes the encoding target block according to the second prediction mode specified in step 150. Turn into. Furthermore, the encoding apparatus 100 encodes the encoding target block according to the third prediction mode specified in step 124.

  In step 170a, the encoding device 100 calculates the cost values of the first prediction mode, the second prediction mode, and the third prediction mode by the rate-distortion optimization method (RDO method).

  In step 180a, the encoding apparatus 100 selects a prediction mode to be used when encoding the current block based on the cost values of the first prediction mode, the second prediction mode, and the third prediction mode. Specifically, the encoding apparatus 100 selects the prediction mode that minimizes the cost value calculated in step 170 as the prediction mode used when encoding the current block.

[Comparison result]
Hereinafter, a comparison result between the prediction mode selection method according to the embodiment of the present invention and the prediction mode selection method according to the related art will be described with reference to the drawings. FIG. 5 is a diagram illustrating a comparison result between a prediction mode selection method according to an embodiment of the present invention and a prediction mode selection method according to the related art.

  In FIG. 5, (a) image quality priority method, (b) speed priority method, and (c) SAD selection method are listed as prediction modes according to the prior art. (A) The image quality priority method is a prediction mode selection method (High Complex Mode) in which image quality is prioritized. Specifically, in (a) image quality priority method, the Rate-Distortion optimization method is applied to all prediction modes.

  The (b) speed priority method is a prediction mode selection method (Low Complex Mode) in which processing speed is prioritized. Specifically, in (b) speed priority method, SAD (or SATD) and the bit amount for specifying the prediction mode are calculated for all prediction modes, and the cost value calculated based on these is calculated. A prediction mode is selected as a criterion.

  Further, (c) in the SAD selection method, among the prediction modes of adjacent blocks, rate-distortion optimization is performed for a prediction mode with the smallest mode number and a prediction mode with the smallest SAD calculated for all prediction modes. The law applies.

  As shown in FIG. 5, in (a) the image quality priority method, the encoding efficiency is high, but the processing time is long. In addition, in (b) the speed priority method, the processing time is short, but the encoding efficiency is low. Furthermore, in (c) SAD selection method, a certain degree of encoding efficiency was obtained while suppressing an increase in processing time.

  On the other hand, (d) the prediction mode selection method of the first embodiment and (e) the prediction mode selection method of the second embodiment suppress the increase in processing time, while (c) the SAD selection method. High coding efficiency was also obtained.

It is a block diagram which shows the structure of the encoding apparatus 100 which concerns on 1st Embodiment of this invention. It is a figure which shows the kind of prediction mode which concerns on 1st Embodiment of this invention. It is a flowchart which shows the selection method of the prediction mode which concerns on 1st Embodiment of this invention. It is a flowchart which shows the selection method of the prediction mode which concerns on 2nd Embodiment of this invention. It is a figure which shows the comparison result of this invention and a prior art.

Explanation of symbols

  DESCRIPTION OF SYMBOLS 100 ... Encoding apparatus, 101 ... Encoding control part, 102 ... Difference calculation part, 103 ... Motion vector detection part, 104 ... Orthogonal transformation / quantization part, 105 ... Entropy Encoding unit 106 ... Inverse quantization unit / inverse orthogonal transform 107 ... Adder unit 108 ... Intra-frame prediction predicted image generation unit 109 ... Deblock filter 110 ... Frame Prediction image generation unit for inter prediction, 111... Switch

Claims (3)

An encoding device that encodes an encoding target block, which is a block included in the encoding target image, using pixels included in the encoding target image to be encoded,
According to a plurality of types of prediction modes, a prediction block generation unit that generates a prediction block based on pixels included in the encoding target image;
Orthogonal transformation is performed on a prediction error that is a difference between the prediction block generated by the prediction block generation unit and the encoding target block, and an orthogonal transformation value sum that is a sum of absolute values of values obtained by the orthogonal transformation is obtained. A calculation unit for calculating for each prediction mode;
A first prediction mode specified by a prediction mode applied when an adjacent block that is a block adjacent to the encoding target block in the encoding target image is encoded, and the orthogonality calculated by the calculation unit A prediction mode selection unit that selects a prediction mode to be applied when the encoding target block is encoded by applying the Rate-Distortion optimization method to the second prediction mode that is the prediction mode in which the total sum of the transform values is the minimum. An encoding device comprising: The calculation unit calculates a prediction error sum that is a sum of absolute values of the prediction errors for each prediction mode,
In addition to the first prediction mode and the second prediction mode, the prediction mode selection unit performs rate-distortion optimization in a third prediction mode that is a prediction mode in which the prediction error sum calculated by the calculation unit is minimized. The encoding apparatus according to claim 1, wherein a prediction mode to be applied when the encoding target block is encoded is selected by applying an encoding method. A program applied to an encoding apparatus for encoding an encoding target block that is a block included in the encoding target image using pixels included in the encoding target image to be encoded, the computer In addition,
Generating a prediction block based on pixels included in the encoding target image according to a plurality of types of prediction modes; and
A prediction error that is a difference between the prediction block generated in the step A and the encoding target block is orthogonally transformed, and an orthogonal transformation value sum that is a sum of absolute values of values obtained by orthogonal transformation is a prediction mode. Step B to calculate for each,
A first prediction mode identified by a prediction mode applied when an adjacent block that is a block adjacent to the encoding target block in the encoding target image is encoded, and the orthogonality calculated in the step B Applying a rate-distortion optimization method to the second prediction mode, which is a prediction mode in which the sum of the converted values is minimum, and selecting a prediction mode to be applied when the encoding target block is encoded; A program characterized by being executed.

Images