JP2015180007A - Prediction mode selection method and prediction mode selection program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide a prediction mode selection method capable of further reducing a calculation cost by increasing the number of prediction modes requiring no processing.SOLUTION: A prediction mode selection method for video coding that omits cost calculation by previously obtaining prediction modes whose prediction residual costs are the same includes: a selection step of selecting at least one most probable prediction mode from prediction mode information on an adjacent block; a first prediction mode search step of acquiring a prediction mode number whose scan order does not agrees with that of the prediction mode number of the selected most probable prediction mode; a second prediction mode search step of acquiring a prediction mode number whose prediction image does not agrees with that of the prediction mode number of the selected most probable prediction mode; and a determination step of determining a prediction mode whose prediction processing is to be omitted on the basis of the prediction number whose scan order does not agree and the prediction number whose prediction image does not agree.

Description

  The present invention relates to a prediction mode selection method and a prediction mode selection program for selecting a prediction mode for video coding.

  H. In a video encoding standard such as H.264, an encoding target image is divided into small regions, a prediction mode and a prediction block size are determined for each divided region (small region), and a residual with the predicted image is orthogonally transformed to be orthogonal A method of encoding transform coefficients together with prediction mode and prediction block size information has become mainstream. The prediction modes are roughly classified into two types: intra prediction for generating a prediction image from in-screen information and inter prediction for generating a prediction image from another already encoded frame.

  As an example of Intra prediction, H. H.264 Intra4 × 4 prediction will be described. FIG. It is explanatory drawing which shows an example of H.264 intra4x4 prediction. H. In H.264, as shown in FIG. 6, an intra prediction image is created from 13 pixels of four encoded blocks A to D adjacent to the encoding target block. Intra4x4 prediction has a total of nine modes. FIG. 7 is an explanatory diagram showing nine types of prediction modes. As shown in FIG. 7, each predicted image is created in such a manner that adjacent pixels are copied in the direction indicated by the arrow. In Intra4 × 4 prediction, the mode with the highest coding efficiency is selected from these nine types of modes, and the mode number and the prediction residual signal (information on the difference between the predicted image and the original image) are encoded.

  Among these, in the encoding of mode numbers, the amount of codes is reduced by using the concept of “most probable mode” (hereinafter referred to as MPM). Of the mode numbers selected in adjacent encoded blocks A and B shown in FIG. 6, the smaller mode number is defined as MPM, and when the same mode number as MPM is selected, the header code amount is minimum. Become. FIG. It is explanatory drawing which shows an example of H.264 intra4x4 prediction. As shown in FIG. 8, when the prediction mode of the left adjacent block A is mode 5 and the prediction mode of the upper adjacent block B is mode 2, the MPM is mode 2. Therefore, in the encoding target block, when mode 2 is selected, the code amount of the mode number, that is, the header code amount R_H is the smallest.

In prediction mode and prediction block size selection by the RD optimization method, a method that is determined by using the following cost function is common.
Cost = D + λ × (R_T + R_H) (1)
Here, D is distortion, R_T is orthogonal transform coefficient code amount (DCT coefficient code amount), R_H is header code amount, and λ is a constant (Lagrange constant) determined from the quantization width and the prediction mode. Using the cost function of equation (1), calculating the cost for each prediction mode and prediction block size, and selecting a combination of the prediction mode and the prediction block size with the lowest cost, thereby performing video compression with high coding efficiency. Can do.

  FIG. 9 is a block diagram illustrating a configuration of an apparatus (Intra prediction mode determination unit) that performs Intra prediction mode selection using a cost function. This apparatus includes a header cost calculation unit 1, a control unit 2, an Intra prediction image generation unit 3, a prediction residual cost calculation unit 4, a cost calculation unit 5, and a minimum cost determination unit 6. When the prediction mode information of the adjacent block is input, the header cost calculation unit 1 calculates and outputs a header cost R_H based on the prediction mode information of the adjacent block. The control unit 2 sequentially outputs all intra prediction mode numbers from 0. The intra predicted image generation unit 3 generates and outputs a predicted image related to the input prediction mode number based on adjacent pixels.

  The prediction residual cost calculation unit 4 calculates and outputs the distortion D and the DCT coefficient code amount R_T from the difference between the Intra prediction image and the original image. The cost calculation unit 5 calculates λ from the prediction mode number and the quantization width, and uses this λ and the input header cost R_H, distortion D, and DCT coefficient code amount R_T according to the calculation of equation (1). Calculate the cost of prediction mode. The minimum cost determination unit 6 receives the prediction mode number and the prediction mode cost corresponding thereto for all prediction modes, obtains the prediction mode number having the smallest prediction mode cost, and outputs it together with the prediction mode cost.

  Next, the processing operation of the apparatus shown in FIG. 9 will be described with reference to FIG. FIG. 10 is a flowchart showing the processing operation of the apparatus shown in FIG. When the process starts, the next iterative process (steps S21 to S25) is performed for all intra prediction modes n. First, the header cost calculation unit 1 calculates the header code amount R_H of the target prediction mode n (step S21). Further, the Intra predicted image generation unit 3 calculates an Intra predicted image in the target prediction mode n (step S22). Subsequently, the prediction residual cost calculation unit 4 calculates the DCT coefficient code amount R_T and the distortion D from the difference signal from the original image (steps S23 and S24). And the cost calculation part 5 calculates the cost of object prediction mode n using (1) Formula (step S25). Next, the minimum cost determination unit 6 compares the costs of all intra prediction modes obtained by the above iterative process, and determines the mode with the lowest cost as the intra prediction mode (step S26).

  However, a large amount of calculation is required to calculate the cost for many prediction modes and prediction block sizes. In particular, with respect to the distortion D and the DCT coefficient code amount R_T, since it is necessary to perform orthogonal transformation on the prediction residual to obtain the code amount and distortion, the processing amount is very large. Therefore, as a method for reducing the processing amount without reducing the encoding efficiency as much as possible, a method using a technique such as Patent Document 1 is conceivable. This method is a technique for detecting a prediction mode in which prediction images match between different prediction modes. In the modes in which the predicted images match, the distortion D in equation (1) and the DCT coefficient code amount R_T match, so these values are calculated only once and the result is shared to reduce the amount of calculation. Can do.

  Incidentally, in the video encoding standard HEVC standardized in 2013, H.264 Compared to H.264, the options for the prediction block size, prediction mode, and orthogonal transform size have increased significantly. Focusing on intra prediction here, the prediction block size is 4x4, 8x8, and 16x16 plus 4 types in total of 32x32, the prediction mode is 9 types, 35 types, and the orthogonal transform sizes are 4x4 and 8x8 so far. What was two types has increased to a maximum of four types (4 × 4 to 32 × 32) that are smaller than the predicted block size. For the orthogonal transform sizes 4x4 and 8x8, the scan order of the prediction residual coefficient is applied in the vertical direction, the horizontal direction, or the diagonal direction depending on the prediction mode in order to improve the coding efficiency. Yes. For example, when the prediction mode is 6 to 14, vertical scanning is performed, and when 22 to 30, horizontal scanning is performed. Otherwise, oblique scanning is performed.

  When the acceleration by the above-described prior art method is applied to HEVC, the distortion D and the DCT coefficient code amount R_T match only when the predicted image and the scan order match in the 4x4 and 8x8 blocks. In addition, it is necessary to consider the matching of scan order.

  FIG. 11 is a block diagram illustrating a configuration of an apparatus (high-speed intra prediction mode determination unit) that uses a high-speed technique. In FIG. 11, the same parts as those of the apparatus shown in FIG. 9 are denoted by the same reference numerals, and the description thereof is omitted. The apparatus shown in FIG. 11 is different from the apparatus shown in FIG. 9 in that a prediction mode necessity list generation unit 21 and a control unit 22 with a prediction necessity function are provided in place of the control unit 2.

  The prediction mode necessity list generation unit 21 creates and outputs a prediction mode necessity list from prediction mode information of adjacent blocks and adjacent pixels. Upon receipt of the prediction mode necessity list, the control unit with a prediction necessity function 22 outputs only the prediction mode numbers that are “process required” among the prediction modes in order.

  Next, the processing operation of the apparatus shown in FIG. 11 will be described with reference to FIG. FIG. 12 is a flowchart showing the processing operation of the apparatus shown in FIG. In FIG. 12, the same reference numerals are given to the same portions as the processing operation shown in FIG. 10, and the description thereof is omitted. When the process starts, the prediction mode necessity list generation unit 21 determines a mode that requires processing and an unnecessary mode among all intra prediction modes, and creates and outputs a prediction mode necessity list (step S31). . Then, the control unit with a prediction necessity function 22 refers to the prediction mode necessity list, and only costs the intra prediction mode determined to be “processing necessary” according to the processing operation similar to the processing operation illustrated in FIG. The mode with the lowest prediction mode cost calculated is determined as the Intra prediction mode (steps S21 to S26).

  Next, the internal configuration of the prediction mode necessity list generation unit 21 shown in FIG. 11 will be described with reference to FIG. FIG. 13 is a block diagram showing an internal configuration of the prediction mode necessity list generation unit 21 shown in FIG. The prediction mode necessity list generation unit 21 includes a scan order match list output unit 211, a prediction image match determination unit 212, a two-set synthesis unit 213, a header cost calculation unit 214, a minimum header cost determination unit 215, and a prediction mode necessity list holding. Part 216.

  The scan order match list output unit 211 holds in advance a list of modes that match the scan order, and outputs the list as a scan order match mode list. The predicted image match determination unit 212 obtains a mode in which the predicted image matches from adjacent pixels, and creates and outputs a predicted image match mode list. The two-set combining unit 213 combines the two input lists and obtains and outputs a set of modes that are common to both lists. FIG. 15 is an explanatory diagram of an example of the matching mode list. FIG. 16 is an explanatory diagram illustrating an example of a set of predicted image matching and scan order matching modes. When the scan order match mode list and the predicted image match mode list shown in FIG. 15 are input, the two-set combining unit 213 obtains the pairs shown in the table of FIG. Output a set of numbers.

  When the prediction mode number is input, the header cost calculation unit 214 calculates and outputs a header code amount R_H for the input prediction mode number based on the prediction mode information of the adjacent block. The minimum header cost determination unit 215 inquires the header cost calculation unit 214 for the header cost of each prediction mode for the set of mode numbers sent from the two-set combining unit 213, and the prediction mode with the smallest header cost in this set. Print the number.

  The prediction mode necessity list holding unit 216 holds and outputs the necessity determination result of the prediction mode process for all prediction modes. At the start of processing, the necessity determination result of all the prediction modes held is “processing unnecessary”. When the prediction mode necessity list holding unit 216 receives the prediction mode number of the minimum header cost from the minimum header cost determination unit 215, the prediction mode necessity list holding unit 216 changes the held necessity prediction result of the target prediction mode to “processing required”.

  Next, the processing operation of the prediction mode necessity list generation unit 21 shown in FIG. 13 will be described with reference to FIG. FIG. 14 is a flowchart showing the processing operation of the prediction mode necessity list generation unit 21 shown in FIG. When the process starts, first, the prediction mode necessity list holding unit 216 sets all the modes in the prediction mode necessity list to “processing unnecessary” (step S41).

  Next, the scan order match list output unit 211 obtains a set of modes whose scan orders match (step S42). In parallel with this, the predicted image match determination unit 212 obtains a set of modes in which the predicted images match (step S43). FIG. 15 shows the matching mode list obtained by these processes. For simplicity, an example in which the prediction mode is 12 types of modes 1 to 12 is shown. The upper table shown in FIG. 15 is an example of the predicted image matching mode list, and shows results when the predicted images of modes 1 to 5 and the predicted images of modes 6 to 10 are equal. In this case, there are substantially only four types of predicted images of patterns 1 to 4. The lower table shown in FIG. 15 is an example of the scan order coincidence mode list. The scan orders of modes 1 to 3, 6, 7, and 11 are the same, and the scan orders of modes 4, 5, 8 to 10, and 12 are the same. An example of the case is shown.

  Next, the two-set synthesizing unit 213 obtains a set of prediction modes in which the scan order and the prediction image match based on this classification, and sets it as a set S (step S44). Next, the header cost calculation unit 214 calculates the header code amount R_H of each prediction mode belonging to the set S (step S45). And the minimum header cost determination part 215 calculates | requires the prediction mode with few header code amounts among the prediction modes which belong to these S. Then, the prediction mode necessity list holding unit 216 changes the item of the prediction mode necessity list to “processing necessary” for the prediction mode with the smallest header code amount (step S46).

  FIG. 17 is a diagram illustrating an example of a result of the prediction mode necessity list according to this process. If the example shown in FIG. 15 is divided into a set in which the scan order and the predicted image match, it is classified into six sets shown in the upper table shown in FIG. Among these, the prediction mode with the smallest header code amount in each group is circled. In the same set of prediction modes, the distortion D and the DCT coefficient code amount R_T in the equation (1) are equal, so the size of the header code amount R_H is directly the cost. Therefore, it can be seen that the cost of the mode with a circle is the smallest in each group, and the cost of the other prediction modes does not need to be calculated. A list of the results is a prediction mode necessity list below. At the time of encoding, the amount of calculation can be suppressed while maintaining the encoding efficiency by performing the cost calculation only for the modes marked “necessary” in this table.

  Next, another example of the speed-up method will be described. FIG. 18 is a block diagram illustrating a configuration of an apparatus (high-speed intra prediction mode determination unit) that realizes another high-speed technique. In FIG. 18, the same parts as those in the apparatus shown in FIG. The apparatus shown in FIG. 18 differs from the apparatus shown in FIG. 11 in that a control unit 2 and a necessity determination unit 23 are provided instead of the control unit 22 with a prediction necessity function.

  The control unit 2 outputs a prediction mode number. The necessity determination unit 23 determines the necessity of the prediction mode number sent from the control unit 2 based on the prediction mode necessity list sent from the prediction mode necessity list generation unit 21. The necessity determination unit 23 outputs the prediction mode number determined as “processing required”, but does not output the prediction mode number determined as “processing unnecessary”.

  Next, the processing operation of the apparatus shown in FIG. 18 will be described with reference to FIG. FIG. 19 is a flowchart showing the processing operation of the apparatus shown in FIG. In FIG. 19, the same parts as those in the processing operation shown in FIG. When the process starts, the prediction mode necessity list generation unit 21 determines a mode that requires processing and an unnecessary mode among all intra prediction modes, and creates and outputs a prediction mode necessity list (step S31). . Then, iterative processing of step S32 and steps S21 to S25 is performed. The necessity determination unit 23 determines whether or not the processing of the target prediction mode is necessary based on the necessity result (step S32). As a result of this determination, if it is determined that it is necessary, the processes of steps S21 to S25 are executed. On the other hand, if it is determined that the process is unnecessary, the process of steps S21 to S25 is not performed, and the process proceeds to the next prediction mode iterative process. By this processing operation, the processing speed can be increased.

Japanese Patent No. 479709

  As described above, in HEVC, processing can be reduced only in a mode in which both the predicted image and the scan order match. For this reason, in the prediction mode selection method described in Patent Document 1, there is a problem that processing that can be reduced is limited, and calculation cost cannot be reduced.

  The present invention has been made in view of such circumstances, and provides a prediction mode selection method and a prediction mode selection program that can further reduce the calculation cost by increasing the number of prediction modes that do not require processing. Objective.

  The present invention is a video encoding prediction mode selection method in which prediction modes in which prediction residual costs are equal are obtained in advance and cost calculation is omitted, and at least one of the most powerful prediction modes is selected from prediction mode information of adjacent blocks. A selection step for selecting, a first prediction mode search step for obtaining a prediction mode number in which the scan order does not match the prediction mode number of the selected most powerful prediction mode, and the prediction mode number and prediction of the selected most powerful prediction mode A second prediction mode search step for obtaining a prediction mode number in which images do not match, a prediction mode in which prediction processing should be omitted based on the prediction mode number in which the scan order does not match and the prediction mode number in which the prediction images do not match And a determination step for determining the above.

  The present invention relates to a prediction mode selection method for video coding in which a prediction mode in which prediction residual costs are equal is obtained in advance and cost calculation is omitted, and the prediction mode is grouped into a plurality of groups in advance. A selection step of selecting at least one of the most probable prediction modes from the prediction mode information of adjacent blocks; and a step of obtaining a prediction mode number in which the selected prediction mode number of the most probable prediction mode does not match the group to which the group belongs A first prediction mode search step, a second prediction mode search step for obtaining a prediction mode number in which the prediction mode number of the selected most probable prediction mode and the prediction image do not match, and the prediction mode number in which the group does not match, Prediction mode in which prediction processing should be omitted based on the prediction mode number where the prediction images do not match And having a determining step.

  The present invention is characterized in that, in the determination step, the prediction mode in which the scan order and the predicted image do not coincide with each other, or the prediction mode in which the belonging group and the predicted image do not coincide with each other is omitted is determined as a prediction mode.

  The present invention is characterized in that, in the determination step, a prediction mode in which the scan order does not match is determined as a prediction mode that omits the prediction mode.

  The present invention is characterized in that, in the determining step, a prediction mode in which the prediction image or the group to which the group belongs does not match is determined as a prediction mode.

  The present invention is characterized in that, in the selection step, the most probable prediction mode is selected by selecting all of the plurality of prediction modes of the adjacent block or by selecting one prediction mode that appears most frequently. .

  The present invention is characterized in that, in the selection step, the most probable prediction mode is selected by obtaining one or all of the most probable modes obtained from the adjacent blocks.

  The present invention is a prediction mode selection program for causing a computer to execute the prediction mode selection method.

  According to the present invention, the number of prediction modes determined as “processing unnecessary” can be increased as compared with the prior art, so that it is possible to further reduce the calculation cost.

It is a block diagram which shows the apparatus structure of 1st Embodiment of this invention. It is explanatory drawing which shows an example of a process result. It is a flowchart which shows the processing operation of the apparatus shown in FIG. It is a block diagram which shows the apparatus structure of 2nd Embodiment of this invention. 5 is a flowchart showing a processing operation of the apparatus shown in FIG. H. It is explanatory drawing which shows an example of H.264 intra4x4 prediction. It is explanatory drawing which shows nine types of prediction modes. H. It is explanatory drawing which shows an example of H.264 intra4x4 prediction. It is a block diagram which shows the structure of the apparatus (Intra prediction mode determination part) which performs Intra prediction mode selection by a cost function. 10 is a flowchart showing a processing operation of the apparatus shown in FIG. 9. It is a block diagram which shows the structure of the apparatus (high-speed intra prediction mode determination part) using the speed-up method. It is a flowchart which shows the processing operation of the apparatus shown in FIG. It is a block diagram which shows the internal structure of the prediction mode necessity list production | generation part 21 shown in FIG. It is a flowchart which shows the processing operation of the prediction mode necessity list production | generation part 21 shown in FIG. It is explanatory drawing which shows an example of a matching mode list | wrist. It is explanatory drawing which shows an example of the group of a prediction image coincidence and scan order coincidence mode. It is a figure which shows an example of the result of a prediction mode necessity list. It is a block diagram which shows the structure of the apparatus (high-speed intra prediction mode determination part) which implement | achieves other speed-up methods. It is a flowchart which shows the processing operation of the apparatus shown in FIG.

  Hereinafter, a prediction mode selection method according to an embodiment of the present invention will be described. First, the principle of the prediction mode selection method of the present invention will be described. The prediction mode selection method of the present invention has a function for obtaining one or a plurality of prediction mode numbers from prediction mode information of adjacent blocks, and a function for obtaining a mode number whose scan order does not coincide with the obtained one or more prediction mode numbers. And a function for obtaining a mode number in which one or a plurality of obtained prediction mode numbers do not match the predicted image. Then, a function for determining a prediction mode for “processing unnecessary” based on a mode number that does not match the scan order and a mode number that does not match the prediction image is newly added.

  A specific example of the function for obtaining one or a plurality of mode numbers from the prediction mode information of the adjacent block is, for example, a function for obtaining any one or all of the prediction modes of the adjacent blocks. As a method of narrowing down any one, for example, the most frequent mode or the mode with the smallest mode number can be cited. Another specific example is a function for obtaining one or all of MPMs obtained from adjacent blocks. Moreover, as a method of narrowing down any one, the mode with the smallest mode number is mentioned, for example.

  As a specific example of the function for determining the prediction mode to be “processing unnecessary” based on the mode number in which the scan order does not match and the mode number in which the prediction image does not match, for example, a prediction mode in which neither the scan order nor the prediction image matches It is determined that the prediction mode is “processing unnecessary”. Alternatively, only a mode in which the scan order does not match is determined as a prediction mode that does not require processing. Alternatively, only the prediction mode in which the prediction images do not match is determined as the prediction mode “processing unnecessary”. Alternatively, in the conventional method, the determination is based on the match between the scan order and the predicted image, but instead of the classification based on the scan order, the prediction direction is grouped, and whether or not the prediction direction groups match is used as a determination criterion. .

<First Embodiment>
Next, a prediction mode selection method according to the first embodiment of the present invention will be described. FIG. 1 is a block diagram showing a device configuration of the embodiment. In this figure, the same parts as those of the conventional apparatus shown in FIG. The apparatus shown in this figure is different from the conventional apparatus in that a most powerful mode number determination unit 217, two mismatch mode collation units 218 and 219, an unnecessary mode addition determination unit 220, and a prediction mode necessity list holding / updating unit 221 are provided. This is the point.

  The most probable mode number determination unit 217 obtains the MPM from the prediction mode information of the adjacent block, and outputs it as the most probable mode number. The inconsistent mode collating unit 218 receives the most probable mode number and the scan order coincidence mode list as inputs, and outputs prediction mode numbers that are classified into categories different from the most probable mode number in the list. When the scan order match mode list is input, the mismatch mode check unit 218 outputs a prediction mode number whose scan order is different from the most probable mode number. The mismatch mode collating unit 219 receives the most probable mode number and the predicted image match mode list as inputs, and outputs the prediction mode numbers that are classified in a different category from the most probable mode number in the list. When the predicted image match mode list is input, the mismatch mode check unit 219 outputs a prediction mode number in which the most probable mode number and the predicted image are different.

  The unnecessary mode addition determination unit 220 obtains a list of unnecessary prediction mode numbers by obtaining a logical product of the two input prediction mode lists. The prediction mode necessity list holding / updating unit 221 holds and outputs the necessity determination result of the prediction mode process for all prediction modes. At the start of processing, the necessity determination result of all the prediction modes held is “processing unnecessary”. When the prediction mode number of the minimum header cost is received from the minimum header cost determination unit 215, the necessity determination result of the prediction mode held is changed to “processing required”. Further, when the mode number is received from the unnecessary mode addition determination unit 220, the necessity determination result of the held prediction mode is changed to “processing unnecessary”. When the mode number input from the minimum header cost determination unit 215 and the same mode number are input from the unnecessary mode addition determination unit 220, the result of the unnecessary mode addition determination unit 220 is prioritized and “processing unnecessary” is set. change.

  In the prediction mode necessity list generation unit 21 shown in FIG. 1, all MPM scan orders are used as a function for obtaining one or a plurality of mode numbers from prediction mode information of adjacent blocks. In addition, as a function of determining a prediction mode that “does not require processing” based on a mode number that does not match the scanning order and a mode number that does not match the prediction image, a prediction mode that does not match both the scanning order and the prediction image is “processing unnecessary”. It is assumed that the prediction mode is correct.

  For example, the prediction mode necessity list generation unit 21 shown in FIG. 1 receives the set of the predicted image match mode list and the scan order match mode list shown in FIG. Shown in FIG. 2 is an explanatory diagram illustrating an example of a processing result. Of the prediction modes that have been determined to be “processing required” in the prior art, mode 1 and mode 11 are determined to be “processing unnecessary” because both the scan order and the predicted image are different from MPM.

  Next, the processing operation of the apparatus shown in FIG. 1 will be described with reference to FIG. FIG. 3 is a flowchart showing the processing operation of the apparatus shown in FIG. When the process starts, the prediction mode necessity list holding / updating unit 221 first sets all the modes in the prediction mode necessity list to “processing unnecessary” (step S1).

  Next, the scan order match list output unit 211 obtains a set of modes whose scan orders match (step S2). In parallel with this, the predicted image match determination unit 212 obtains a set of modes in which the predicted images match (step S3).

  Next, the two-set combining unit 213 obtains a set of prediction modes in which the scan order and the prediction image match based on this classification, and sets it as a set S (step S4). Next, the header cost calculation unit 214 calculates the header code amount R_H of each prediction mode belonging to the set S (step S5).

  Next, the minimum header cost determination unit 215 obtains the prediction mode with the smallest header code amount among the prediction modes belonging to S. Then, the prediction mode necessity list holding unit 216 changes the item of the prediction mode necessity list to “processing necessary” for the prediction mode with the smallest header code amount (step S6).

  Next, the inconsistent mode collating unit 218 inputs the most probable mode number and the scan order coincidence mode list, and outputs a prediction mode number having a different scan order from the most probable mode number. Further, the mismatch mode matching unit 219 receives the most probable mode number and the predicted image match mode list, and outputs a prediction mode number in which the most probable mode number and the predicted image are different.

  Next, the unnecessary mode addition determination unit 220 calculates and outputs a list of unnecessary prediction mode numbers by calculating the logical product of the two input prediction mode lists. In response to this, the prediction mode necessity list holding / updating unit 221 changes the item of the prediction mode different from the MPM in the scan order and the prediction image in the prediction mode necessity list to “processing unnecessary” (step S7).

  As described above, since the number of prediction modes determined as “processing unnecessary” is increased as compared with the conventional technique, the calculation cost can be further reduced.

Second Embodiment
Next, a prediction mode selection method according to the second embodiment of the present invention will be described. FIG. 4 is a block diagram showing a device configuration of the embodiment. In this figure, the same parts as those in the apparatus shown in FIG. The apparatus shown in this figure is different from the apparatus shown in FIG. 1 in that a prediction direction group match list output unit 222 is provided instead of the scan order match list output unit 211. The second embodiment uses a prediction direction group match list in which prediction modes are grouped in advance according to the prediction direction, instead of using the scan order match mode list in which the modes with the same scan order are listed in the first embodiment. Different.

  The prediction direction group match list output unit 222 outputs a prediction mode group match list in which prediction modes are grouped in advance according to the prediction direction.

  Next, the processing operation of the apparatus shown in FIG. 4 will be described with reference to FIG. FIG. 5 is a flowchart showing the processing operation of the apparatus shown in FIG. When the process starts, first, the prediction mode necessity list holding / updating unit 221 sets all the modes in the prediction mode necessity list to “processing unnecessary” (step S11).

  Next, the prediction direction group match list output unit 222 performs a match determination on the prediction direction group and outputs a prediction direction group match mode list (step S12). In parallel with this, the predicted image match determination unit 212 obtains a set of modes in which the predicted images match (step S13).

  Next, the two-set combining unit 213 obtains a set of prediction modes in which the prediction direction group and the prediction image match based on this classification, and sets it as a set S (step S14). Next, the header cost calculation unit 214 calculates the header code amount R_H of each prediction mode belonging to the set S (step S15).

  Next, the minimum header cost determination unit 215 obtains the prediction mode with the smallest header code amount among the prediction modes belonging to S. Then, the prediction mode necessity list holding unit 216 changes the item of the prediction mode necessity list to “processing necessary” for the prediction mode with the smallest header code amount (step S16).

  Next, the inconsistent mode collating unit 218 inputs the most probable mode number and the scan order coincidence mode list, and outputs a prediction mode number having a different scan order from the most probable mode number. Further, the mismatch mode matching unit 219 receives the most probable mode number and the predicted image match mode list, and outputs a prediction mode number in which the most probable mode number and the predicted image are different.

  Next, the unnecessary mode addition determination unit 220 calculates and outputs a list of unnecessary prediction mode numbers by calculating the logical product of the two input prediction mode lists. Receiving this, the prediction mode necessity list holding / updating unit 221 changes the item of the prediction mode different from the MPM in the scan order and the prediction image in the prediction mode necessity list to “processing unnecessary” (step S17).

  As described above, since the number of prediction modes determined as “processing unnecessary” is increased as compared with the prior art, the calculation cost can be further reduced. In general, since the scan order has a high correlation between adjacent blocks, a prediction mode having a scan order with a low correlation with an adjacent block in a set of prediction modes with the same predicted image is set as “no processing required”. However, the degradation of the encoding efficiency can be extremely reduced.

  You may make it implement | achieve the prediction mode necessity list production | generation part 21 in embodiment mentioned above with a computer. In that case, a program for realizing this function may be recorded on a computer-readable recording medium, and the program recorded on this recording medium may be read into a computer system and executed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Furthermore, the “computer-readable recording medium” dynamically holds a program for a short time like a communication line when transmitting a program via a network such as the Internet or a communication line such as a telephone line. In this case, a volatile memory inside a computer system serving as a server or a client in that case may be included and a program held for a certain period of time. Further, the program may be for realizing a part of the functions described above, and may be a program capable of realizing the functions described above in combination with a program already recorded in the computer system. It may be realized using hardware such as PLD (Programmable Logic Device) or FPGA (Field Programmable Gate Array).

  As mentioned above, although embodiment of this invention has been described with reference to drawings, the said embodiment is only the illustration of this invention, and it is clear that this invention is not limited to the said embodiment. is there. Therefore, additions, omissions, substitutions, and other modifications of the components may be made without departing from the technical idea and scope of the present invention.

  By increasing the number of prediction modes that do not require processing, the present invention can be applied to applications where it is indispensable to further reduce the calculation cost.

  21 ... Prediction mode necessity list generation unit, 211 ... Scan order match list output unit, 212 ... Predictive image match determination unit, 213 ... Two-set synthesis unit, 214 ... Header cost calculation unit 215: Minimum header cost determination unit, 217: Most probable mode number determination unit, 218, 219 ... Mismatch mode verification unit, 220 ... Unnecessary mode addition determination unit, 221 ... Prediction mode required Reject list holding / updating unit, 222...

The present invention relates to a prediction mode selection method for selecting a prediction mode when the movies picture coding, a selection step of selecting at least one of the most promising prediction mode from the prediction mode information of adjacent blocks, the selected top A first prediction mode search step for obtaining a prediction mode number of a prediction mode in which the scan order of the leading prediction mode and the prediction residual coefficient does not match, and a prediction mode in which the prediction mode number of the selected leading prediction mode and the prediction image do not match A second prediction mode search step for obtaining a number; a determination step for determining a prediction mode in which a prediction process should be omitted based on the prediction mode number in which the scan order does not match and the prediction mode number in which the prediction images do not match; It is characterized by having.

The present invention relates to a prediction mode selection method for selecting a prediction mode when the movies picture encoding, and grouping steps to be grouped in advance a plurality of groups prediction mode, most of the prediction mode information of adjacent blocks A selection step for selecting at least one dominant prediction mode, a first prediction mode search step for obtaining a prediction mode number for which the prediction mode number of the selected most powerful prediction mode and the group to which the group belongs do not match, The second prediction mode search step for obtaining a prediction mode number in which the prediction mode number of the most probable prediction mode and the prediction image do not match, the prediction mode number in which the group does not match, and the prediction mode number in which the prediction image does not match And a determination step for determining a prediction mode in which the prediction process should be omitted, That.

The present invention, in the above determination step is characterized by determining the prediction mode in which the predictive image or the membership group does not match the prediction mode is omitted.

Claims (8)

A prediction mode selection method for video coding that eliminates cost calculation by obtaining a prediction mode in which prediction residual costs are equal in advance,
A selection step of selecting at least one most probable prediction mode from prediction mode information of adjacent blocks;
A first prediction mode search step for obtaining a prediction mode number whose scan order does not match the prediction mode number of the selected most probable prediction mode;
A second prediction mode search step for obtaining a prediction mode number in which the prediction mode number of the selected most probable prediction mode and the prediction image do not match;
A prediction mode selection method comprising: a determination step of determining a prediction mode in which a prediction process should be omitted based on the prediction mode number in which the scan order does not match and the prediction mode number in which the prediction images do not match. A prediction mode selection method for video coding that eliminates cost calculation by obtaining a prediction mode in which prediction residual costs are equal in advance,
A grouping step for previously grouping the prediction modes into a plurality of groups;
A selection step of selecting at least one most probable prediction mode from prediction mode information of adjacent blocks;
A first prediction mode search step for obtaining a prediction mode number in which the selected prediction mode number of the most probable prediction mode and the group to which the group belongs do not match;
A second prediction mode search step for obtaining a prediction mode number in which the prediction mode number of the selected most probable prediction mode and the prediction image do not match;
A prediction mode selection method comprising: a determination step of determining a prediction mode in which a prediction process should be omitted based on the prediction mode number in which the group does not match and the prediction mode number in which the prediction image does not match.   2. The determination step is characterized in that the prediction mode in which the scan order and the predicted image do not match each other or the group belonging to the predicted image and the predicted image do not match is determined as a prediction mode that omits the prediction mode. Or the prediction mode selection method of 2.   The prediction mode selection method according to claim 1, wherein in the determination step, a prediction mode in which the scan order does not match is determined as a prediction mode that omits the prediction mode.   The prediction mode selection method according to claim 1 or 2, wherein in the determination step, a prediction mode in which the prediction image or the group to which the group belongs does not match is determined as a prediction mode that omits the prediction mode.   The selection step includes selecting the most probable prediction mode by selecting all of the plurality of prediction modes of the adjacent block or by selecting one prediction mode that appears most frequently. 6. The prediction mode selection method according to any one of 5 above.   The prediction according to any one of claims 1 to 5, wherein, in the selection step, the most probable prediction mode is selected by obtaining one or all of most probable modes obtained from the adjacent blocks. Mode selection method.   A prediction mode selection program for causing a computer to execute the prediction mode selection method according to any one of claims 1 to 7.

Images