KR101337345B1 - Encoder and method for skipping sub-pel motion estimation - Google Patents

Abstract

An encoder and a subpel motion estimation skipping method are disclosed. The encoder according to the present invention includes: a determining unit that determines whether a size of a current prediction unit matches a size of a previous prediction unit existing at a same position in a previous frame in which an encoding mode is determined; The subpel search unit may perform a motion estimation of a sub-pel unit only when the size of the current prediction unit coincides with the size of the previous prediction unit. According to the present invention, there is an effect that the video encoding time can be shortened in all devices using the encoding device.

Description

Encoder and method for skipping sub-pel motion estimation

The present invention relates to an encoder and a method for omitting subpel motion estimation.

Due to the development of various multimedia devices and the generalization of digital multimedia broadcasting service, the demand for high quality video service is increasing. To this end, the development and dissemination of devices capable of playing and storing high-definition video content has been accelerated, and the development of video codecs for effectively encoding or decoding high-definition video content is underway.

In the case of the video codec according to the related art, a limited encoding method is applied based on a macroblock having a predetermined size in encoding video data.

In order to overcome these limitations and enable more efficient encoding, MPEG-H High Efficiency Video Coding (HEVC), a next-generation video encoding technology, is a coding unit (CU) that is a basic unit of prediction and transformation. ) Is used for example, from 64x64 to 8x8, and also has three block concepts: Coding Unit (CU), Prediction Unit (PU), and Transform Unit (TU). A hierarchical video compression scheme is employed. This concept is also present in some of the articles “Improved Video Compression Efficiency Through Flexible Unit Representation and Corresponding Extension of Coding Tools” (W.-J. Han, J. Min, I.-K. Kim, E. Alshina, A.) Alshin, T. Lee, J. Chen, V. Seregin, S. Lee, YM Hong, M.-S. Cheon, N. Shlyakhov, K. McCann, T. Davies, J.-H. Park, “Improved Video Compression Efficiency Through Flexible Unit Representation and Corresponding Extension of Coding Tools, ”IEEE Trans. Circuits Syst. Video Technol., Vol. 20, no. 12, pp. 1709-1720, Dec. 2010).

FIG. 1A is a diagram illustrating a coding unit CU and a prediction unit PU according to depths presented in MPEG-H HEVC, and FIG. 1B is a diagram illustrating a relationship between a coding unit and a transformation unit (TU). 2 is a diagram illustrating a process for determining the size of a coding unit for each data unit according to the prior art, and FIG. 3 is a diagram illustrating an example in which the size of a coding unit is determined for each data unit.

Here, the depth means a step in which the coding unit CU is hierarchically divided, and as the depth increases, the coding unit according to depths is determined from the maximum coding unit (for example, 64x64) to the minimum coding unit (eg, 64x64). For example, up to 8x8). The depth of depth may be expressed in a direction from an upper depth (eg, depth 0) to a lower depth (eg, depth 3). As the depth becomes deeper, the number of times of division of the maximum coding unit increases, and the total number of times of division of the maximum coding unit corresponds to the maximum depth. The maximum size and the maximum depth of the coding unit may be preset.

FIG. 1A illustrates a case where the maximum height and width of the coding unit are 64 and the maximum depth is 3 as a hierarchical structure of the coding unit. As shown, the size of the largest coding unit 110 at the highest depth (ie, depth 0) may be represented as 64 × 64.

The depth is deeper along the vertical axis of the hierarchical structure of the coding unit, and thus the height and width of the coding unit according to depths are respectively divided. Therefore, 64x64, which is the size of the largest coding unit at depth 0, which is the highest depth, is divided 120 into the size of the coding unit having a depth of 32x32 at depth 1, and is 130 divided into the size of the coding unit having a 16x16 at depth 2, at the lowest level. A depth 140 is divided into a size of a coding unit having a depth of 8 × 8 at a depth of 3. Here, an 8x8 sized coding unit of depth 3 may be referred to as a minimum coding unit.

Further, a prediction unit PU, which is a partial data unit that is the basis of the prediction encoding of each depth-decoding coding unit, is shown along the horizontal axis of the hierarchical structure of the coding unit. That is, the prediction unit of the coding unit 110 having a size of 64x64 having a depth of 0 includes a 64x64 sized partial data unit, a 64x32 sized partial data unit, a 32x64 sized partial data unit, and a 32x32 sized data included in a 64x64 sized coding unit. Partial data units, and so on. Therefore, the coding unit may be represented by a data unit that is a square of a minimum size including each partial data unit.

Similarly, the prediction unit of a 32x32 sized coding unit of depth 1 may be a 32x32 partial data unit, a 32x16 partial data unit, a 16x32 partial data unit, a 16x16 partial data unit, or the like, and a depth 2 16x16 size. The prediction unit of the coding unit of may be a 16x16 partial data unit, a 16x8 partial data unit, an 8x16 partial data unit, an 8x8 partial data unit, or the like.

1B shows a relationship between a coding unit and a transform unit.

The encoding apparatus encodes an image with a coding unit having a size smaller than or equal to the maximum coding unit for each maximum coding unit. In the encoding process, the size of a transform unit (TU) for frequency transformation may be selected based on a data unit that is not larger than each coding unit.

For example, assuming that the current coding unit is 64x64 size, frequency transformation may be performed using a 32x32 size transform unit. In addition, after encoding the data of the 64x64 size coding unit into 32x32, 16x16, 8x8, and 4x4 size transform units each having 64x64 size or less, the transform unit having the least error with the original may be selected. .

The encoding apparatus considering the depth-based encoding unit illustrated in FIG. 1A should perform encoding for each coding unit of each depth included in the maximum coding unit 110 to determine the coded depth for the maximum coding unit 110. As a result, as shown in FIG. 2, the size of the coding unit for each data unit may be determined.

At this time, the number of coding units according to depths for including data of the same range and size increases as the depth increases. For example, four data units of depth 2 are required for data included in one coding unit of depth 1. Therefore, in order to compare the encoding results of the same data for each depth, each of the coding units having one depth 1 and four coding units having four depths 2 should be encoded.

That is, as shown in FIG. 2, in the MPEG-H HEVC, both the prediction mode and the transform mode at each depth are applied to obtain the best coding gain. For each data unit (that is, from a 64x64 data unit to an 8x8 data unit), the size of an optimal coding unit CU, a prediction unit PU, and a transform unit TU are determined. A coding unit for each data unit that is optimally determined by this is illustrated in FIG. 3.

However, according to the method of determining the size of the coding unit (CU) in the manner proposed in the prior art, although the optimal coding unit size can be determined by considering various coding unit sizes, the complexity increases in the determination process. Problem occurs. For example, as shown in FIG. 2, when the lowest depth is 3 and the size of the largest coding unit is 64x64, the encoding apparatus determines at least 1xCU0 to determine the size of the coding unit for each data unit by determining an optimal cost. The processing corresponding to the quantity of + 4xCU1 + 16xCU2 + 64xCU3 should be performed.

In addition, in order to increase the performance of the motion estimation, the motion estimation is performed not only in 1/2 pixel but also in 1/4 pixel or 1/8 pixel units. This is because the actual motion is not made by the unit of integer pixel, so the motion vector of the unit of integral multiple is inferior in accuracy.

In order to perform motion estimation in units of sub-pels (eg, sub-pels) such as only 1/2 pixel, 1/4 pixel, and 1/8 pixel, a two-step process is generally performed. In the first step, the motion estimation is performed in integer units in the given search region to obtain an integer motion vector. In the second step, the motion vector is predicted in units of subpels based on the integer motion vector to obtain a final motion vector.

In addition to searching for more search points when estimating subpel motion, a large amount of computation is required in the interpolation process to produce 1/2, 1/4, 1/8 pixels, and the like.

As such, the demand for a large amount of computation has problems such as weakening of the encoder's real-time processing performance. Therefore, it is necessary to develop an encoder with higher efficiency by minimizing degradation of image quality and omitting motion estimation in subpel units. .

An object of the present invention is to provide an encoder and a method for eliminating subpel motion estimation, which can reduce encoding time for video data in all devices using an encoder.

The present invention determines encoding mode by referring to a depth of a coding unit (CU) of a previous frame, a partition size and / or an index of a prediction unit (PU) in a motion estimation (ME) process. The present invention provides an encoder and a method of eliminating subpel motion estimation, which enable a faster encoding process by allowing motion estimation of a subpel unit for a prediction unit of a current frame to be omitted.

An object of the present invention is to provide an encoder and a method of skipping subpel motion estimation, which can perform a faster encoding process by interpolation and a subpel unit search point, which are required in the process of making a subpel.

Other objects of the present invention will become readily apparent from the following description.

According to an aspect of the present invention, an encoder for performing a subpel motion estimation skipping method is provided.

According to an embodiment of the present invention, an encoder includes: a determining unit that determines whether a size of a current prediction unit matches a size of a previous prediction unit existing at a same position in a previous frame in which an encoding mode is determined; An encoder is provided that includes a subpel search unit that performs motion estimation on a sub-pel basis only when the size of the current prediction unit matches the size of the previous prediction unit.

An index may be used for the current prediction unit to identify the position in the current frame and the current coding unit, and the previous prediction unit to the position in the previous frame and the previous coding unit, respectively.

The determination unit may match a depth of a current coding unit including the current prediction unit and a depth of a previous coding unit including the previous prediction unit, and a partition size of the current prediction unit and a partition size of the previous prediction unit match. It may be determined whether the size of the current prediction unit coincides with the size of the previous prediction unit with reference to the information.

The encoder may further include a previous frame information storage unit that stores information about a depth of the previous coding unit and a partition size of the previous prediction unit included in the previous coding unit.

The encoder may further include an integer pixel search unit that obtains an integer multiple motion vector by performing motion estimation in units of multiples in a designated search region. Here, the integer motion vector or the motion vector of the subpel unit may be output as the final motion vector depending on whether the subpel motion estimation is performed.

The encoder may further include a cost calculator configured to calculate a cost value for each prediction mode of the motion estimation prediction unit using the final motion vector.

The prediction mode may be at least one of Inter 2Nx2N, Inter 2NxN, Inter Nx2N, and Inter NxN.

The cost value may be calculated using one or more of a sum of absorptive difference (SAD), a sum of squared difference (SSD), and a sum of absolute transformed difference (SATD).

According to another aspect of the present invention, a subpel motion estimation skipping method provided by an encoder is provided.

According to an embodiment of the present invention, in the subpel motion estimation omitting method, (a) the determining unit whether the size of the current prediction unit matches the size of the previous prediction unit existing at the same position in the previous frame in which the encoding mode is determined. Determining whether or not; And (b) the sub-pel searcher performing motion estimation of the subpel unit only when the partition size of the current prediction unit matches the partition size of the previous prediction unit. A method is provided.

The determination unit may match a depth of a current coding unit including the current prediction unit and a depth of a previous coding unit including the previous prediction unit, and a partition size of the current prediction unit and a partition size of the previous prediction unit match. It may be determined whether the size of the current prediction unit coincides with the size of the previous prediction unit with reference to the information.

The step of storing information on the depth of the previous coding unit and the partition size of the previous prediction unit corresponding to the previous coding unit may be stored in a previous frame information storage unit before step (a).

An index may be used for the current prediction unit to identify the position in the current frame and the current coding unit, and the previous prediction unit to the position in the previous frame and the previous coding unit, respectively.

Before the step (b), a step of acquiring an integer multiple motion vector may be performed by performing motion estimation in units of integral multiples in a designated search area that performs motion estimation in the subpel units, and whether or not the subpel motion estimation is performed. As a result, the integer motion vector or the motion vector of the subpel unit may be output as the final motion vector.

A cost value for each prediction mode of the motion estimation prediction unit may be calculated using the final motion vector.

The prediction mode may be at least one of Inter 2Nx2N, Inter 2NxN, Inter Nx2N, and Inter NxN.

The cost value may be calculated using one or more of a sum of absorptive difference (SAD), a sum of squared difference (SSD), and a sum of absolute transformed difference (SATD).

Other aspects, features, and advantages will become apparent from the following drawings, claims, and detailed description of the invention.

According to an embodiment of the present invention, all devices using the encoder can reduce the encoding time for video data.

In the motion estimation process, the current mode for determining the encoding mode may be determined by referring to the depth of the coding unit CU of the previous frame, the partition size and / or the index of the prediction unit PU. By allowing the motion estimation of the subpel unit with respect to the prediction unit of the frame to be omitted, there is an effect that a faster encoding process can be performed.

In addition, since interpolation and subpel unit search points required in the process of making a subpel can be omitted, a faster encoding process can be performed.

1A is a diagram illustrating depth-specific coding units (CUs) and prediction units (PUs) presented in MPEG-H HEVC.
1B is a diagram illustrating a relationship between a coding unit and a transform unit (TU).
2 is a diagram illustrating a process for determining the size of a coding unit for each data unit according to the prior art.
3 is a diagram illustrating an example in which a size of a coding unit is determined for each data unit.
4 is a diagram schematically illustrating a configuration of an encoder according to an embodiment of the present invention.
5 illustrates a general procedure for selecting a prediction mode in the current prediction unit.
FIG. 6 illustrates a prediction unit for performing subpel motion estimation in MPEG-H HEVC. FIG.
7 illustrates a prediction unit for performing subpel motion estimation in an encoder according to an embodiment of the present invention.
8 is a diagram illustrating a method of omitting subpel motion estimation performed by an encoder according to an embodiment of the present invention.

The present invention is capable of various modifications and various embodiments, and specific embodiments are illustrated and described in the drawings. However, this is not intended to limit the present invention to specific embodiments, it should be understood to include all transformations, equivalents, and substitutes included in the spirit and scope of the present invention. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS Hereinafter, the present invention will be described in detail with reference to the accompanying drawings.

The terms first, second, etc. may be used to describe various components, but the components should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another.

The terminology used herein is for the purpose of describing particular example embodiments only and is not intended to be limiting of the present invention. Singular expressions include plural expressions unless the context clearly indicates otherwise. In this application, the terms "comprise" or "have" are intended to indicate that there is a feature, number, step, operation, component, part, or combination thereof described in the specification, and one or more other features. It is to be understood that the present invention does not exclude the possibility of the presence or the addition of numbers, steps, operations, components, components, or a combination thereof.

The terms "part", "unit", "unit", "module", "block", and the like described in the specification mean units for processing at least one function or operation. Software or a combination of hardware and software.

DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENT Hereinafter, embodiments of the present invention will be described in detail with reference to the accompanying drawings. Referring to the accompanying drawings, the same or corresponding components are denoted by the same reference numerals, .

Hereinafter, a description will be given of an encoder for quickly determining a size of a coding unit according to depths, but it is natural that the same or similar technical concepts may be applied to a decoder.

In addition, in applying the size determination method of the coding unit, a hierarchical coding unit is used to consider image characteristics, and the maximum height, width, and maximum depth of the coding unit may be adaptively determined according to the characteristics of the image. It can be implemented to be set in various ways depending on the needs. However, when necessary for the convenience of explanation, the case where the maximum height and width of the coding unit are 64 and the maximum depth is 3 will be described as an example.

4 is a diagram schematically illustrating a configuration of an encoder according to an embodiment of the present invention, and FIG. 5 is a diagram illustrating a general process for selecting a prediction mode in a current prediction unit. FIG. 6 is a diagram illustrating a prediction unit that performs subpel motion estimation in MPEG-H HEVC, and FIG. 7 is a diagram illustrating a prediction unit that performs subpel motion estimation in an encoder according to an embodiment of the present invention.

Referring to FIG. 4, the encoder includes a previous frame information storage unit 410, a motion estimation unit 420, and an optimum mode determination unit 450. It is obvious to those skilled in the art that the encoder may further include a transform / scaling / quantization unit, an intra prediction unit, an entropy encoder unit, etc., but since the components are somewhat separated from the gist of the present invention, the description thereof will be provided. Will be omitted.

The previous frame information storage unit 410 stores depth information of the coding unit CU of the previous frame in which the optimum coding mode is determined before the current frame and the prediction unit PU for each coding unit CU. Stores partition size information. In addition, the previous frame information storage unit 410 may further store one or more of an index of each coding unit and an index of each prediction unit. Here, the index may be identification information for identifying the position of each coding unit in the frame.

The motion estimator 420 includes an integer pixel searcher 425, a determiner 430, a subpel searcher 435, and a cost calculator 440.

In general, since the motion estimator of the encoder according to the prior art performs sub-pel motion estimation for all prediction units (see FIG. 6), a large amount of computation is required and as a result, the real-time performance of the encoder is degraded. Had a problem.

In contrast, the motion estimation unit 420 according to the present embodiment relates to a position in the frame with respect to a coding unit (hereinafter, referred to as a 'current coding unit') to be encoded of the current frame, as described below. Locates the current coding unit by using information index (that is, CU index information), and grasps the size of the current coding unit by using depth information and partition size information, which is information split inside the current coding unit, A coding unit (hereinafter referred to as 'previous') that is located at the same position in a previous frame by using an index (that is, a CU internal index) that is information about a position where a prediction unit (hereinafter, referred to as a 'current prediction unit') exists within the coding unit. Coding Unit ') recognizes prediction units (hereinafter referred to as' previous prediction units') that exist at the same position within the By size and performs a sub-pel motion estimation only if the size of the current prediction unit matches the prediction unit (see Fig. 7) has a characteristic that a fast coding processing is possible.

The integer pixel search unit 425 obtains an integer multiple motion vector by performing motion estimation on an integer multiple unit in the designated search region.

The determination unit 430 refers to the depth information of the previous coding unit and the size information of the previous prediction unit stored in the previous frame information storage unit 410, and the size of the current prediction unit to be encoded coincides with the size of the previous prediction unit. Determine.

When the determination unit 430 determines that the size of the current prediction unit and the size of the previous prediction unit match, the subpel search unit 435 moves an integer multiple obtained by the integer pixel search unit 425 with respect to the current prediction unit. A final motion vector is obtained by predicting a motion vector of a subpel unit based on the vector. The subpel search unit 435 may perform motion estimation in units of one or more subpels of, for example, 1/2 pixel, 1/4 pixel, 1/8 pixel, and the like.

However, if the determination unit 430 determines that the size of the current prediction unit does not match the size of the previous prediction unit, the integer motion vector obtained by the integer pixel search unit 425 is output as the final motion vector.

The cost calculator 440 costs the cost of each prediction mode of the motion estimation prediction unit using the final motion vector or the like provided from the integer pixel search unit 425 or the subpel search unit 435. Calculate the value. As a method of calculating a cost value, one or more of a sum of absolute difference (SAD), a sum of squared difference (SSD), a sum of absolute transformed difference (SATD), and the like may be used, but is not limited thereto. Cost value calculation process is obvious to those skilled in the art, so a description thereof will be omitted.

The motion estimation unit 420 separately calculates cost values for each prediction mode (for example, SKIP, Inter 2Nx2N, Inter NxN, Inter 2NxN, Inter Nx2N, Intra 2Nx2N, Intra NxN, etc.), and calculates the calculated cost value. The best mode determiner 450 is provided to the best mode determiner 450, and the best mode determiner 450 determines a prediction mode of which the cost value is the minimum as the optimum coding mode of the macroblock.

Hereinafter, the distinctive features of the encoder according to the present embodiment as compared to the encoder according to the prior art will be briefly described with reference to FIGS. 5 to 7.

As shown in FIG. 5, in order to determine an optimal encoding mode, a cost value for each of a plurality of prediction modes is generally calculated, and a prediction mode having a minimum cost value is determined as an optimal encoding mode.

For reference, a pseudo code for determining an optimal coding unit applied in MPEG-H HEVC adopts a recursive structure.

Recursive_CU_Processing (depth, index) {

parent_cost = CU_processing (depth, index)

for from index = 0 to index = 3 do

   children_cost + = Recursive_CU_Processing (depth + 1, index)

end

if (parent_cost <children_cost)

   Best_CU = CU (depth)

else

   Best_CU = CU (depth + 1)

if (leaf node)

   return

}

As described above with reference to FIG. 2, the encoder performs RD cost estimation for four coding units corresponding to the size of a coding unit of a current depth and subsequent depths. By performing this process recursively and repeatedly, optimal coding unit sizes are determined for one maximum coding unit size (Maximum CU size). An example in which the size of the coding unit is determined for each data unit by the above process is as illustrated in FIG. 3.

In addition, for each prediction unit in the coding unit, an optimal coding mode is determined by calculating RD costs for each of a plurality of predetermined prediction modes as shown in FIG. 5.

When evaluating rate-distortion (RD) cost, prediction is performed, and prediction is performed for three modes such as SKIP, Inter, and Intra. Prediction of Inter and Intra modes is performed in six ways by partition size: Inter 2Nx2N, Inter NxN, Inter 2NxN, Inter Nx2N, Intra 2Nx2N, Intra NxN. In this way, each prediction unit PU obtains a total of seven methods of RD costs, and determines a mode having the best RD cost among them as an optimal coding mode.

Here, in order to calculate the RD cost for each prediction mode corresponding to the inter mode, the encoder according to the prior art has a problem that a large amount of computation is required by performing subpel motion estimation for all prediction units as shown in FIG. 6.

In contrast, the motion estimation unit 420 according to the present embodiment uses the depth information of the previous coding unit and the partition size information of the previous prediction unit to determine the size of the previous prediction unit in which the size of the current prediction unit in the current coding unit is finally determined. By calculating the subpel motion only when it is equal to, it is possible to minimize the amount of computation for the encoding process.

That is, as shown in FIG. 7, when the encoding mode of the reference frame (that is, the previous frame) is presented as 710, the subpel motion estimation is performed only when the sizes 720, 730, and 740 of the current prediction unit corresponding thereto are matched. By minimizing the calculation amount for the encoding process.

This process can be represented by the following pseudo code.

Motion_Estimation (depth, index, partition) {

Integer_Pel_motion_estimation ();

if (pre_PU_Size (depth, index, partition) == cur_PU_Size (depth, index, partition)

Sub_Pel_motion_estimation ();

return

}

Table 1 and Table 2 show the experimental results using the encoder with the subpel motion estimation omitting method according to the present embodiment and the encoder according to the prior art.

For reference, Table 1 is a comparison experiment with the encoder according to the prior art using the encoder according to the present embodiment, Table 2 is compared with the encoder according to the prior art using an encoder that does not perform any subpel motion estimation This is the result of the experiment.

△ Bitrate (%) △ Y-PSNR △ U-PSNR ΔV-PSNR △ time (%) Kimono 0.35 -0.03 0.00 0.00 -33.04 Parkscene 1.06 -0.05 -0.02 -0.01 -36.72 Cactus 1.00 -0.03 0.00 -0.01 -36.90 Basketballdrive 0.70 -0.02 0.02 0.02 -33.92 BQTarras 1.68 -0.04 -0.02 -0.03 -39.14 Average 0.96 -0.03 0.00 -0.01 -35.95

△ Bitrate (%) △ Y-PSNR △ U-PSNR ΔV-PSNR △ time (%) Kimono 0.94 -0.05 0.00 -0.02 -43.44 Parkscene 4.79 -0.11 0.00 -0.02 -43.64 Cactus 3.10 -0.07 0.02 0.02 -42.88 Basketballdrive 3.67 -0.07 0.00 0.00 -41.90 BQTarras 17.37 -0.13 0.00 0.01 -44.40 Average 5.97 -0.09 0.00 0.00 -43.25

Judging from Tables 1 and 2, it can be seen that when the encoder using the subpel motion estimation elimination method according to the present embodiment is used, the encoding process speed is maximized while the deterioration of image quality is minimized compared to the encoder according to the prior art. have.

8 is a diagram illustrating a method of omitting subpel motion estimation performed by an encoder according to an embodiment of the present invention.

Referring to FIG. 8, in step 810, the encoder determines whether the prediction mode of the current prediction unit for calculating the RD cost is an Inter mode.

If the prediction mode of the current prediction unit is not Inter mode (e.g., SKIP mode, Intra mode), the process proceeds to step 860 where the encoder determines the RD cost for the prediction mode according to the RD cost calculation method according to the prior art. To calculate.

However, if the prediction mode of the current prediction unit is Inter mode, the flow proceeds to step 820 where the encoder performs motion estimation on an integer multiple of the size of the current prediction unit.

In step 830, it is determined whether the size of the current prediction unit matches the size of the last determined previous prediction unit. As described above, the previous prediction unit is a prediction unit in the previous frame that exists at the same position as the current prediction unit identified using the stored index or the like.

If the size of the current prediction unit does not match the size of the previous prediction unit as a result of the determination in step 830, the encoder performs only a motion estimation on an integer multiple of the current prediction unit in step 840 (step 820). Is generated as the final motion vector. That is, if the size of the current prediction unit does not match the size of the previous prediction unit, the subpel unit motion estimation is omitted.

However, if the size of the current prediction unit matches the size of the last determined previous prediction unit as a result of the determination of step 830, the encoder obtains a motion vector obtained by further performing motion estimation of a subpel unit with respect to the current prediction unit in step 850. Is generated as the final motion vector.

Then, in step 860, the encoder calculates the RD cost for the specified prediction mode using the final motion vector generated by steps 820 through 850.

In step 870, the encoder determines whether the RD costs for all the specified prediction modes have been calculated to determine the optimal encoding mode. If the RD costs for all the specified prediction modes have not been calculated, go back to step 810.

However, if the RD cost is calculated for all the specified prediction modes, the process proceeds to step 880 where the encoder determines the prediction mode with the least RD cost as the optimal encoding mode.

It is apparent that the aforementioned subpel motion estimation omission method may be performed by an automated procedure according to a time series order by a software program or the like embedded in the encoder. The codes and code segments that make up the program can be easily deduced by a computer programmer in the field. In addition, the program is stored in a computer readable media, and read and executed by a computer to implement the method. The information storage medium includes a magnetic recording medium, an optical recording medium, and a carrier wave medium.

It will be apparent to those skilled in the art that various modifications and variations can be made in the present invention without departing from the spirit or scope of the invention as defined in the following claims And changes may be made without departing from the spirit and scope of the invention.

410: previous frame information storage unit
420: motion estimation unit
425 integer pixel search unit
430: judgment
435: subpel search unit
440: cost calculator
450: optimum mode determiner

Claims (16)

In the encoder,
A determination unit that determines whether the size of the current prediction unit matches the size of the previous prediction unit existing at the same position in the previous frame in which the encoding mode is determined; And
And a subpel search unit that performs motion estimation in sub-pel units only when the size of the current prediction unit matches the size of the previous prediction unit.
And an index is used for the current prediction unit to identify the position in the current frame and the current coding unit, and the previous prediction unit to identify the position in the previous frame and the previous coding unit, respectively.
delete The method of claim 1,
The determination unit may match a depth of a current coding unit including the current prediction unit and a depth of a previous coding unit including the previous prediction unit, and a partition size of the current prediction unit and a partition size of the previous prediction unit match. Determine whether the size of the current prediction unit matches the size of the previous prediction unit.
The method of claim 3,
And a previous frame information storage unit which stores information about a depth of the previous coding unit and a partition size of the previous prediction unit included in the previous coding unit.
The method of claim 1,
An integer pixel search unit may further include an integer pixel search unit configured to obtain an integer multiple motion vector by performing motion estimation in units of multiples of a designated search area.
And an integer motion vector or a motion vector of a subpel unit is output as a final motion vector according to whether the subpel motion estimation is performed.
The method of claim 5,
And a cost calculator configured to calculate a cost value for each prediction mode of the prediction unit of the motion estimation using the final motion vector.
The method according to claim 6,
The prediction mode is at least one of Inter 2Nx2N, Inter 2NxN, Inter Nx2N and Inter NxN.
The method according to claim 6,
And the cost value is calculated using at least one of a sum of absolut difference (SAD), a sum of squared difference (SSD), and a sum of absolute transformed difference (SATD).
In the method of omitting subpel motion estimation,
(a) the determining unit determining whether the size of the current prediction unit matches the size of the previous prediction unit existing at the same position in the previous frame in which the encoding mode is determined; And
(b) the sub-pel searcher performing a motion estimation of the subpel unit only if the partition size of the current prediction unit matches the partition size of the previous prediction unit,
The determination unit may match a depth of a current coding unit including the current prediction unit and a depth of a previous coding unit including the previous prediction unit, and a partition size of the current prediction unit and a partition size of the previous prediction unit match. And determining whether the size of the current prediction unit coincides with the size of the previous prediction unit.
delete 10. The method of claim 9,
Subpel movement, characterized in that the information on the depth of the previous coding unit and the partition size of the previous prediction unit corresponding to the previous coding unit is stored in a previous frame information storage unit before step (a). Estimation skip method.
10. The method of claim 9,
An index is used for the current prediction unit to identify the position in the current frame and the current coding unit and the previous prediction unit to identify the position in the previous frame and the previous coding unit, respectively. How to omit motion estimation.
10. The method of claim 9,
Prior to the step (b), a step of acquiring an integer multiple motion vector by performing motion estimation in integral units in a designated search area for performing motion estimation in the subpel unit,
The subpel motion estimation omission method according to claim 1, wherein the integer motion vector or the motion vector of a subpel unit is output as a final motion vector according to whether the subpel motion estimation is performed.
The method of claim 13,
And calculating a cost value for each prediction mode of the prediction unit of the motion estimation using the final motion vector.
15. The method of claim 14,
And the prediction mode is at least one of Inter 2Nx2N, Inter 2NxN, Inter Nx2N, and Inter NxN.
15. The method of claim 14,
And the cost value is calculated using at least one of a sum of absolut difference (SAD), a sum of squared difference (SSD) and a sum of absolute transformed difference (SATD).

Images