KR20170126817A - Fast video encoding method and apparatus for the same - Google Patents

Abstract

A method for encoding an image at high speed is disclosed. The method for encoding an image at high speed in a high-speed image encoding apparatus comprises the steps of: performing unidirectional prediction on an encoded current block; determining whether a motion vector derived by the unidirectional prediction or a motion vector difference of the motion vector is an integer pixel unit; and selecting a prediction mode on the current block from prediction modes including the unidirectional prediction except bidirectional prediction when the motion vector or the motion vector difference of the motion vector is the integer pixel unit. Accordingly, the image encoding speed can be improved.

Description

[0001] FAST VIDEO ENCODING METHOD AND APPARATUS FOR THE SAME [0002]

The present invention relates to a high-speed image coding method and apparatus, and more particularly, to a method and apparatus for high-speed image coding by omitting bi-directional prediction which greatly increases coding complexity in an image coding process.

High Efficiency Video Coding (HEVC), also referred to as H.265, is the next generation of video coding technology that has begun development due to the success of H.264 / MPEG-4 AVC. ISO / IEC MPEG, which developed H.264 / MPEG-4 AVC, and Video Coding Experts Group of ITU-T formed a team as Joint Collaborative Team on Video Coding (JCT-VC) And ITU has approved HEVC as the next generation final standard in Geneva, Switzerland.

In order to achieve high coding efficiency, HEVC has applied additional techniques such as supporting a block structure from a minimum of 4 × 4 units to 64 × 64 and supporting 35 directional and non-directional modes in the intra prediction mode, The computational complexity has greatly increased. The increase in computational complexity places a heavy load on the encoding device, which is a major obstacle to the spread of HEVC.

In particular, motion estimation (ME) is one of the processes that greatly affect computational complexity in image coding. In motion estimation, bidirectional prediction using a maximum of two reference pictures and unidirectional prediction using only one reference picture .

The existing high - speed image coding method mainly omits some prediction processes in unidirectional prediction, and a method of improving the speed of bidirectional prediction having a large computational complexity is not studied.

In some existing studies that reduce the computational complexity in bi-directional prediction, the distortion cost of the other prediction mode (unidirectional or merge mode) is used, or only a part of the process is omitted using the association between encoding units. The effect of contributing is not large. Therefore, there is a need for a method capable of greatly reducing the computational complexity by omitting the entire bi-directional prediction in some cases.

In order to solve the above problems, an object of the present invention is to provide a high-speed image coding method.

It is another object of the present invention to provide a high-speed image encoding apparatus.

According to one aspect of the present invention, there is provided a high-speed image encoding method in a high-speed image encoding apparatus.

Here, the high-speed image coding method may include performing unidirectional prediction on a current block to be coded, determining whether a motion vector or a motion vector difference derived from unidirectional prediction is an integer pixel unit And a step of selecting a prediction mode for a current block among the prediction modes including unidirectional prediction, except for bidirectional prediction, if the prediction mode is an integer pixel unit.

The step of performing unidirectional prediction includes the steps of determining an initial search start position using motion information of a block located around a position of a current block in a reference picture, Obtaining a first reference block by measuring the correlation between the reference block and the current block based on the initial search start position and measuring the correlation between the neighboring reference block and the current block on a subpixel basis, And a step of acquiring.

Here, the step of acquiring the first reference block by measuring in units of integer pixels may include measuring a sum of absolute difference (SAD) between the current block and the reference block.

Here, the step of measuring the sub-pixel and obtaining the second reference block may include measuring a Sum of Transform Difference (SATD) between the current block and the reference block.

Here, the subpixel unit may include at least one of a 1/2 subpixel unit and a 1/4 subpixel unit.

Herein, the performing the unidirectional prediction may include: selecting a block having an optimal rate-distortion cost as a prediction block for the current block by comparing the first reference block and the second reference block, It is possible to generate motion information including one.

Here, the current block may include at least one of 2N × 2N, 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N and nR × 2N block sizes .

According to another aspect of the present invention, there is provided a high-speed image encoding method in a high-speed image encoding apparatus.

Here, the high-speed image coding method may include selecting one of prediction modes including unidirectional prediction and bidirectional prediction for a current block having a block size of 2Nx2N on a pixel-by-pixel basis, and if the selected prediction mode is not bidirectional prediction, And selecting a prediction mode for a block having a size different from that of the current block from prediction modes other than prediction.

Here, in the step of selecting the prediction mode, a prediction mode having a reference block with the highest correlation between the current block and the reference block of the reference picture can be selected.

Here, the block having a size different from that of the current block may have at least one size of 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N, and nR × 2N.

According to another aspect of the present invention, there is provided a high-speed image encoding apparatus.

Here, the high-speed image encoding apparatus may include at least one processor and a memory storing instructions for instructing the at least one processor to perform at least one step.

Here, at least one step may include performing unidirectional prediction on a current block to be coded, determining whether a motion vector or a motion vector difference derived from the unidirectional prediction is an integer pixel unit And selecting the prediction mode for the current block from among the inter-picture prediction modes including unidirectional prediction, except for bidirectional prediction, if the prediction mode is an integer pixel unit.

The step of performing unidirectional prediction includes the steps of determining an initial search start position using motion information of a block located around a position of a current block in a reference picture, Obtaining a first reference block by measuring the correlation between the reference block and the current block based on the initial search start position and measuring the correlation between the neighboring reference block and the current block on a subpixel basis, And a step of acquiring.

Here, the step of acquiring the first reference block by measuring in units of integer pixels may include measuring a sum of absolute difference (SAD) between the current block and the reference block.

Here, the step of measuring the sub-pixel and obtaining the second reference block may include measuring a Sum of Transform Difference (SATD) between the current block and the reference block.

Here, the subpixel unit may include at least one of a 1/2 subpixel unit and a 1/4 subpixel unit.

Herein, the performing the unidirectional prediction may include: selecting a block having an optimal rate-distortion cost as a prediction block for the current block by comparing the first reference block and the second reference block, It is possible to generate motion information including one.

Here, the current block may include at least one of 2N × 2N, 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N and nR × 2N block sizes .

Here, at least one of the prediction modes is a step of, if the current block has a block size of 2N x 2N on a pixel basis and the prediction mode of the current block is selected from among the prediction modes excluding bidirectional prediction, And selecting a prediction mode for a block of a different size from the prediction mode.

Here, in the step of selecting the prediction mode, a prediction mode having a reference block with the highest correlation between the current block and the reference block of the reference picture can be selected.

Here, the block having a size different from that of the current block may have at least one size of 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N, and nR × 2N.

When the high-speed image encoding method and apparatus according to the present invention as described above is used, the computational complexity of the image encoding apparatus can be reduced.

In addition, there is an advantage that the coding rate can be improved.

1 is a conceptual diagram of an image encoding and decoding system according to an embodiment of the present invention.
2 is a block diagram of an image encoding apparatus according to an embodiment of the present invention.
3 is a configuration diagram of an image decoding apparatus 30 according to an embodiment of the present invention.
4 is an exemplary diagram for explaining a prediction unit in the HEVC.
5 is a flowchart for explaining a coding method according to reference software in HEVC.
6 is a first flowchart of a high-speed image coding method according to an embodiment of the present invention.
7 is a second flowchart of a high-speed image encoding method according to an embodiment of the present invention.
8 is a block diagram of a high-speed image encoding apparatus according to an embodiment of the present invention.

While the invention is susceptible to various modifications and alternative forms, specific embodiments thereof are shown by way of example in the drawings and will herein be described in detail. It should be understood, however, that the invention is not intended to be limited to the particular embodiments, but includes all modifications, equivalents, and alternatives falling within the spirit and scope of the invention. Like reference numerals are used for like elements in describing each drawing.

The terms first, second, A, B, etc. may be used to describe various elements, but the elements should not be limited by the terms. The terms are used only for the purpose of distinguishing one component from another. For example, without departing from the scope of the present invention, the first component may be referred to as a second component, and similarly, the second component may also be referred to as a first component. And / or < / RTI > includes any combination of a plurality of related listed items or any of a plurality of related listed items.

It is to be understood that when an element is referred to as being "connected" or "connected" to another element, it may be directly connected or connected to the other element, . On the other hand, when an element is referred to as being "directly connected" or "directly connected" to another element, it should be understood that there are no other elements in between.

The terminology used in this application is used only to describe a specific embodiment and is not intended to limit the invention. The singular expressions include plural expressions unless the context clearly dictates otherwise. In the present application, the terms "comprises" or "having" and the like are used to specify that there is a feature, a number, a step, an operation, an element, a component or a combination thereof described in the specification, But do not preclude the presence or addition of one or more other features, integers, steps, operations, elements, components, or combinations thereof.

Unless defined otherwise, all terms used herein, including technical or scientific terms, have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs. Terms such as those defined in commonly used dictionaries are to be interpreted as having a meaning consistent with the contextual meaning of the related art and are to be interpreted as either ideal or overly formal in the sense of the present application Do not.

A moving picture may be composed of a series of pictures and the term picture may be replaced with another term having an equivalent meaning to an image or a frame.

Further, each picture (or frame) can be divided into predetermined areas. Here, the divided area may be referred to as a block or a macro block, but it may be a coding tree unit (CTU), a coding unit (CU), a prediction unit (PU) ), A transform unit (TU), and the like. Accordingly, it is to be understood that the meaning of the blocks used throughout this specification may include all of the above terms.

Here, the encoding unit may mean a basic unit for performing intra-picture prediction or inter-picture prediction, and the prediction unit may be generated by dividing an encoding unit as a basic unit of prediction, and one prediction unit may generate a specific prediction mode Lt; / RTI > The conversion unit may be generated by dividing the encoding unit and may be a basic unit for converting a residual block generated as a difference value between the original image block and the prediction block generated in the prediction unit.

In addition, the divided area may be composed of several pixels or pixels again. For example, the encoding tree unit may have a size of one of 64 x 64, 32 x 32, and 16 x 16 in units of pixels.

In a coding tree unit, a block may be divided into a quad-tree type, in which the divided blocks may be referred to as a coding unit, and this division may be recursively performed, It does not need to be divided into the form. However, the encoding unit may have a maximum size and a minimum size. The encoding unit of the largest size may be referred to as a maximum encoding unit (LCU), the encoding unit of the smallest size may be referred to as a minimum encoding unit SCU).

Here, depth or depth can be used as an index representing the number or degree of division from the maximum encoding unit. For example, if the size is the same as the maximum encoding unit (LCU), the depth can be expressed as 0, and the depth can be increased by one as divided into a quad tree. If divided to the same size as the minimum encoding unit, Can be expressed as a maximum depth (Max depth).

Here, each coding unit (CU) can be further divided into prediction units (PU). The prediction units can be divided into asymmetric structures without being divided into a quad-tree form, and can be divided only once in the coding unit. And may have a predetermined form according to inter-picture prediction and intra-picture prediction. For example, intra prediction may have a size of 2N x 2N or N x N.

In addition, each coding unit (CU) can be divided again into a conversion unit (TU), where the conversion unit can be recursively partitioned on the basis of a quadtree structure and can have maximum and minimum sizes And can be represented by a corresponding depth.

The division from the encoding unit into the prediction unit or the conversion unit may be performed independently, but there may be restrictions. For example, when the encoding mode of the encoding unit is intra-picture prediction, the conversion unit divided from the encoding unit may not be larger than the size of the prediction unit divided from the encoding unit.

Each unit may consist of one luminance block and two color difference blocks, which may be otherwise configured according to the color format. Further, the size of the luminance block and the color difference block can be determined according to the color format. For example, in the case of 4: 2: 0, the size of the color difference block may have a length of 1/2 of the luminance block.

Conventional HEVC or H.264 / AVC terms can be referred to for the above unit or block unit. In the present invention, blocks and the terms may be used interchangeably and understood differently according to the standard technique, and they should be understood as terms or units corresponding to the encoding and decoding processes according to such standard techniques.

A reference picture, a reference block, or a reference pixel may be referred to as a reference picture, a block, or a pixel that is referred to in encoding or decoding a current block or a current pixel. Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

1 is a conceptual diagram of an image encoding and decoding system according to an embodiment of the present invention.

1, the image encoding apparatus 105 and the decoding apparatus 100 may be implemented as a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP) Player, a PlayStation Portable (PSP), a wireless communication terminal, a smart phone, a TV, a server terminal such as an application server and a service server, A communication device such as a communication modem for performing communication with a wired / wireless communication network, a memory (memories 120 and 125) for storing various programs for inter or intra prediction for encoding or decoding an image and memories (memories 120 and 125) And a processor (processor, 110, 115) for controlling the processor. The video encoded by the video encoding apparatus 105 can be transmitted through a wired or wireless communication network such as the Internet, a short-range wireless communication network, a wireless network, a WiBro network, or a mobile communication network in real time or non- And then transmitted to the image decoding apparatus 100 through various communication interfaces such as a serial bus (USB: Universal Serial Bus), etc., decoded by the image decoding apparatus 100, and restored and reproduced as an image. In addition, the image encoded by the image encoding apparatus 105 as a bit stream can be transferred from the image encoding apparatus 105 to the image decoding apparatus 100 via a computer-readable recording medium.

2 is a block diagram of an image encoding apparatus according to an embodiment of the present invention.

2, the image encoding apparatus 20 according to the present embodiment includes a predictor 200, a subtractor 205, a transformer 210, a quantizer 215, an inverse quantizer 220, An inverse transform unit 225, an adder 230, a filter unit 235, a decoded picture buffer 240, and an entropy encoding unit 245.

The prediction unit 200 can determine a prediction block through intra-picture prediction or inter-picture prediction with respect to a block to be coded. Here, the prediction block is a block which is understood to closely match the block to be encoded in terms of pixel difference, and can be determined by various methods including a sum of absolute difference (SAD) and a sum of square difference (SSD). At this time, syntaxes indicating various information necessary for decoding the blocks of the image slice can be generated.

The subtraction unit 205 may generate a residue image block by subtracting the pixel values of the prediction block from the pixel values of the current block to be encoded to derive pixel difference values.

The transforming unit 210 transforms the residual image block received from the subtracting unit 205 into residual transform coefficient values using various methods of transforming signals in the spatial domain such as Hamadard transform and discrete cosine transform into frequency domain signals Can be generated. That is, the residual image block can be converted from the pixel value signal to the frequency domain signal.

The quantization unit 215 may quantize the transform coefficients transformed by the transform unit 210 to reduce the bit rate. Through quantization it is possible to reduce the bit depth or amount of bits associated with some or all of the coefficients. The quantization error according to the degree of quantization can be changed by adjusting a quantization parameter (QP).

Dead Zone Uniform Threshold Quantization (DZUTQ), a quantization weighted matrix or an improved technique can be used for the quantization technique, and a detailed description of this technique can be found in the technical field of the present invention So that it can be easily understood by a person skilled in the art.

The entropy encoding unit 245 can encode the quantized transform coefficients through the quantization unit 215. [ For example, techniques such as context adaptive variable length coding (CAVLC), context adaptive binary arithmetic coding (CABAC), syntax based context adaptive binary arithmetic coding (SBAC), probability interval partitioning entropy (PIPE) .

In addition, the entropy encoding unit 245 can encode various information necessary for decoding the bit stream obtained by encoding the quantization coefficient.

The inverse quantization unit 220 and the inverse transform unit 225 may apply the inverse quantization and inverse transform to reconstruct the residual block with the pixel signal. The reconstructed residual block may be added to the prediction block in the adder 230 to reconstruct the reconstructed image block and may be stored in the decoded picture buffer 240 and used as a reference block.

The filter unit 235 may apply a deblocking filter to the restored image block to remove a blocking phenomenon if necessary, and may further apply a loop filter to improve the video quality .

3 is a configuration diagram of an image decoding apparatus 30 according to an embodiment of the present invention.

3, the image decoding apparatus 30 includes an encoding picture buffer 300, an entropy decoding unit 305, a predicting unit 310, an inverse quantization unit 315, an inverse transform unit 320, 325, a filter 330, and a decoded picture buffer 335.

In addition, the prediction unit 310 may include an intra prediction module and an inter prediction module.

First, when an image bitstream transmitted from the image encoding apparatus 20 is received, the image bitstream can be stored in the encoding picture buffer 300.

The entropy decoding unit 305 may decode the bitstream to generate quantized coefficients, motion vectors, and other syntax. The generated data may be transmitted to the predicting unit 310.

The prediction unit 310 may generate a prediction block for a block of the current image slice based on the data transmitted from the entropy decoding unit 305. [ At this time, based on the reference image stored in the decoded picture buffer 335, a reference picture list using a default construction technique may be constructed.

The inverse quantization unit 315 is provided as a bitstream and can dequantize the quantized transform coefficients decoded by the entropy decoding unit 305. [

The inverse transform unit 320 may apply the inverse DCT, inverse integer transform, or similar inverse transformation techniques to the transform coefficients to generate a residual block having the pixel signal.

The inverse quantization unit 315 and the inverse transformation unit 320 inversely perform the processes performed by the transform unit 210 and the quantization unit 215 of the image encoding apparatus 20 described above and may be implemented in various ways have. For example, the same process and inverse transformation that are shared with the transform unit 210 and the quantization unit 215 may be used, and information on the transform and quantization process (for example, transform size, transform Shape, quantization type, etc.), the transformation and quantization processes can be performed inversely.

The residual video block subjected to the inverse quantization and inverse transform process may be added to the prediction block derived by the prediction unit 310 to generate a reconstructed image block. This addition may be performed by the adder / subtracter 325.

The filter 330 may apply a deblocking filter to the reconstructed image block to remove a blocking phenomenon if necessary, and may further add other loop filters to improve the video quality before and after the decoding process. It can also be used.

The reconstructed and filtered image block may be stored in the decoded picture buffer 335.

4 is an exemplary diagram for explaining a prediction unit in the HEVC.

Referring to FIG. 4, when a coding unit is coded in the inter picture prediction mode in the HEVC, a prediction unit generated by division from the coding unit can be described.

More specifically, it can be equally divided as shown in 4a to 4d, and can be divided into a nonuniform division such as 4e to 4h or a form called an AMP (Asymmetric Motion Partition).

Here, when the size of the encoding unit is 2N x 2N, the N x N blocks of 4d may be possible only when the encoding unit is the minimum size, and the AMPs such as 4e to 4h may have the minimum size , It may not be used in consideration of the complexity of encoding and decoding.

Here, n of 4e to 4h may mean a value of 1/4 of 2N. For example, when 2N is 16, the value of n may be 4.

In video encoding, block types of predictive units, such as 4a to 4h, may be various. The image coding apparatus needs to determine which block type prediction method has the most optimal coding efficiency. At this time, by performing prediction according to various prediction methods for each type of prediction block, a block type having an optimal coding efficiency And a prediction mode can be selected.

Here, the prediction method may include inter prediction and intra prediction, and inter-picture prediction may include unidirectional prediction, bidirectional prediction, and weighted prediction when divided in more detail according to the prediction method . In addition, the inter picture prediction may further include a merge mode and an AMVP mode according to a method of expressing motion information.

The prediction method and the prediction mode may be used in combination, but it should be understood that the term includes both the prediction method and the prediction mode described above. It should be understood that the merge mode and the AMVP mode can be additionally applied to unidirectional prediction or bidirectional prediction .

Herein, the unidirectional prediction can determine the prediction block of the current block by selecting the reference block having the highest correlation with the current block from one reference picture, and bidirectional prediction can determine the prediction block of the current block using the maximum of two reference pictures You can decide. The weighted prediction can generate one prediction block by applying a weight and an offset to the reference block.

The merge mode is a technique for deriving a reference direction, a reference picture index, and a motion vector prediction value from a neighboring block. The AMVP mode is a technique for deriving only a motion vector prediction value from a neighboring block.

Representation of motion information and generation of syntax information according to each prediction method and mode can be easily understood by a person skilled in the art to which the present invention pertains, so a detailed description will be omitted.

5 is a flowchart for explaining a coding method according to reference software in HEVC.

As for the coding according to the reference software in the HEVC, the optimum prediction method / mode of the coding unit can be selected as shown in the flowchart of FIG.

Referring to FIG. 5, when determining a prediction method / mode of a coding unit (CU), it is first determined whether a merge skip mode (skip mode or skip / merge) is applied to a 2N × 2N block size coding unit You can decide. The merge skip mode is a special case of the merge mode. After quantization, transmission of the residual signal is omitted and only the selection information of neighboring blocks is transmitted in order to increase the efficiency when the transform coefficients for entropy encoding are all close to zero Lt; / RTI >

Unidirectional prediction (Uni-L0 pred, Uni-L1 pred), bidirectional prediction (Bi-Pred), and merging of the merge skip mode and the merge skip mode are performed for the prediction units of each of the block types (4a to 4h) Mode (Merge Pred), respectively.

In addition, when inter-picture prediction according to each block type is completed, Intra Pred can be performed, and a prediction method and a block type having the highest coding efficiency can be finally selected.

Referring to FIG. 5, unidirectional prediction for the reference picture list L1 can be omitted because L0 and L1, which are reference picture lists, are the same for a generalized B picture of an HEVC.

Further, in the case of unidirectional prediction or bidirectional prediction, Integer Motion Estimation / integer ME is performed in units of integer, and motion prediction (1/2 ME) Pixel-by-pixel motion prediction (1/4 ME).

As can be seen from the driving process of the HEVC reference software according to FIG. 5, since the bidirectional prediction as well as the unidirectional prediction is performed according to each block type (2N x 2N to AMP), the computational complexity for coding is very large.

Therefore, in the following, a description will be given of a method of omitting bidirectional prediction, which has a particularly high computational amount, among the inter picture prediction modes in some cases.

6 is a first flowchart of a high-speed image coding method according to an embodiment of the present invention.

Referring to FIG. 6, a high-speed image coding method in a high-speed image coding apparatus includes performing unidirectional prediction on a current block to be encoded (S100), calculating a difference between a motion vector or a motion vector derived from unidirectional prediction A step S110 of determining whether a motion vector difference is an integer pixel unit and a step S220 of selecting a prediction mode for a current block among prediction modes including unidirectional prediction except bi- ).

Here, the high-speed image encoding apparatus can designate the encoding apparatus 20 according to FIGS. 1 and 2, and FIG. 8 to be described later can be referred to together.

Here, the current block may indicate a prediction unit in the HEVC, but is not limited thereto.

Here, if motion information is derived in units of integer pixels in unidirectional prediction, it can be determined that the current block is a non-complex block, and if it is derived in units of pixels, the current block can be determined as a complex block. Therefore, if the motion information in the unidirectional prediction is in the case of the test pixel unit, the bidirectional prediction can be skipped boldly and the prediction according to the next prediction mode (for example, the merge mode) And the unidirectional prediction may be selected as the prediction mode.

The step of performing unidirectional prediction includes the steps of determining an initial search start position using motion information of a block located around a position of a current block in a reference picture, Obtaining a first reference block by measuring the correlation between the reference block and the current block based on the initial search start position and measuring the correlation between the neighboring reference block and the current block on a subpixel basis, And a step of acquiring.

Here, the determination of the initial search position is intended to limit the search range because it has excessive computational complexity when searching all the blocks in the reference picture. Specifically, when there are a plurality of pieces of motion information of neighboring blocks, a motion vector of a neighboring block having a minimum rate-distortion cost can be determined as a start position of a motion search. If there is no motion vector in a neighboring block, It can be determined as the start position of the search. Using a zero motion vector may mean using the same position in the reference picture as the prediction unit to be currently encoded as the start position of the motion search.

Here, the process of searching the neighboring reference block at the initial search position may be rhombic search, square search, sequential search, and the like.

Here, the step of acquiring the first reference block by measuring in units of integer pixels may include measuring a sum of absolute difference (SAD) between the current block and the reference block.

Here, the SAD can be derived from the following equation (1).

Referring to Equation (1), Bcur denotes a current block, Bref denotes a reference block, i, j denotes pixel coordinates of the current block, and k and l denote pixel coordinates of the reference block.

Here, the step of measuring the sub-pixel and obtaining the second reference block may include measuring a Sum of Transform Difference (SATD) between the current block and the reference block.

Here, SATD can be derived from the following equation (2).

Referring to Equation (2), the T () operator may mean Hadamard Transform.

Here, since the subpixel unit does not actually have a subpixel unit, it is possible to further apply the Hamadard transform by comparing it with Equation 1 for more precise consideration than the integer pixel unit. However, the Hamadard transform is not the same as Descret Cosine Transform (DCT) and entropy coding, but it can be used because it can derive a similar encoding result and can greatly reduce the computational complexity than DCT and entropy coding.

Here, the sub-pixel unit may include at least one of a ½ sub-pixel unit and a ¼ sub-pixel unit.

Herein, the performing the unidirectional prediction may include: selecting a block having an optimal rate-distortion cost as a prediction block for the current block by comparing the first reference block and the second reference block, It is possible to generate motion information including one.

Here, the current block may include at least one of 2N × 2N, 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N and nR × 2N block sizes .

7 is a second flowchart of a high-speed image encoding method according to an embodiment of the present invention.

Referring to FIG. 7, a high-speed image encoding method in a high-speed image encoding apparatus includes a step of selecting one of prediction modes including unidirectional prediction and bidirectional prediction for a current block having a block size of 2Nx2N on a pixel- If the selected prediction mode is not bi-directional prediction, selecting a prediction mode for a block having a size different from that of the current block among the prediction modes other than bi-directional prediction.

Here, the prediction mode includes unidirectional prediction and bidirectional prediction, and may further include merge mode, and may further include an AMVP mode.

Here, if the current block has a block size of 2N × 2N and bidirectional prediction is selected as an optimal prediction method / mode, the current block can be determined as a motion-rich or complex area, and 2N × N, N × 2N, It is reasonable to consider bidirectional prediction because blocks with sizes of NxN, 2NxnU, 2NxnD, nLx2N, and nRx2N are likely to be areas with more motion or more complexity.

On the other hand, if the current block has a block size of 2N × 2N and the bidirectional prediction is not selected as the optimal prediction method / mode, the current block can be regarded as a relatively simple area. Therefore, 2N × N, N × 2N, N Speed coding can be made possible by not considering bidirectional prediction for blocks having sizes of N, 2N x nU, 2N x nD, nL x 2N, and nR x 2N.

Here, in the step of selecting the prediction mode, a prediction mode having a reference block with the highest correlation between the current block and the reference block of the reference picture can be selected.

Here, the block having a size different from that of the current block may have at least one size of 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N, and nR × 2N.

8 is a block diagram of a high-speed image encoding apparatus according to an embodiment of the present invention.

8, a high-speed image coding apparatus 20 includes at least one processor 21 and a memory 21 for storing instructions for instructing the at least one processor 21 to perform at least one step, 22).

The high-speed image encoding apparatus 20 may further include an input / output module 23 for receiving an input from a user and displaying an encoding result or an image. The input / output module 23 may include a keyboard, a mouse, Device or the like.

Here, the high-speed image encoding apparatus 20 may further include a communication module 24 for receiving an image to be encoded by wire / wireless communication or transmitting an encoded image.

Here, the high-speed image encoding apparatus 20 may be, for example, a desktop computer, a laptop computer, a notebook, a smart phone, a tablet PC, A mobile phone, a smart watch, a smart glass, an e-book reader, a portable multimedia player (PMP), a portable game machine, a navigation device, a digital camera, a DMB digital multimedia broadcasting player, a digital audio recorder, a digital audio player, a digital video recorder, a digital video player, a PDA (Personal Digital Assistant) have.

Here, at least one step may include performing unidirectional prediction on a current block to be coded, determining whether a motion vector or a motion vector difference derived from the unidirectional prediction is an integer pixel unit And selecting the prediction mode for the current block from among the inter-picture prediction modes including unidirectional prediction, except for bidirectional prediction, if the prediction mode is an integer pixel unit.

The step of performing unidirectional prediction includes the steps of determining an initial search start position using motion information of a block located around a position of a current block in a reference picture, Obtaining a first reference block by measuring the correlation between the reference block and the current block based on the initial search start position and measuring the correlation between the neighboring reference block and the current block on a subpixel basis, And a step of acquiring.

Here, the step of acquiring the first reference block by measuring in units of integer pixels may include measuring a sum of absolute difference (SAD) between the current block and the reference block.

Here, the step of measuring the sub-pixel and obtaining the second reference block may include measuring a Sum of Transform Difference (SATD) between the current block and the reference block.

Here, the subpixel unit may include at least one of a 1/2 subpixel unit and a 1/4 subpixel unit.

Herein, the performing the unidirectional prediction may include: selecting a block having an optimal rate-distortion cost as a prediction block for the current block by comparing the first reference block and the second reference block, It is possible to generate motion information including one.

Here, the current block may include at least one of 2N × 2N, 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N and nR × 2N block sizes .

Here, at least one of the prediction modes is a step of, if the current block has a block size of 2N x 2N on a pixel basis and the prediction mode of the current block is selected from among the prediction modes excluding bidirectional prediction, And selecting a prediction mode for a block of a different size from the prediction mode.

Here, in the step of selecting the prediction mode, a prediction mode having a reference block with the highest correlation between the current block and the reference block of the reference picture can be selected.

Here, the block having a size different from that of the current block may have at least one size of 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N, and nR × 2N.

Table 1 is a table showing an experimental environment in which a high-speed image encoding method or a high-speed image encoding apparatus according to an embodiment of the present invention is performed.

Referring to Table 1, images (BasketballDrive, ParkScene, etc.) corresponding to the respective resolutions (1920 × 1080, 832 × 480, and 1280 × 720) were used as the experimental images, and the encoded frames were recorded for 10 seconds And reference software was HM 15.0 of HEVC. The quantization parameters 22, 27, 32, and 37 are used and low-delay is applied to the configuration.

Table 2 shows how a function for performing bidirectional prediction is omitted compared with the reference software HM 15.0 except for RoIMBC, which is a conventional fast encoding method, when a high-speed image encoding method and apparatus according to the present invention is applied .

Referring to Table 2, in comparison with the prior art UCU-EBS (upper CU-based early bi-prediction skipping method), the high-speed image encoding method and apparatus according to the present invention perform an average unnecessary bidirectional prediction process of 84% Can be omitted.

Table 3 shows experimental results showing the difference in compression rate and encoding time between the high-speed image encoding method and apparatus according to the present invention and the existing technology (UCU-EBS).

Referring to Table 3, a BD-rate measurement unit is used for each image sequence, and a positive real number value indicates an increase in bit amount (compression rate loss) required for the same image quality, and a negative real number value is the same (Compression rate gain) required for image quality. It has been confirmed that the use of the high-speed image encoding method and apparatus according to the present invention further reduces an average of 6% (%) in encoding time.

Table 4 shows how the UCU-EBS method and the high-speed image encoding method and apparatus according to the present invention reduce the number of memory accesses compared to RoIBMC, which is a technique for reducing memory issues that are critical to encoding complexity.

Referring to Table 4, it is confirmed that the high-speed image encoding method and apparatus according to the present invention have a 69% reduction in memory access and a 19% reduction in UCU-EBS compared to RoIBMC.

The methods according to the present invention can be implemented in the form of program instructions that can be executed through various computer means and recorded on a computer readable medium. The computer readable medium may include program instructions, data files, data structures, and the like, alone or in combination. The program instructions recorded on the computer readable medium may be those specially designed and constructed for the present invention or may be available to those skilled in the computer software.

Examples of computer-readable media include hardware devices that are specially configured to store and execute program instructions, such as ROM, RAM, flash memory, and the like. Examples of program instructions may include machine language code such as those produced by a compiler, as well as high-level language code that may be executed by a computer using an interpreter or the like. The hardware devices described above may be configured to operate with at least one software module to perform the operations of the present invention, and vice versa.

Furthermore, the above-mentioned method or apparatus may be implemented by combining all or a part of the structure or function, or may be implemented separately.

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 present invention as defined by the following claims It can be understood that

Claims (20)

In a high-speed image encoding method in a high-speed image encoding apparatus,
Performing unidirectional prediction on a current block to be encoded;
Determining whether a motion vector difference or a motion vector difference derived from the unidirectional prediction is an integer pixel unit; And
And selecting a prediction mode for the current block among the prediction modes including the unidirectional prediction except bidirectional prediction if the prediction mode is an integer pixel unit. In claim 1,
Wherein the performing the unidirectional prediction comprises:
Determining an initial search start position using motion information of a block located around a position of the current block in a reference picture;
Obtaining a first reference block by measuring a degree of correlation between the neighboring reference block and the current block based on the initial search start position in units of the integer pixels; And
And obtaining a second reference block by measuring a degree of correlation between a neighboring reference block and the current block on a subpixel basis based on the initial search start position. In claim 2,
Wherein the step of acquiring the first reference block by measuring in units of integer pixels comprises:
And measuring a sum of absolute difference (SAD) between the current block and the reference block. In claim 2,
Wherein the step of acquiring the second reference block by measuring the sub-
And measuring SATD (Sum of Transform Difference) between the current block and the reference block. In claim 2,
The sub-
Pixel sub-pixel unit and a 1/4 sub-pixel unit. In claim 2,
Wherein the performing the unidirectional prediction comprises:
The method of claim 1, further comprising: comparing the first reference block with the second reference block to select a block having an optimal rate-distortion cost as a prediction block for the current block and generating a motion including at least one of the motion vector or the motion vector difference Speed information encoding method. In claim 1,
Wherein the current block comprises:
Wherein the block size includes at least one of 2N × 2N, 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N and nR × 2N block sizes. In a high-speed image encoding method in a high-speed image encoding apparatus,
Selecting one of prediction modes including unidirectional prediction and bidirectional prediction for a current block having a block size of 2Nx2N on a pixel-by-pixel basis;
And selecting a prediction mode for a block having a size different from the current block from among the prediction modes other than the bidirectional prediction, if the selected prediction mode is not bi-directional prediction. In claim 8,
Wherein the step of selecting the prediction mode comprises:
Wherein a prediction mode having a reference block with the highest correlation between the current block and a reference block of the reference picture is selected. In claim 8,
The block having a size different from that of the current block,
2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N, and nR × 2N. At least one processor; And
A memory for storing instructions for directing the at least one processor to perform at least one step,
Wherein the at least one step comprises:
Performing unidirectional prediction on a current block to be encoded;
Determining whether a motion vector difference or a motion vector difference derived from the unidirectional prediction is an integer pixel unit; And
And selecting the prediction mode for the current block from among the inter-picture prediction modes including the unidirectional prediction, except bidirectional prediction, if the prediction mode is an integer pixel unit. In claim 11,
Wherein the performing the unidirectional prediction comprises:
Determining an initial search start position using motion information of a block located around a position of the current block in a reference picture;
Obtaining a first reference block by measuring a degree of correlation between the neighboring reference block and the current block based on the initial search start position in units of the integer pixels; And
And obtaining a second reference block by measuring a degree of correlation between a neighboring reference block and the current block based on the initial search start position on a subpixel basis. In claim 12,
Wherein the step of acquiring the first reference block by measuring in units of integer pixels comprises:
And measuring a sum of absolute difference (SAD) between the current block and the reference block. In claim 12,
Wherein the step of acquiring the second reference block by measuring the sub-
And measuring SATD (Sum of Transform Difference) between the current block and the reference block. In claim 12,
The sub-
Pixel sub-pixel unit, and a 1/4 sub-pixel unit. In claim 12,
Wherein the performing the unidirectional prediction comprises:
The method of claim 1, further comprising: comparing the first reference block with the second reference block to select a block having an optimal rate-distortion cost as a prediction block for the current block and generating a motion including at least one of the motion vector or the motion vector difference Speed information encoding apparatus. In claim 11,
Wherein the current block comprises:
And a block size of 2N × 2N, 2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N and nR × 2N blocks. In claim 11,
Wherein the at least one step comprises:
Wherein when the current block has a block size of 2N x 2N on a pixel basis and a prediction mode of the current block is selected from prediction modes other than the bidirectional prediction,
Further comprising the step of selecting a prediction mode for a block of a size different from the current block from among the prediction modes other than the bidirectional prediction. In claim 18,
Wherein the step of selecting the prediction mode comprises:
And selects a prediction mode having a reference block having the highest correlation between the current block and a reference block of the reference picture. In claim 18,
The block having a size different from that of the current block,
2N × N, N × 2N, N × N, 2N × nU, 2N × nD, nL × 2N, and nR × 2N.

Images