KR20150069110A - Method for video parallel processing based on dependency of intra prediction and apparatus for video decoding using the method - Google Patents

Abstract

Disclosed is a method for a high-speed parallel processing based on a dependency of an intra prediction, and a video decoding apparatus using the same. The video parallel processing method based on a dependency of an intra prediction comprises the steps of: detecting an intra-coded coding unit among at least one coding unit which divides coding tree units constituting a picture; determining an actual data dependency of an intra prediction with regard to the coding tree unit based on the intra-coded coding unit on the coding tree unit; and performing a parallel processing with regard to the coding tree unit by using the actual data dependency. Accordingly, it is possible to detect an actual data dependency between coding tree units in an intra prediction and to perform a parallel processing of a video based on the actual data dependency only which is detected.

Description

BACKGROUND OF THE INVENTION 1. Field of the Invention The present invention relates to a video parallel processing method based on dependency of intra prediction and a video decoding apparatus using the video parallel processing method and a video decoding apparatus using the video parallel processing method.

The present invention relates to encoding and decoding video, and more particularly, to a method for high-speed parallel processing of video based on intra-prediction dependency and a video decoding apparatus using the same.

In recent years, the use of video data through wired and wireless communication networks has been exploding due to the emergence of smart phones and smart TVs. Video data has superior information transmission capability than plain text data, but has a very large capacity, which makes it difficult to transmit, reproduce and store data in a network channel having a limited bandwidth. In addition, since a large amount of moving picture information is appropriately processed according to the demand of an application, a system for processing a moving picture also requires a high specification.

The standardization of HEVC (High Efficiency Video Coding) has been recently completed with the next generation video compression standard technology known to have a compression efficiency twice as high as H.264 / AVC.

The HEVC defines a coding unit (CU), a prediction unit (PU) and a transform unit (TU) having a quadtree structure. The sample adaptive offset (SAO: Sample Adaptive Offset), and deblocking filters. In addition, the conventional intra prediction and inter prediction are improved to improve compression encoding efficiency.

On the other hand, since the video codec decoder performs many operations, a parallel processing technique is frequently used for high speed decoding. Particularly, parallel processing in a video codec such as H.264 or HEVC is limited by data dependency caused by Intra prediction which is one of coding tools.

That is, conventional parallel processing techniques have performed parallel processing, assuming that all data dependencies that may be generated by the intra prediction algorithm exist.

However, a large number of data dependencies according to the conventional parallel processing techniques are not actually existed, but there is also a problem that it hinders parallel processing unnecessarily.

It is an object of the present invention to solve the above problems and provide a method of parallel processing video based on actual data dependency in performing intra prediction.

It is another object of the present invention to solve the above problems and provide an apparatus for fast decoding video using a multi-processor based on actual data dependency in performing intra prediction.

According to an aspect of the present invention, there is provided a method of video parallel processing based on intra-prediction dependency, comprising: detecting an intra-coded coding unit among at least one coding unit dividing a coding tree unit constituting a picture; Determining an actual data dependence of the intra prediction on the coding tree unit based on the location of the intra-coded coding unit on the coding tree unit, performing parallel processing on the coding tree unit using the actual data dependency .

Here, the coding tree unit may be decoded according to a raster scan order.

Wherein determining the actual data dependency of the intra prediction for the coding tree unit may determine the actual data dependency of the intra prediction based on the portion of the intra coded coding unit in contact with the boundary of the coding tree unit.

Wherein the step of determining the actual data dependency of the intra prediction for the coding tree unit comprises the steps of: if the intra-coded unit exists at a position adjacent to the left boundary of the coding tree unit, And the intra-coded unit exists at a position adjacent to the upper boundary of the coding tree unit, the coding tree unit determines that the coding tree unit is in a dependency relation with the coding tree unit located at the upper side of the coding tree unit .

Wherein the step of determining the actual data dependency of the intra prediction for the coding tree unit comprises the steps of: if an intra-coded unit exists at a position adjacent to the left and upper boundaries of the coding tree unit, When there is an intra coded unit in dependence on a coding tree unit positioned on the upper left side and in a position in contact with upper and right boundaries of the coding tree unit, the coding tree unit includes a coding tree unit located on the upper right side of the coding tree unit, Can be determined to be in a dependency relationship.

Wherein determining the actual data dependency of the intra prediction for the coding tree unit comprises determining whether the intra-coded unit does not exist at a position adjacent to the boundary of the coding tree unit or at least one of the lower or right side of the coding tree unit If there is an intra-coded unit at a location that is close together, the coding tree unit may determine that there is no dependency on another coding tree unit.

Wherein the step of determining the actual data dependency of the intra prediction for the coding tree unit comprises the steps of: if the intra-coded unit is equivalent to the coding tree unit, the coding tree unit determines the left, upper left, It can be determined that there is a dependency relation with the coding tree unit located at the side

The step of performing the parallel processing on the coding tree unit may include a step of generating a directed acyclic graph indicating a dependency between a plurality of coding tree blocks included in the picture using the actual data dependency of the intra prediction And parallelizing the plurality of coding tree blocks based on the directional bicycling graph.

Here, the parallel processing of the plurality of coding tree blocks may perform intra prediction by allocating each of a plurality of coding tree blocks having no actual data dependency of intra prediction to different processors.

Here, the video parallel processing method may be performed in a video decoding process.

According to another aspect of the present invention, there is provided an apparatus for decoding a video based on intra-prediction dependency according to an embodiment of the present invention. The apparatus includes a coding unit that detects an intra-coded coding unit among at least one coding unit that divides a coding- A unit detection module, a dependency determination module for determining an actual data dependence of the intra prediction on the coding tree unit based on the position of the intra-coded coding unit on the coding tree unit, And a parallel processing module for performing processing.

The video parallel processing method based on dependency of intra prediction according to an embodiment of the present invention as described above detects an actual data dependence between coding tree units in intra prediction and considers only the detected actual data dependency So that video can be processed in parallel.

In addition, there is an advantage that the multiprocessor can be effectively utilized by using the directional uncyclic graph based on the detected actual data dependency.

1 is a conceptual diagram for explaining a hierarchical structure of pictures according to HEVC.
FIG. 2 is a conceptual diagram for explaining a provisional data dependency of a CTU level. FIG.
FIG. 3 is a conceptual diagram for explaining an actual data dependence relation of a current coding tree unit according to an embodiment of the present invention.
4 is a conceptual diagram for explaining a process of inferring actual data dependency according to an embodiment of the present invention.
5 is an exemplary diagram illustrating an algorithm for determining an actual data dependency according to an embodiment of the present invention.
FIG. 6 is a conceptual diagram for explaining a directional bicyclic graph showing dependency between coding tree units according to conventional parallel processing and a processor allocation using the graph.
FIG. 7 is a conceptual diagram for explaining a directional bicyclic graph showing a dependency relation between coding tree units according to an embodiment of the present invention and a processor allocation using the graph. FIG.
8 is a conceptual diagram for explaining a video decoding apparatus based on intra-prediction dependency according to an embodiment of the present invention.
9 is a block diagram for explaining a configuration of a video decoding apparatus based on intra prediction dependency 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.

The Video Encoding Apparatus and the Video Decoding Apparatus to be described below may be implemented as a personal computer (PC), a notebook computer, a personal digital assistant (PDA), a portable multimedia player (PMP) Such as a portable multimedia player (PSP), a PlayStation Portable (PSP), a wireless communication terminal, a smart phone, a TV application server and a service server. A communication device such as a communication modem for performing communication with a user terminal or a wired or wireless communication network, a memory for storing various programs and data for inter-screen or intra-screen prediction for coding or decoding a picture, coding or decoding, And a microprocessor for executing and operating and controlling It can mean a variety of devices.

In addition, the image encoded by the video encoding apparatus can be transmitted in real time or in non-real time through a wired or wireless communication network such as the Internet, a local area wireless communication network, a wireless LAN network, a WiBro network, a mobile communication network, A serial bus, and the like, and can be decoded and reconstructed into an image and reproduced by an image decoding apparatus.

In general, a moving picture may be composed of a series of pictures, each picture may be divided into a slice or a tile, and a slice or tile may be divided into a coding tree unit (CTU) . Further, the CTU can be divided into a coding unit (CU). Here, the coding tree unit (CTU) is similar to the macro block of the existing codec (H.264), but can perform coding / decoding while varying the size. Also, in the present invention, the coding unit (CU) can be understood as equivalent to an existing block.

A block or pixel to be referred to in encoding or decoding a current block or a current pixel is referred to as a reference block or a reference pixel.

It is also to be understood that the term "picture" described below may be used in place of other terms having equivalent meanings such as image, frame, etc., If you are a child, you can understand.

In addition, although the video parallel processing method based on intra-prediction dependency according to the embodiment of the present invention and the video decoding apparatus using the method are described based on the HEVC, it can be applied to other codecs utilizing the intra prediction technique to be.

For example, the present invention can be applied to other codecs such as H.264. In this case, the coding tree unit CTU corresponds to a macro block and the coding unit CU corresponds to a block ). ≪ / RTI >

Hereinafter, preferred embodiments according to the present invention will be described in detail with reference to the accompanying drawings.

FIG. 1 is a conceptual diagram for explaining a hierarchical structure of a picture according to an HEVC, and FIG. 2 is a conceptual diagram for explaining a temporal data dependency of a CTU level.

The hierarchical structure of a picture will be described with reference to FIG. A video is composed of a sequence of pictures (or frames), and each picture may have a hierarchical structure. A picture may be divided into a plurality of slices or tiles.

Referring to FIG. 1, one picture may be divided into a slice boundary line 12 or a tile boundary line 11. For example, a picture may be divided into three slices or two tiles. Here, the tiles may have the shape of a rectangle.

In addition, each slice or tile may include a plurality of coding tree units (CTU), and the coding tree units (CTU) may be encoded or decoded in a raster scan order having zigzag orientation .

The coding tree unit (CTU) can be divided into a coding unit (CU), and the block size of the coding unit (CU) can be variably and adaptively determined.

The video data has temporal and spatial redundancy. Temporal redundancy means redundancy between consecutive pictures, and pixels of consecutive pictures have a very high correlation. Inter prediction is a method of compressing an image by eliminating temporal redundancy. It can compress an image by performing motion estimation using a picture located temporally before and after a current picture as a reference picture.

Spatial redundancy means redundancy existing in a picture, and a brightness value of one pixel has a high correlation with a brightness value of neighboring pixels. Intra prediction is a method of compressing an image by removing spatial redundancy using a degree of pixel correlation between blocks in one picture. The intra prediction is a method of compressing an image, And generates a predicted value of the current block using a reference pixel, and compresses the difference value between the generated predicted value and the pixel value of the current block.

Hereinafter, the data dependency in the description of the present invention means that the intra-prediction depends on the inter-CTU or inter-CU dependency.

Referring to FIG. 2, the current CTU to be currently encoded or decoded may be intra-predicted by referring to adjacent CTUs. Here, the arrow indicates the provisional data dependency of the CTU level. That is, the current CTU can perform intra-prediction with reference to the left CTU, the upper-left CTU, the upper CTU, and the upper-right CTU, and thus may have potential data dependency with four different CTUs.

However, there are not always data dependencies with four different CTUs in performing intra prediction for the current CTU.

FIG. 3 is a conceptual diagram for explaining an actual data dependence relation of a current coding tree unit according to an embodiment of the present invention.

Referring to FIG. 3A, a current CTU may be composed of a plurality of CUs, and each CU may be divided into intra-coded CUs through intra-prediction and inter-coded CUs through inter-prediction. There may also be coded CUs skipping intra prediction and inter prediction. Thus, the data dependency due to the intra prediction for the current CTU can be determined on the basis of the intra-coded CU. Here, the arrows can indicate the data dependency of the CU level.

Only the CUs located in A, B, and C in FIG. 3A may be intra-coded CUs. The CU located at A is in contact with the upper boundary of the CTU, the CU located at B is in contact with the left boundary of the current CTU, and the CU located in C is in contact with the lower and right boundaries of the CTU.

The determination of the actual data dependence on the current CTU based on the data dependence of the CU level is shown in FIG. 3B.

Referring to FIG. 3B, a CU located at A performs intra-prediction with reference to a CTU located at an upper side of the current CTU, and a CU located at B performs intra-prediction with reference to a CTU located at a left side of the current CTU. Also, a CU located in C does not refer to another adjacent CTU. This is because the CTU is encoded or decoded in raster scan order.

Therefore, the actual data dependence of the intra prediction for the current CTU can be represented as shown in FIG. 3B. That is, the current CTU is in a data dependency relationship with the CTUs located on the upper and left sides of the current CTU.

Table 1 below is a table for describing the actual data dependency based on the position of the CU according to an embodiment of the present invention.

Actual dependency (AD) in Table 1 represents the adjacent CTU with the actual data dependency in intra prediction at the CTU level.

For example, type number 0 indicates that there is no adjacent CTU with an actual data dependency if the intra-coded CU does not touch the left and top boundaries of the current CTU.

The type number 1 indicates that the adjacent CTU having the actual data dependency is the CTU located on the left side (L: Left) of the current CTU when the intra-coded CU touches only the left boundary of the current CTU.

In the case where the intra-coded CU is tangent to the left and upper boundaries of the current CTU, the adjacent CTU having the actual data dependency has the left (L: Left), upper left (UL: Up Left) : Up).

Type number 3 indicates that the adjacent CTU having the actual data dependency is the CTU located at the upper side (U: Up) of the current CTU when the intra-coded CU is touched only at the upper boundary of the current CTU.

In the case of type number 4, when the intra-coded CU touches the right and upper boundaries of the current CTU, the adjacent CTU having the actual data dependency is the CTU located on the upper side (U: Up) and the upper right side (UR) .

In the case where the intra-coded CU is equal to the current CTU, the adjacent CTU having the actual data dependency is classified into the left (L), the left upper (UL), the upper (U) (UR: Up Right).

Therefore, according to the embodiment of the present invention, it is possible to derive an actual dependency (AD) excluding a virtual dependency (VD) from a potential dependency (PD).

In detail, only the intra-coded CUs can be selected by analyzing the inside of the CTU at the CU level, and the data dependency of the CU level occurring at this time can be converted into the data dependency of the CTU level. Thus, video can be parallelized based on intra-prediction dependencies by assigning CTUs that are not dependent on each other to different processors based on the actual data dependence of the CTU level.

4 is a conceptual diagram for explaining a process of inferring actual data dependency according to an embodiment of the present invention. The arrows in FIG. 4 indicate the actual data dependency in intra prediction.

Referring to FIG. 4A, when the intra-coded CU is equal to the current CTU, the left (L), upper left (UL), upper (U: Up) and upper right ) Has a real data dependency with the current CTU. That is, if the intra-coded CU is equal to the current CTU, then the current CTU is actually dependent on four different CTUs.

In detail, according to the embodiment of the present invention, an actual dependency deduction (ADD) can be performed by hierarchically dividing a coding tree unit (CTU) or a coding unit (CU).

Referring to FIG. 4B, a current CTU corresponding to a type number 5 as shown in FIG. 4A is divided into four child nodes by a parent node. Here, each of the four child nodes dividing the parent node is located in the north-west (NW), northeast (NE), southwest (SW), and southeast (SE) directions from the parent node.

In Figure 4B, the child node located in the northwest (NW) has the actual data dependency corresponding to the type number 2, the child node located in the NE (NE) has the actual data dependency corresponding to the type number 4, It can be seen that the located child node has the actual data dependency corresponding to the type number 1 and the child node located at the SE is the actual data dependency corresponding to the type number 0.

Referring to FIG. 4C, the divided child nodes may be divided into parent nodes in FIG. 4B. However, a child node located in the SE (SE) corresponding to the type number 0 may be left as type number 0 without being divided.

A child node located at the north-west (NW) of type number 2 can be further divided into four child nodes (grandchild nodes), each of which has a type number 1, a type number 2, a type number 3, and a type number 0 .

In addition, the child node located at the NE of the type number 4 can be further divided into four child nodes (grandchild nodes), and each of these grandchildren corresponds to the type number 3, the type number 4, and the type number 0 can do.

In addition, the child node located at the SW in the type number 1 can be further divided into four child nodes (grandchild nodes), and each of these grandchild nodes may correspond to the type number 1 and the type number 0.

The relationship between the actual data dependency between the parent node and the child nodes according to FIG. 4 is summarized in Table 2 below.

That is, referring to FIG. 4, Actual Dependency Deduction (ADD) types between the parent node and the child nodes are summarized in Table 2.

5 is an exemplary diagram illustrating an algorithm for determining an actual data dependency according to an embodiment of the present invention.

An algorithm for determining actual data dependency according to an embodiment of the present invention will be described with reference to FIG.

In line 01, d means the set of actual data dependencies of the CTU level that make up AD. First, d is initialized with an empty set.

Line 02 performs depth-first search (DFS) from the root node (CTU). The ADD type of the root node is always 5 as shown in Table 1.

On lines 08 to 11, DDS is performed while deriving the ADD type of the child nodes (child nodes).

In line 09, the ADD type (denoted by x) of the lower nodes is calculated using an array called ADD_type_of_child. This arrangement is based on Table 2. The first index in the array, t, represents the ADD type of the parent node. The second index represents the position (NE, NW, SW, SE) of the child unit w. The return value of the array indicates the ADD type of the child node.

On line 10, the child node is called and the DFS recursively.

On lines 03 to 04, when arriving at the lowest-order node CU, it is checked whether the CU is an intra-coded CU.

In line 05, if the current node is identified as an intracoded CU, the set of actual data dependencies of the CTU level defined by the ADD type of the current node is aggregated with the set d. Array real data dependencies are given in Table 1.

The index of the array can represent the number of the ADD type, and the array can return the actual data dependency of the CTU level defined by the ADD type.

FIG. 6 is a conceptual diagram for explaining a directional bicyclic graph showing dependency between coding tree units according to conventional parallel processing and a processor allocation using the graph.

In FIG. 6A, each block represents a CTU, and a number assigned to each CTU may mean a sequence number of a CTU.

Referring to FIG. 6A, each CTU has a latent data dependency with another CTU. Conventionally, CTUs are processed in parallel based on temporal dependency (PD) including virtual dependency (VD) as well as actual dependency (AD).

For example, CTU # 5 has a data dependency with CTU # 0 and CTU # 1. Therefore, in order to perform intra prediction for the 5th CTU, it is necessary to perform decoding for the 0th and 1st CTUs first. Therefore, the core (or processor) performing the intra prediction for the 0th and 1st CTUs in the 5th CTU, A delay time may occur until the processing for the 0th and 1st CTUs is completed.

That is, Referring to Figure 6b, time 0 and time 1 CTU may be assigned a 5 CTU after processing by assigning to the first core to the second core, and therefore idle time by T _PD.1 (idle ) May occur.

Furthermore, idle time may occur in the process of allocating the 10th CTU or the 15th CTU to the second core or the third core, respectively.

FIG. 7 is a conceptual diagram for explaining a directional bicyclic graph showing a dependency relation between coding tree units according to an embodiment of the present invention and a processor allocation using the graph. FIG.

In FIG. 7A, each block represents a CTU, and a number assigned to each CTU may mean a sequence number of a CTU.

Referring to FIG. 7A, each CTU has an actual data dependency with another CTU. That is, according to an embodiment of the present invention, the CTU can be parallelized based on the actual dependency (AD) from which the virtual dependency (AD) is removed from the potential dependency (PD).

For example, CTU # 5 has no data dependencies with other CTUs. Therefore, the intra prediction for the 5th CTU can be performed simultaneously with the intra prediction for the 0th and 1st CTU.

That is, referring to FIG. 7B, the 0th and 1st CTUs may be assigned to the first core and processed and the 5th CTU may be allocated to the second core. Thus, video can be effectively processed in parallel without causing idle time as in Fig. 6B.

In particular, a video parallel processing method based on intra-prediction dependency according to an embodiment of the present invention can detect an intra-coded coding unit among at least one coding unit that divides a coding tree unit constituting a picture.

Next, the actual data dependence of the intra prediction on the coding tree unit can be determined based on the position of the intra-coded coding unit on the coding tree unit.

In particular, the intra-coded coding unit can determine the actual data dependency of the intra prediction based on the boundary of the coding tree unit with the boundary.

For example, if an intra-coded unit exists at a position adjacent to the left boundary of the coding tree unit, the coding tree unit can determine that it is in dependency with the coding tree unit located at the left side of the coding tree unit, If an intra-coded unit exists at a position adjacent to the upper boundary, the coding tree unit may determine that the coding tree unit is in dependency with a coding tree unit located on the upper side of the coding tree unit.

In addition, when an intra-coded unit exists at a position adjacent to the left and upper boundaries of the coding tree unit, the coding tree unit can determine that the coding tree unit is in dependency with the coding tree unit located at the upper left of the coding tree unit, The coding tree unit may determine that the coding tree unit is in a dependency relation with the coding tree unit positioned at the upper right side of the coding tree unit when the intra-coded unit exists at the position adjacent to the upper and the right boundary of the coding tree unit.

If there is no intra-coded unit at a position adjacent to the boundary of the coding tree unit, or if there is an intra-coded unit at a position adjacent to at least one of the lower or right side of the coding tree unit, It can be determined that there is no dependency with the tree unit.

Furthermore, if the intra-coded unit is equivalent to the coding tree unit, the coding tree unit can determine that it is in dependency with the coding tree unit located on the left, top left, top and right side of the coding tree unit.

Finally, parallel processing of the coding tree unit can be performed using the actual data dependency.

Specifically, a directed acyclic graph is generated that shows the dependency between a plurality of coding tree blocks included in a picture by using the actual data dependency of intra prediction, and a directed acyclic graph is generated based on the directional non- Tree blocks can be processed in parallel. That is, it is possible to perform intra prediction by allocating each of a plurality of coding tree blocks having no actual data dependency of intraprediction to each other to different processors.

FIG. 8 is a conceptual diagram for explaining a video decoding apparatus based on dependence of intraprediction according to an embodiment of the present invention, FIG. 9 is a view for explaining a configuration of a video decoding apparatus based on dependency of intra prediction according to an embodiment of the present invention Fig.

8, the decoding apparatus according to an embodiment of the present invention includes an entropy decoding unit 110, an inverse transform unit 140, an intra prediction unit 120, an inter prediction unit 130, an adder 150, A deblocking filter unit 160, and an SAO performing unit 170. [ However, FIG. 8 illustrates essential components for explaining a decoding apparatus according to an embodiment of the present invention, and it is apparent to those skilled in the art that the decoding apparatus may further include other components.

In the first step, the entropy decoding unit 110 receives entropy-encoded bitstreams from the encoding apparatus and entropy-decodes them. Here, the entropy encoding can entropy-encode quantization result values calculated by quantization using Context-Adaptive Variable Length Coding (CAVLC) or Context-Adaptive Binary Arithmetic Coding (CABAC) Information necessary for decoding can be entropy-encoded.

In the second step, the inverse transform unit 140 may restore the residual image by transforming the frequency domain values (frequency coefficients) provided in the inverse quantization unit (not shown) from the frequency domain to the spatial domain. In addition, the intra prediction unit 120 performs intra prediction, the inter prediction unit 130 performs inter prediction, and the addition unit 150 performs inverse transformation on the prediction image generated by intra prediction or inter prediction, The reconstructed image of the input image may be generated by adding the residual image reconstructed by the reconstructing unit 140 to the frame memory (not shown).

In the third step, the deblocking filter unit 160 performs deblocking filtering on the restored image, and the SAO performing unit 170 performs a sample adaptive offset (SAO) on the restored image. Here, the deblocking filtering can smooth the boundary between the blocks to improve the picture quality of the encoded or decoded picture, and the SAO can reduce the distortion between the original picture and the restored picture generated through the encoding process such as quantization, The distortion can be compensated by using the difference value of the pixels of the reconstructed image or using a filter using the coefficient (SAO coefficient) based thereon.

9, the intraprediction unit 120 according to an embodiment of the present invention includes a coding unit detection module 121, a dependency determination module 123, and a parallel processing module 125 for parallel processing of video . 9, the coding unit detection module 121, the dependency determination module 123, and the parallel processing module 125 are included in the intra prediction unit 120. However, the coding unit detection module 121, to be.

The coding unit detection module 121 can detect an intra coded coding unit among at least one coding unit that divides a coding tree unit constituting a picture.

Dependency determination module 123 may determine the actual data dependence of the intra prediction on the coding tree unit based on the location of the intra coding unit on the coding tree unit. That is, the dependency determination module 123 can determine the actual data dependency of the intra prediction based on the portion of the intra-coded coding unit in contact with the boundary of the coding tree unit.

The parallel processing module 125 may perform parallel processing on the coding tree unit using the actual data dependency.

In detail, the parallel processing module 125 generates a directed acyclic graph showing a dependency between a plurality of coding tree blocks included in a picture using the actual data dependency of the intra prediction, The plurality of coding tree blocks can be processed in parallel based on the graph. That is, the parallel processing module 125 can perform intra prediction by allocating each of the plurality of coding tree blocks, which have no actual dependency of intra prediction, to different processors.

Although the respective components of the decoding apparatus according to the above-described embodiments of the present invention have been described as being arranged in the respective constituent parts for convenience of explanation, at least two of the constituent parts may be combined to form one constituent part, It is to be understood that the invention is not limited to the disclosed embodiments, but, on the contrary, is intended to cover various modifications and equivalent arrangements included within the spirit and scope of the appended claims.

The video parallel processing method based on intra-prediction dependency according to the embodiment of the present invention and the video decoding apparatus using the intra-prediction dependency of the intra-prediction method according to the present invention detect the actual data dependency between the coding tree units in intra prediction, Video can be processed in parallel.

In addition, the multiprocessor can be effectively utilized by using a directional bicyclic graph based on the detected actual data dependency.

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

110: Entropy decoding unit 120: Intra prediction unit
121: Coding unit detection module 123: Dependency determination module
125: parallel processing module 130: inter prediction unit
140: Inverse transform unit 150:
160: deblocking filter unit 170: SAO performing unit

Claims (12)

A video parallel processing method comprising:
Detecting an intra coded coding unit among at least one coding unit dividing a coding tree unit constituting a picture;
Determining an actual data dependency of intra prediction on the coding tree unit based on a position of the intra coded coding unit on the coding tree unit; And
And performing parallel processing on the coding tree unit using the actual data dependency. ≪ Desc / Clms Page number 21 > The method according to claim 1,
The coding tree unit comprises:
Wherein the decoding is performed in accordance with a raster scan order. The method according to claim 1,
Wherein determining an actual data dependency of intra prediction for the coding tree unit comprises:
Wherein the intra-coded coding unit determines an actual data dependency of the intra prediction based on a portion in contact with the boundary of the coding tree unit. The method of claim 3,
Wherein determining an actual data dependency of intra prediction for the coding tree unit comprises:
Wherein when the intra-coded unit exists at a position adjacent to the left boundary of the coding tree unit, the coding tree unit is in dependency relation with a coding tree unit located at the left side of the coding tree unit,
Wherein when the intra-coded unit exists at a position adjacent to an upper boundary of the coding tree unit, the coding tree unit determines that the coding tree unit is in a dependency relation with a coding tree unit located on the upper side of the coding tree unit. A video parallel processing method based on prediction dependencies. The method of claim 3,
Wherein determining an actual data dependency of intra prediction for the coding tree unit comprises:
Wherein when the intra-coded unit exists at a position adjacent to the left and upper boundaries of the coding tree unit, the coding tree unit is in dependency with a coding tree unit positioned at the upper left of the coding tree unit,
And determines that the coding tree unit is in a dependency relationship with a coding tree unit located at an upper right side of the coding tree unit when the intra-coded unit exists at a position adjacent to the upper and right boundaries of the coding tree unit Based on the dependency of intra prediction. The method of claim 3,
Wherein determining an actual data dependency of intra prediction for the coding tree unit comprises:
When the intra-coded unit does not exist at a position adjacent to the boundary of the coding tree unit or when the intra-coded unit exists at a position adjacent to at least one of the lower side and the right side of the coding tree unit, Is determined to be independent of other coding tree units. ≪ RTI ID = 0.0 > [10] < / RTI > The method of claim 3,
Wherein determining an actual data dependency of intra prediction for the coding tree unit comprises:
And determines that the coding tree unit is in a dependency relationship with a coding tree unit located on the left, upper left, upper and upper right sides of the coding tree unit when the intra-coded unit is equivalent to the coding tree unit A video parallel processing method based on dependency of intra prediction. The method according to claim 1,
Wherein the step of performing parallel processing on the coding tree unit comprises:
Generating a directed acyclic graph showing a dependency between a plurality of coding tree blocks included in the picture using an actual data dependence of the intra prediction; And
And parallel processing the plurality of coding tree blocks based on the directional bi-cycling graph. The method of claim 8,
Wherein the step of parallel processing the plurality of coding tree blocks comprises:
Wherein intra prediction is performed by allocating each of the plurality of coding tree blocks having no real data dependency of the intra prediction to different processors, and performing intra prediction. The method according to claim 1,
The video parallel processing method includes:
Video decoding process based on intra-prediction dependence of intra-prediction. A coding unit detection module for detecting an intra-coded coding unit among at least one coding unit dividing a coding tree unit constituting a picture;
A dependency determination module that determines an actual data dependency of intra prediction on the coding tree unit based on the position of the intra coded coding unit on the coding tree unit; And
And a parallel processing module for performing parallel processing on the coding tree unit using the actual data dependency. The method of claim 11,
The dependency determination module includes:
Wherein the intra-coded coding unit determines an actual data dependency of the intra prediction based on a portion in contact with the boundary of the coding tree unit.

Images