KR20120096848A - Apparatus and method for video decoding - Google Patents

Abstract

PURPOSE: An image encoding device and a method thereof are provided to efficiently decode an encoded image without overhead and to increase the performance of decoding. CONSTITUTION: A thread allocation unit(510) allocates two or more threads to first to third syntax elements. The thread allocation unit allocates one thread to fourth to fifth syntax elements. A decoding unit(520) processes two or more threads according to a pipeline scheme. The decoding unit decodes the first to fifth syntax elements.

Description

[0001] APPARATUS AND METHOD FOR VIDEO DECODING [0002]

Embodiments of the present invention relate to an image decoding apparatus and method, and more particularly, to an image decoding apparatus and method capable of efficient parallel decoding of an encoded image.

As digital TVs are becoming more common, users are demanding high resolution video services. For this reason, researches on video compression processing to provide a high resolution video service have been actively conducted.

Regarding video compression processing, H.264 / AVC is a video codec standard with the highest compression ratio. It is widely used in multimedia service fields such as digital broadcasting, multimedia player, and video conferencing. H.264 / AVC uses a higher complexity algorithm than other video codecs to support high compression rates. Therefore, a high performance processor should be used to encode / decode data according to H.264 / AVC.

Existing single core-based processors improve the performance by increasing the clock speed, there is a problem that can not increase the processing performance by a certain level due to the heat generation and power consumption of the processor. In order to solve this problem, researches on multi-core based processors that increase the number of processor cores and process data in parallel have been continuously conducted.

In order to improve decoding performance in H.264 / AVC decoders, various studies on parallel processing methods that can effectively operate in a multicore system have been conducted. Representatively, the data level parallelization method is a decoding method in which H.264 / AVC data is divided to be parallelized and processed through multiple threads.

By the way, the data level parallelization method has to decode according to parallel processing while maintaining dependency between data of H.264 / AVC, but the entropy encoding part has to decode sequentially, thus performing decoding according to parallel processing. Causes difficulties. In particular, CABAC (Context Adaptive Binary Arithmetic Coding) entropy encoding, which has a high compression ratio, is preferred due to an increasing demand for high resolution video, but has a disadvantage in that the encoding time is long.

Therefore, parallelism of H.264 / AVC data decoding is closely related to parallelism of entropy decoding. In this regard, various algorithms for parallelizing entropy decoding have been proposed, but most of the conventional algorithms have been studied for improving the performance of hardware decoders and focused on partial performance improvement of entropy decoding. There was a difficulty.

In order to solve the problems of the prior art as described above, the present invention is to propose an image decoding apparatus and method capable of efficiently parallel decoding the encoded video without a large overhead.

Other objects of the present invention may be derived by those skilled in the art through the following examples.

According to an embodiment of the present invention to achieve the above object, in the apparatus for decoding an image encoded according to the syntax element segmentation technique, the first syntax element included in the MBINFO group, the second included in the PRED group Allocate two or more threads for decoding the syntax elements and the third syntax element included in the CBP group, and one for the decoding of the fourth syntax element included in the SIGMAP group and the fifth syntax element included in the COEFF group. A thread allocator for allocating threads; And parallel processing the two or more threads and the one thread according to a pipeline technique to decode the first syntax element, the second syntax element, the third syntax element, the fourth syntax element and the fifth syntax element. Provided is an image decoding apparatus including a decoding unit.

The thread allocator allocates a first thread for decoding the first syntax element, allocates a second thread for decoding the second syntax element and the third syntax element, and assigns the fourth syntax element and the first thread. The third thread is allocated to the decoding of five syntax elements, and the decoding unit sequentially initiates the processing of the first thread, the processing of the second thread, and the processing of the third thread with a predetermined time gap. One thread, the second thread, and the third thread may be processed in parallel.

The decoding unit processes the first thread, the second thread, and the third thread in macroblock units, respectively, and starts processing of the second thread one time slot after the start of processing of the first thread; The processing of the third macroblock may be started one time slot after the start of the processing of the second thread.

The decoding unit decodes two or more first syntax elements in one time slot, decodes one or more second syntax elements and one or more third syntax elements in one time slot, and one or more of the fourth syntax elements. And decode one or more of the fifth syntax elements in one time slot.

The thread allocator allocates a first thread for decoding the first syntax element, allocates a second thread for decoding the second syntax element, and allocates a third thread for decoding the third syntax element. And assigning a fourth thread to the decoding of the fourth syntax element and the fifth syntax element, wherein the decoding unit processes the first thread, the second thread, and the third thread by a predetermined time gap. Processing of the thread and processing of the fourth thread may be sequentially started to process the first thread, the second thread, the third thread, and the fourth thread in parallel.

The decoding unit processes the first thread, the second thread, the third thread, and the fourth thread on a time slot basis, respectively, and processes the second thread after a time slot from a start time of processing of the first thread. To start the processing of the third macroblock one time slot after the start of the processing of the second thread, and to process the fourth macroblock one time slot from the starting of the processing of the third thread. May be initiated.

The decoding unit decodes two or more first syntax elements in one time slot, decodes two or more second syntax elements in one time slot, decodes two or more third syntax elements in one time slot, and At least one of the fourth syntax element and at least one of the fifth syntax elements may be decoded in one time slot.

Further, according to another embodiment of the present invention, in a method for decoding an image encoded according to a syntax element segmentation technique, the first syntax element included in the MBINFO group, the second syntax element included in the PRED group, and the CBP group Allocating two or more threads for decoding the included third syntax element, and allocating one thread for decoding the fourth syntax element included in the SIGMAP group and the fifth syntax element included in the COEFF group; And parallel processing the two or more threads and the one thread according to a pipeline technique to decode the first syntax element, the second syntax element, the third syntax element, the fourth syntax element and the fifth syntax element. There is provided a video decoding method comprising the step of.

An image decoding apparatus and method according to the present invention has an advantage of efficiently parallel decoding an encoded image without large overhead.

1 to 4 are diagrams for explaining a concept of decoding according to the H.264 / AVC standard.
5 is a block diagram illustrating a general configuration of an image decoding apparatus according to an embodiment of the present invention.
6 to 12 are diagrams for describing a decoding concept of an image decoding apparatus according to an embodiment of the present invention.
13 is a flowchart illustrating the overall flow of an image decoding method 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.

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

1 to 4 are diagrams for explaining a concept of decoding according to the H.264 / AVC standard. A concept of decoding according to the H.264 / AVC standard will be described below with reference to FIGS. 1 to 4.

H.264 / AVC is a video compression standard created by ISO / IEC and ITU. It has a higher compression rate than the existing video compression standard and is suitable for streaming over a network and is used in various multimedia applications. Among them, the decoding apparatus according to the H.264 / AVC standard includes entropy decoding (ED), inverse transformation (IT), and inverse quantization (IQ) according to its function. , Hereinafter referred to as "IQ", intra prediction (IP: Intra Prediction, hereinafter "IP"), motion compensation (MC: Motion Compensation, hereinafter "MC"), deblocking filter (DF: Deblocking Filter, hereinafter referred to as "DF").

The data generated according to the H.264 / AVC standard has a structure as shown in FIG. Referring to FIG. 1, a video sequence of H.264 / AVC is composed of a group of pictures (GOP), and each GOP is composed of a plurality of pictures or frames. Here, the frames constituting the GOP are classified into I frames using spatial redundancy, P frames using temporal redundancy, and B frames using reference frames in both directions. One picture is composed of one or more slices. In addition, one slice includes a plurality of macroblocks, and one macroblock includes a luma block and a chrominance block.

Meanwhile, the data level parallelization method, which is a typical conventional parallel decoding processing method of H.264 / AVC data, is a method of dividing and processing H.264 / AVC data to enable parallelism. More specifically, the data level parallelization method is H.264. According to the / AVC data unit, it is roughly classified into a parallel unit method of a frame unit, a parallel unit method of a slice unit, and a parallel unit method of a macroblock unit.

However, in the case of the frame-by-frame parallelization method, there is a limit to the improvement of the parallel decoding performance by the reference frame dependency of the P frame or the B frame. In addition, in the case of the parallelization method of a slice unit, there is no data dependency between slices, but there is a limitation that only an image encoded by dividing a frame into multiple slices and decoded in slice units can be parallelly decoded.

Due to the above problems, various studies have been conducted on the parallelization method in macroblock units. In the parallelization method of macroblock units, parallel decoding is performed by allocating macroblocks that can be processed simultaneously to each thread. At this time, there is a dependency as shown in FIG. 2 between the macroblocks. For example, for the IP processing of "Current MB", macro macro 1, macro block 3, macro block 3, and macro block 4 must first be IP processed.

In this regard, a representative 2D-wave parallelization method among macroblock parallelism methods allocates a thread to each arrow to process the macroblock along the arrow as shown in FIG. 3. In other words, parallel processing is performed by allocating threads horizontally in the direction of the arrow. For example, in FIG. 3, MB (4,0), MB (2,1), and MB (0,2) are processed simultaneously at the fifth time T5 while maintaining data dependency.

In addition, the 3D-wave method, which is an extension of the 2D-wave parallelization method described above, performs parallel processing of MC + IT / IQ IP + IT / IQ after preprocessing entropy decoding in order to proceed with data level parallelization as shown in FIG. 4. And perform parallel processing of the DF. The reason why the entropy decoding process is not preprocessed is that the ED process is not parallelized.

In particular, the higher the resolution of the image, the higher the compression rate of CABAC is used, CABAC has a disadvantage that the time required to perform entropy decoding because the compression ratio is good but the complexity is high. Accordingly, various studies have been conducted to parallelize CABAC entropy decoding in order to reduce the time required for decoding CABAC.

However, the H.264 / AVC encoding apparatus using CABAC performs encoding by sequentially performing binarization, context modeling, and arithmetic coding. Due to this, there is a problem that it is difficult to perform fast CABAC encoding through parallelism.

In accordance with the above problems, the entropy slice parallelization method has been proposed for the next-generation standard after the H.264 / AVC standard. While entropy slice parallelism has some similarities to conventional slice level parallelism, the existing slice level parallelization method divides a picture into several slices and decodes each slice unit, whereas entropy slice method slices only the CABAC part. Entropy decoding is performed by dividing by. The remaining decoding is performed using reconstructed data. However, the entropy slice method has a disadvantage in that the encoding efficiency is inferior.

Accordingly, the CABAC parallelization method using the syntax element partitioning technique has been proposed to improve the encoding efficiency and the parallelization performance that occur in CABAC parallelism. The syntax element splitting technique is a method of performing CABAC by dividing a syntax element (SE) into groups (syntax element groups). Here, the syntax elements are classified and encoded into MBINFO group, PRED group, CBP group, SIGMAP group, and COEFF group as shown in Table 1 below (hereinafter, for convenience of description, syntax elements included in MBINFO group) To "first syntax element", syntax element included in PRED group to "second syntax element", syntax element included in CBP group to "third syntax element", syntax element included in SIGMAP group " Fourth syntax element "and syntax elements included in the COEFF group are referred to as" fiveth syntax element ").

group Syntax element MBINFO mb_skip_flag, mb_type, sub_mb_type, mb_field_decoded flag,
end of slice flag PRED prev_intra4x4 pred_mode_flag, rem intra4x4_pred_mode, prev_intra8x8_pred_mode_flag, rem intra8x8_pred_mode, intra_chroma pred_mode, ref_idx_l0, ref_idx_l1, mvd_l0, mvd_l1 CBP transform_size_8x8_flag, coded_block_pattern, coded_block_flag SIGMAP significant_coeff flag, last_significant_coeff flag COEFF coeff_abs_level_minus1, coeff_sign_flag

The image decoding apparatus according to the present invention is to decode entropy encoded data in parallel classified into the above MBINFO group, PRED group, CBP group, SIGMAP group, and COEFF group. Hereinafter, an image decoding apparatus according to an embodiment of the present invention will be described in detail with reference to FIG. 5.

5 is a block diagram illustrating a general configuration of an image decoding apparatus according to an embodiment of the present invention.

Referring to FIG. 5, an image decoding apparatus 500 according to an embodiment of the present invention includes a thread allocator 510 and a decoder 520. Hereinafter, the function of each component will be described in detail.

The thread allocator 510 allocates a plurality of threads to a task for decoding data encoded (or more precisely entropy encoded) according to a syntax element splitting technique.

More specifically, the thread allocator 510 defines a decoding operation by grouping syntax element groups that can be decoded simultaneously by properly combining the aforementioned MBINFO group, PRED group, CBP group, SIGMAP group, and COEFF group, and accordingly, A plurality of threads are generated by allocating one thread for each decoding group that occurs.

The decoding unit 520 decodes the syntax elements belonging to each syntax element group by performing parallel processing on a plurality of allocated threads according to a pipeline technique. In other words, the decoding unit 520 starts processing of a plurality of threads with a predetermined time gap and performs decoding in parallel.

Hereinafter, referring to FIG. 6 to FIG. 12, a group of decodeable groups can be bundled and defined as one decoding operation to assign a thread, and the reason and effect thereof for performing parallel processing using a pipeline technique will be described in more detail. Let's go.

First, FIG. 6 illustrates an example of syntax elements classified into a plurality of groups according to a syntax element splitting scheme. That is, the syntax elements are classified into the MBINFO group, the PRED group, the CBP group, the SIGMAP group, and the COEFF group as described above. Therefore, when a video encoded according to a syntax element segmentation scheme is to be decoded, theoretically, one thread may be allocated to each of five syntax element groups, thereby performing decoding in parallel.

However, there is a data dependency between the syntax element groups as shown in FIG. This data dependence is caused by the speed of data due to the encoding order of data when the image is encoded according to the syntax element segmentation technique.

Therefore, referring to FIG. 7, the second syntax element included in the PRED group should be decoded after the first syntax element included in the MBINFO group is decoded, and the fourth syntax element included in the SIGMAP group is included in the PRED group. 2 The syntax element and the third syntax element included in the CBP group must be decoded after decoding, and the fifth syntax element included in the COFF group must be decoded after the fourth syntax element included in the SIGMAP group is decoded.

In order to perform parallel processing of CABAC decoding while maintaining the dependency of such data, as shown in FIG. 8, the first thread (Task 1) to the MBINFO group, the CBP group, the PRED group, the SIGMAP group, and the COEFF group, respectively. After sequentially assigning a fifth thread (Task 5), processing of the first to fifth threads is started at predetermined time intervals to perform parallel decoding in a pipelined manner having five tasks in a multicore system. Can be.

However, the residual data encoded in the SIGMAP group portion and the COEFF group portion is a level in the COEFF group after significant_coeff_flag and last_significant coeff_flag in the SIGMAP group are encoded when the encoding flag coded_block_flag is 1 as shown in FIG. 9. Coeff_abs_level minus1 and coeff_sign_flag, which are information values, are encoded. This encoding process causes strong data binding between the fourth syntax element included in the SIGMAP group and the fifth syntax element included in the COEFF group.

Therefore, as shown in FIG. 8, when parallel decoding is performed according to a pipeline technique by allocating a separate thread for each syntax element group, decoding of the fourth syntax element included in the SIGMAP group (that is, the fourth thread). Processing) and decoding of the fifth syntax element included in the COEFF group (that is, processing of the fifth thread) causes a large overhead, thereby greatly reducing the efficiency of parallel decoding according to the pipeline technique.

Accordingly, the image decoding apparatus 500 according to an embodiment of the present invention defines the decoding operation of the fourth syntax element included in the SIGMAP group and the decoding operation of the fifth syntax element included in the COEFF group as one decoding operation group. After allocating one thread, the parallel processing with the other threads is performed to decode and prevent the above overhead from occurring.

More specifically, according to an embodiment of the present invention, the thread allocator 510 may include a first syntax element included in the MBINFO group, a second syntax element included in the PRED group, and a third syntax element included in the CBP group. Allocate two or more threads for decoding and one thread for decoding the fourth syntax element included in the SIGMAP group and the fifth syntax element included in the COEFF group. In other words, the thread allocator 510 defines two or more decoding operations by appropriately combining the MBINFO group, the PRED group, and the CBP group, and allocates threads to the prescribed two or more decoding operations, respectively, and the SIGMAP group and the COEFF group. It defines a decoding for as one decoding operation and allocates one thread.

In this case, the decoding unit 520 processes the two or more threads and the one thread in parallel according to a pipelined technique, so that the first syntax element, the second syntax element, the third syntax element, the fourth syntax element and the fifth syntax are processed in parallel. Decode the elements.

According to the first embodiment of the present invention, the thread allocator 510 allocates a first thread for decoding the first syntax element, and allocates a second thread for decoding the second syntax element and the third syntax element. In addition, a third thread may be allocated to the decoding of the fourth syntax element and the fifth syntax element. In this case, the decoding unit 520 sequentially initiates the processing of the first thread, the processing of the second thread, and the processing of the third thread with a predetermined time gap, thereby paralleling the first thread, the second thread, and the third thread. Processing to perform decoding.

In other words, according to the first embodiment of the present invention, the image decoding apparatus 500 allocates a first thread Thread 1 to the decoding of the first syntax element included in the MBINFO group as shown in FIG. 10. Allocates a second thread for decoding the second syntax element included in the PRED group and the third syntax element included in the CBP group, and is included in the fourth syntax element and COEFF group included in the SIGMAP group. After allocating a third thread (Thread 3) for decoding the fifth syntax element, processing the first thread, processing the second thread, and processing the third thread with a predetermined time gap (ie, one time slot). Processing may be sequentially started to process the first thread, the second thread, and the third thread in parallel to perform decoding according to a pipeline technique.

Here, according to an embodiment of the present invention, as shown in FIG. 10, the decoding unit 520 may process the first thread, the second thread, and the third thread in units of time slots, respectively. In this case, the decoding unit 520 starts processing of the second thread one time slot after the start time of the processing of the first thread, and processes the third time slot one time slot after the start of the processing of the second thread. May be initiated.

Meanwhile, when decoding is performed according to the pipeline technique as described above, when only one syntax element is processed in one time slot, the overall decoding speed may be slowed down by increasing the number of synchronizations.

Therefore, according to an embodiment of the present invention, the decoding unit 520 can reduce the number of synchronization by processing a plurality of syntax elements in one time slot for each thread to prevent a decrease in decoding speed. . In other words, the decoding unit 520 decodes two or more first syntax elements in one time slot, decodes one or more second syntax elements and one or more third syntax elements in one time slot, and one or more fourths. Syntax elements and one or more fifth syntax elements may be decoded in one time slot.

Further, according to the second embodiment of the present invention, the thread allocator 510 allocates a first thread for decoding the first syntax element, allocates a second thread for decoding the second syntax element, A third thread may be allocated for decoding of three syntax elements, and a fourth thread may be allocated for decoding of the fourth syntax element and the fifth syntax element. In this case, the decoding unit 520 sequentially initiates the processing of the first thread, the processing of the second thread, the processing of the third thread, and the processing of the fourth thread with a predetermined time gap, thereby causing the first thread and the second thread. The third thread and the fourth thread may be processed in parallel to perform decoding.

In other words, according to the second embodiment of the present invention, as shown in FIG. 11, the image decoding apparatus 500 allocates a first thread to the decoding of the first syntax element included in the MBINFO group. Allocate a second thread (Thread 2) to the decoding of the second syntax element included in the PRED group, allocate a third thread (Thread 3) to the decoding of the third syntax element included in the CBP group, After allocating a fourth thread (Thread 4) to the decoding of the fourth syntax element included in the SIGMAP group and the fifth syntax element included in the COEFF group, processing of the first thread with a predetermined time gap and processing of the second thread The processing of the third thread, the processing of the third thread, and the processing of the fourth thread may be sequentially started, and the first thread, the second thread, the third thread, and the fourth thread may be processed in parallel to perform decoding according to the pipeline technique.

Here, according to the exemplary embodiment of the present invention, the decoding unit 520 may process the first thread, the second thread, the third thread, and the fourth thread on a time slot basis as described above. In this case, the decoding unit 520 starts processing of the second thread one time slot after the start time of the processing of the first thread, and processes the third time slot one time slot after the start of the processing of the second thread. In this example, the processing of the fourth time slot may be started one time slot after the start of the processing of the third thread.

In addition, according to an embodiment of the present invention, as described above, the decoding unit 520 decodes two or more first syntax elements in one time slot, decodes two or more syntax elements in one time slot, and The third syntax element may be decoded in one time slot, and one or more fourth syntax elements and one or more fifth syntax elements may be decoded in one time slot. Accordingly, it is possible to solve the problem of slowing down the overall decoding speed by increasing the number of synchronizations by processing only one syntax element in one time slot.

Accordingly, the image decoding apparatus 500 according to the present invention can efficiently decode data according to an image in parallel without great overhead.

Hereinafter, the results of the parallel decoding simulation by the image decoding apparatus 500 according to the present invention will be described with reference to Table 2 and FIG. 12.

In this simulation, the ITU-T SG16 / Q6 Video Coding Experts Group (VCEG) is a series of activities operated by the ITU-T SG16 / Q6 Video Coding Experts Group (VCEG) to explore relevant core technologies as a preliminary work for H.264 + or H.265 standardization. The simulation was performed using a KTA 2.7 decoder developed by the Key Technology Area (KTA). Here, the image encoded using the KTA 2.7 encoder is used as the simulation target image based on the image encoding profile provided by KTA 2.7. The operating system uses Linux ubuntu 9.10, kernel version 2.6.31, CPU uses Intel quad core i5 processor, compiler uses gcc v4.4.1, and parallelization technique uses OpenMP [21,22].

Table 2 below shows the decoding processing time results of the image decoding apparatus 500 according to an embodiment of the present invention, and FIG. 12 is a diagram illustrating Table 2 in graph form.

before MT-SEP MT-SEP (3) MT-SEP (4) MT-SEP (5) μm μm % μm % μm % mobcal HD,
1280 × 720 42976 29947 30% 28481 34% 28481 13% stockholm HD,
1280 × 720 57066 44106 23% 43701 23% 43701 0% shileds HD,
1280 × 720 36418 25091 31% 21804 40% 21804 22% blue_sky FHD,
1920 × 1088 60446 41091 32% 26924 55% 38707 36% pedestrian_area FHD,
1920 × 1088 57079 39180 31% 27768 51% 37608 34% sunflower FHD,
1920 × 1088 61322 38752 37% 27589 55% 38431 37% rush_hour FHD,
1920 × 1088 52754 35706 32% 24734 53% 31552 40%

More specifically, in Table 2 and FIG. 12, "Multi-Threaded Syntax Element Partitioning (MT-SEP) 3" is used for image decoding according to the first embodiment of the present invention, which allocates three threads to a decoding operation. Corresponding to the apparatus 500, "MT-SEG (4)" corresponds to the image decoding apparatus 500 according to the second embodiment of the present invention that allocates four threads for the decoding operation. In addition, the result related to "MT-SEG (5)" means a processing result when image decoding is performed by assigning five threads to the decoding operation.

Referring to Table 2 and FIG. 12, when the decoding of the fourth syntax element included in the SIGMAP group and the decoding of the fifth syntax element included in the COEFF group are processed in separate threads as described above, the decoding processing time This increase can be seen.

In addition, referring to Table 2 and FIG. 12, when decoding is performed by the image decoding apparatus 500 according to the first and second embodiments of the present invention, it is confirmed that decoding performance is improved by about 23% to 55%. Can be. In particular, when decoding is performed using the image decoding apparatus 500 according to the second embodiment of the present invention, CABAC decoding performance may be improved up to 55%.

13 is a flowchart illustrating the overall flow of an image decoding method according to an embodiment of the present invention. Hereinafter, a process performed for each step will be described.

First, in step S1310, two or more threads are allocated to the decoding of the first syntax element included in the MBINFO group, the second syntax element included in the PRED group, and the third syntax element included in the CBP group, and assigned to the SIGMAP group. One thread is allocated for decoding the fourth syntax element included and the fifth syntax element included in the COEFF group.

In operation S1320, the two or more threads and the one thread may be processed in parallel according to a pipeline technique to process a first syntax element, a second syntax element, a third syntax element, a fourth syntax element, and a fifth syntax element. Decode

According to an embodiment of the present invention, in step S1310, the first thread is allocated to the decoding of the first syntax element, the second thread is allocated to the decoding of the second syntax element and the third syntax element, and A third thread may be allocated for decoding of the four syntax elements and the fifth syntax element. In this case, in the step S1320, the decoding may sequentially start the processing of the first thread, the processing of the second thread, and the processing of the third thread with a predetermined time gap, thereby allowing the first thread, the second thread, and the first thread to be processed. 3 threads can be processed in parallel.

According to another embodiment of the present invention, in step S1310, the first thread is allocated to the decoding of the first syntax element, the second thread is allocated to the decoding of the second syntax element, and the decoding of the third syntax element is performed. Assign a third thread for, and assign a fourth thread for decoding the fourth syntax element and the fifth syntax element. In this case, in step S1320, the process of the first thread, the process of the second thread, the process of the third thread, and the process of the fourth thread are sequentially started with a predetermined time gap, thereby allowing the first thread, the second thread, The third thread and the fourth thread can be processed in parallel.

The embodiments of the image decoding method according to the present invention have been described so far, and the configuration of the image decoding apparatus 500 described above with reference to FIG. 5 is also applicable to the present embodiment. Hereinafter, a detailed description will be omitted.

In addition, embodiments of the present invention may be implemented in the form of program instructions that may be executed by various computer means to be recorded on a computer readable medium. The computer readable medium may include program instructions, data files, data structures, etc. alone or in combination. The program instructions recorded on the medium may be those specially designed and constructed for the present invention or may be available to those skilled in the art of computer software. Examples of computer-readable media include magnetic media such as hard disks, floppy disks, and magnetic tape; optical media such as CD-ROMs and DVDs; magnetic media such as floppy disks; Examples of program instructions, such as magneto-optical and ROM, RAM, flash memory and the like, can be executed by a computer using an interpreter or the like, as well as machine code, Includes a high-level language code. The hardware device described above may be configured to operate as one or more software modules to perform the operations of one embodiment of the present invention, and vice versa.

As described above, the present invention has been described with reference to particular embodiments, such as specific elements, and limited embodiments and drawings. However, it should be understood that the present invention is not limited to the above- Various modifications and variations may be made thereto by those skilled in the art to which the present invention pertains. Accordingly, the spirit of the present invention should not be construed as being limited to the embodiments described, and all of the equivalents or equivalents of the claims, as well as the following claims, belong to the scope of the present invention .

Claims (10)

An apparatus for decoding an image encoded according to a syntax element partitioning technique,
Allocating two or more threads for decoding the first syntax element included in the MBINFO group, the second syntax element included in the PRED group, and the third syntax element included in the CBP group, and the fourth included in the SIGMAP group. A thread allocator for allocating one thread for decoding the fifth syntax element included in the syntax element and the COEFF group; And
Parallelizing the two or more threads and the one thread according to a pipeline technique to decode the first syntax element, the second syntax element, the third syntax element, the fourth syntax element and the fifth syntax element Decoding unit
Image decoding apparatus comprising a. The method of claim 1,
The thread allocation unit
Allocating a first thread for decoding the first syntax element, allocating a second thread for decoding the second syntax element and the third syntax element, and assigning a second thread to the decoding of the fourth syntax element and the fifth syntax element. Allocate a third thread for decoding,
The decoding unit sequentially initiates the processing of the first thread, the processing of the second thread, and the processing of the third thread with a predetermined time gap to parallel the first thread, the second thread, and the third thread. Image decoding apparatus characterized in that the processing. The method of claim 2,
The decoding unit
Process the first thread, the second thread, and the third thread on a time slot basis, respectively, and start processing of the second thread one time slot after the start of processing of the first thread; And the processing of the third macroblock is started one time slot after the start of the processing of the thread. The method of claim 3,
The decoding unit
Decode two or more of the first syntax elements in one time slot, decode one or more of the second syntax elements and one or more of the third syntax elements in one time slot, and one or more of the fourth syntax elements and one or more And decoding the fifth syntax element in one time slot. The method of claim 1,
The thread allocation unit
Allocating a first thread for decoding the first syntax element, allocating a second thread for decoding the second syntax element, allocating a third thread for decoding the third syntax element, Allocate a fourth thread for decoding four syntax elements and the fifth syntax element,
The decoding unit sequentially initiates the processing of the first thread, the processing of the second thread, the processing of the third thread, and the processing of the fourth thread with a predetermined time gap, thereby providing the first thread and the second thread. And processing the third thread and the fourth thread in parallel. The method of claim 5,
The decoding unit
Process the first thread, the second thread, the third thread, and the fourth thread on a time slot basis, respectively, and start processing of the second thread one time slot after the start of processing of the first thread; Starting the processing of the third macroblock one time slot after the start of the processing of the second thread, and starting the processing of the fourth macroblock one time slot after the starting of the processing of the third thread. Image decoding apparatus, characterized in that. The method of claim 6,
The decoding unit
Decode two or more of the first syntax elements in one time slot, decode two or more of the second syntax elements in one time slot, decode two or more of the third syntax elements in one time slot, And decode the fourth syntax element and the at least one fifth syntax element in one time slot. A method of decoding an image encoded according to a syntax element splitting technique,
Allocate two or more threads for decoding the first syntax element included in the MBINFO group, the second syntax element included in the PRED group, and the third syntax element included in the CBP group, and the fourth syntax element included in the SIGMAP group, and Allocating one thread for decoding the fifth syntax element included in the COEFF group; And
Parallelizing the two or more threads and the one thread according to a pipeline technique to decode the first syntax element, the second syntax element, the third syntax element, the fourth syntax element and the fifth syntax element step
Image decoding method comprising a. 9. The method of claim 8,
Allocating the thread
Allocating a first thread for decoding the first syntax element, allocating a second thread for decoding the second syntax element and the third syntax element, and assigning a second thread to the decoding of the fourth syntax element and the fifth syntax element. Allocate a third thread for decoding,
The decoding step may sequentially initiate the processing of the first thread, the processing of the second thread, and the processing of the third thread with a predetermined time gap, such that the first thread, the second thread, and the third thread are sequentially started. The video decoding method characterized in that the parallel processing. 9. The method of claim 8,
Allocating the thread
Allocating a first thread for decoding the first syntax element, allocating a second thread for decoding the second syntax element, allocating a third thread for decoding the third syntax element, Allocate a fourth thread for decoding four syntax elements and the fifth syntax element,
The decoding may sequentially start the processing of the first thread, the processing of the second thread, the processing of the third thread, and the processing of the fourth thread with a predetermined time gap, thereby providing the first thread and the first thread. And processing two threads, the third thread, and the fourth thread in parallel.

Images