JP2013162395A - Image coding method, image coding device, and image coding program - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image coding method which can cut back the amount of arithmetic operation when a target region to be coded is quadtree-divided and image coding is executed for each of divided smaller regions.SOLUTION: In an image coding method, an image is divided into a plurality of rectangles and those rectangles are further divided into smaller regions before being coded. The image coding method includes: a feature quantity calculation step of calculating feature quantity in a predetermined region of the image; a quadtree division step in which a target region to be coded is divided by quadtree on the basis of feature quantity in the predetermined region of the image which was calculated in the feature quantity calculation step with regard to the target region to be coded, and for each of the divided regions obtained by the quadtree division, determination is made of divided subregions which are further divided recursively by quadtree on the basis of the feature quantity calculated in the feature quantity calculation; and a coding step in which coding processing is performed for each of divided regions determined.

Description

  The present invention relates to an image encoding method, an image encoding apparatus, and an image encoding program that perform encoding by dividing an image into small rectangles.

  ITU-T standard video coding system, H.264. As a next generation ITU-T standardization technology next to the H.264 system, there is a HEVC (High Efficiency Video Coding) system. H. In the H.264 system and the HEVC system, an input video frame is divided into small rectangles and an encoding process is performed. In the H.264 system, it is called a macroblock, and in HEVC, it is called a Large Coding Unit (hereinafter referred to as LCU). In the HEVC system, an LCU is further divided into quadtrees to define an encoding target area Coding Unit (hereinafter referred to as CU), and encoding is performed in units of CUs. H. In the H.264 method and the HEVC method, a predictive coding method that generates a prediction signal of encoding target pixel data using already encoded peripheral pixels is employed.

  FIG. 5 shows an example of a CU that can be taken based on the quadtree division adopted in the HEVC method. FIG. 5 is a diagram illustrating an example of quadtree division of the encoding target region. In FIG. 5, a 64 × 64 pixel area is an LCU, and each divided area corresponds to a CU. As shown in FIG. 5, in the HEVC method, four CU sizes of 64 × 64 pixels, 32 × 32 pixels, 16 × 16 pixels, and 8 × 8 pixels are defined, and an appropriate CU is selected and encoded. By performing the processing, Compared to H.264, the encoding efficiency is significantly improved.

  A process (for example, see Non-Patent Document 1) for dividing the encoding target area into a quadtree will be described with reference to FIG. FIG. 6 is a flowchart showing a processing operation for dividing the encoding target area into quadtrees. As shown in FIG. 6, all possible CU sizes are selected while repeating quadtree partitioning on the encoding target region (step S41), and prediction processing is performed using the selected CU size (step S42). The quadtree split shape of the conversion target area is determined, and the CU size and the prediction mode are determined based on the quadtree split shape (step S43).

K. McCann, S. Sekiguci, B. Bross, W.-J. Han, “HM4: High Efficiency Video Coding (HEVC) Test Model 4 Encoder Description,” JCTVC-F802, Oct. 2011.

  However, an increase in the number of quadtree divisions in the encoding target area can improve coding efficiency, but there is a problem that the amount of computation required for determining an appropriate quadtree division shape increases. In the prior art of Non-Patent Document 1, when encoding is performed, prediction processing is performed for all CU sizes based on a predetermined number of quadtree divisions in a small rectangle to be encoded. FIG. 7 shows the relationship between the size of the small rectangle and the prediction processing calculation amount at this time. FIG. 7 is a diagram illustrating the relationship between the size of the small rectangle and the amount of prediction processing calculation. As shown in FIG. 7, as the LCU size and the number of divisions increase, the amount of calculation required for the prediction process increases. The numerical value in the figure is a value representing the amount of calculation, and is the amount of intra prediction processing calculation in the conventional method (the amount of calculation when intra prediction processing is performed with a prediction block size of 4 × 4 pixels is 1).

  The present invention has been made in view of such circumstances, and can reduce the amount of calculation when the encoding target area is divided into quadtrees and image encoding is performed for each obtained small area. An object is to provide an image encoding method, an image encoding device, and an image encoding program.

  The present invention is an image encoding method that divides an image into a plurality of rectangles, divides the rectangles into smaller regions and encodes them, and calculates a feature amount of a predetermined region of the image. Based on the feature amount calculated by the feature amount calculation step for the encoding target region, the encoding target region is divided by a quadtree, and each divided region obtained by the division is calculated by the feature amount calculation step. A quadtree partitioning step for recursively determining a divided region further divided by a quadtree based on the feature amount, and an encoding step for performing an encoding process for each of the determined divided regions. Features.

  In the feature amount calculating step, the complexity of the encoding target region is calculated as the feature amount in the feature amount calculating step, and the complexity of the encoding target region is preset in the quadtree dividing step. Comparison with a threshold value is performed, and quadtree division is performed on the encoding target region or the divided region whose complexity is higher than the threshold value.

  The present invention is characterized in that the complexity is calculated based on an average value and a variance value in the divided area.

  The present invention is an image encoding apparatus that divides an image into a plurality of rectangles, divides the rectangle into smaller regions and encodes the image, and includes a feature amount calculation unit that calculates a feature amount of a predetermined region of the image. Then, based on the feature amount calculated by the feature amount calculation unit for the encoding target region, the encoding target region is divided by a quadtree, and each divided region obtained by the division is calculated by the feature amount calculation unit. A quadtree dividing means that recursively further divides the divided area by a quadtree based on the feature amount, and an encoding means that performs an encoding process for each of the determined divided areas. Features.

  According to the present invention, a feature amount calculating step of calculating a feature amount of a predetermined region of the image is performed on a computer on an image encoding device that divides the image into a plurality of rectangles and divides the rectangle into smaller regions and encodes them. Then, based on the feature amount calculated by the feature amount calculation step for the encoding target region, the encoding target region is divided by a quadtree, and for each divided region obtained by the division, the feature amount calculation step Based on the calculated feature amount, a quadtree partitioning step for recursively determining a divided region further divided by a quadtree, and an encoding step for performing an encoding process for each of the determined divided regions are performed. It is characterized by that.

  According to the present invention, the CU size can be determined in advance without impairing the image quality based on the feature amount of the encoding target area of the original image. For this reason, when encoding is performed, the CU size can be determined without performing a prediction process, so that the amount of calculation related to encoding can be reduced.

It is a block diagram which shows the structure of one Embodiment of this invention. It is a flowchart which shows operation | movement of the image coding apparatus 1 shown in FIG. It is a flowchart which shows the detail of the processing operation of the image coding apparatus 1 shown in FIG. It is a schematic diagram of the division | segmentation of an encoding object area | region. It is explanatory drawing which shows the quadtree division | segmentation example of an encoding object area | region. It is a flowchart which shows the processing operation which divides the encoding object area | region by a prior art into quadtree. It is a figure which shows the relationship between the magnitude | size of the small rectangle by a prior art, and the amount of prediction processing calculations.

  Hereinafter, an image encoding device according to an embodiment of the present invention will be described with reference to the drawings. FIG. 1 is a block diagram showing the configuration of the embodiment. In the following description, it is assumed that the encoding method to be used is the HEVC intra-screen predictive encoding method. As shown in FIG. 1, the image encoding device 1 includes an intra-screen prediction encoding unit 2. Furthermore, the intra prediction encoding unit 2 includes an encoding process size determining unit 3 that determines an encoding process size before encoding. The encoding process size determination unit 3 calculates the complexity in the encoding target area from the original image before performing the encoding process, and encodes according to the calculated complexity. A comparison target of whether or not to divide a CU based on the result of the in-coding region complexity determination unit 5 that determines whether or not to divide the target region into quadtrees and the in-coding region complexity determination unit 5 Result of the quadtree split shape determining unit 7 for determining the quadtree split shape of the CU from the result of the threshold value setting unit 6 and the coding area complexity determination unit, and the result of the quadtree split shape determining unit 7 Includes an encoding processing unit 8 that performs intra-screen encoding.

  Next, the operation of the image encoding device shown in FIG. 1 will be described with reference to FIG. FIG. 2 is a flowchart showing the operation of the image encoding device shown in FIG. First, the in-coding region complexity calculation unit 4 calculates the degree of complexity of the encoding target region (step S1). The degree of complexity of the encoding target region is determined based on the average or variance of pixel values in the region. FIG. 4 is a schematic diagram of the division of the encoding target area. The in-coding region complexity calculation unit 4 divides the encoding target region having a size of 2N × 2N pixels into four sub-blocks of N × N shown in FIG. 4 and sCU_ {n} (n = 0.3).

  Next, the in-coding region complexity calculation unit 4 sets the pixel value in sCU_ {n} to I_ {n} (i, j), and within sCU_ {n} of I_ {n} (i, j). Is the average value of ave (I_ {n}). Subsequently, for the calculated four average values ave (I_ {n}), a variance var (ave (I_ {n})) thereof is calculated.

That is, var (ave (I_ {n}))
= Var (ave (I_ {0}), ave (I_ {1}), ave (I_ {2}), ave (I_ {3}))
= (Σ_ {n = 0} ^ {3} (ave (ave (I_ {0}), ave (I_ {1}), ave (I_ {2}), ave (I_ {3})) − ave ( I_ {n})) ^ 2) / (4-1)
Calculate Here, “^” represents a power operation.

  Next, the in-coding region complexity determination unit 5 is based on a comparison result between the complexity degree calculated by the in-coding region complexity calculation unit 4 and the threshold value TH preset by the threshold value setting unit 6. Then, it is determined whether or not the encoding target area is to be divided into quadtrees (step S2). As a result of this determination, the quadtree split shape determination unit 7 determines that the 2N × 2N encoding processing size is split into N × N if var (ave (I_n)) ≧ TH, and var (ave (ave ( If I_n)) <TH, it is determined that the encoding processing size remains 2N × 2N.

  Next, the encoding processing unit 8 performs encoding processing of the encoding target region based on the quadtree dividing shape of the encoding target region determined by the quadtree dividing shape determination unit 7 (step S3). Here, the encoding process is performed only once for each small area after the encoding target area is divided into quadtrees.

  Next, details of the processing operation of the image encoding device 1 shown in FIG. 1 will be described with reference to FIG. FIG. 3 is a flowchart showing details of the processing operation of the image encoding device 1 shown in FIG. Here, the size of the CU that is the encoding target area is 64 × 64 pixels, 32 × 32 pixels, 16 × 16 pixels, and 8 × 8 pixels, and the number of quadtree divisions in the encoding target area is 3. The operation will be described.

  First, the in-coding region complexity calculation unit 4 calculates the complexity by the above-described calculation method in the LCU size of 64 × 64 pixels (step S11). Subsequently, the in-coding region complexity determination unit 5 determines the complexity at CU = 64 × 64 based on the processing operation described above (step S12). And when dividing | segmenting CU, it progresses to the process of step S13, and when not dividing | segmenting CU, it progresses to the process of step S22.

  Next, the quadtree partition shape determination unit 7 divides the CU into blocks of 32 × 32 pixels to obtain four sub blocks (step S13). Subsequently, for each of the four divided sub-blocks (step S14), complexity calculation is performed again (step S15), and complexity determination is performed again (step S16).

  If the CU is divided as a result of the complexity determination in step S16, the process proceeds to step S17. If the CU is not divided, the process for the next sub-block is performed. When the processing of the four sub-blocks is completed, the process proceeds to step S22.

  Next, the 32 × 32 pixel CU is divided into 16 × 16 pixel blocks to obtain four sub-blocks (step S17). Subsequently, for each of the four divided sub-blocks (step S18), complexity calculation is performed again (step S19), and complexity determination is performed again (step S20).

  As a result of the complexity determination in step S20, if the CU is divided, the process proceeds to step S21. If the CU is not divided, the process for the next sub-block is performed. When the processing of the four sub-blocks is completed, the process returns to step S14.

  Next, the 16 × 16 pixel CU is divided into 8 × 8 pixel blocks to obtain four sub-blocks (step S21). Thereafter, the process returns to step S18.

  Through the above processing operation, the CU is divided from 64 × 64 pixels to 8 × 8 pixels, a combination of CU divisions is determined (step S22), and encoding is performed only with the obtained CU size (step S23). ).

  In the above description, the complexity in the region is used as the feature quantity, but other feature quantities may be calculated to determine the quadtree division shape.

  In this way, in the encoding method in which the encoding target area is divided into small areas by a quadtree and encoding is performed for each small area, an extra encoding process is performed by determining the CU size before encoding. As a result, the amount of computation required for the encoding process can be reduced. Further, the CU size can be determined without actually encoding the encoding target area.

  As described above, in the present embodiment, when encoding is performed, the CU size is set first, instead of narrowing down to one after encoding with all the CU sizes as in the conventional method. I made it. In order to realize this, a combination of a means for calculating the complexity of the encoding target area, a means for dividing the encoding target area based on the calculated complexity, and a combination of the CU sizes of the encoding target area is determined. And means for encoding a moving picture according to the determined CU size. As a result, when the video is divided into small rectangles and encoded, processing based on the complexity of the region to be encoded is used, thereby reducing the amount of computation required for the process of determining the divided shape into small rectangles. Can be realized.

  2 is recorded on a computer-readable recording medium, and the program recorded on the recording medium is read by a computer system and executed, thereby executing image coding processing. May be performed. Here, the “computer system” includes an OS and hardware such as peripheral devices. The “computer system” includes a WWW system having a homepage providing environment (or display environment). The “computer-readable recording medium” refers to a storage device such as a flexible medium, a magneto-optical disk, a portable medium such as a ROM and a CD-ROM, and a hard disk incorporated in a computer system. Further, the “computer-readable recording medium” refers to a volatile memory (RAM) in a computer system that becomes a server or a client when a program is transmitted via a network such as the Internet or a communication line such as a telephone line. In addition, those holding programs for a certain period of time are also included.

  The program may be transmitted from a computer system storing the program in a storage device or the like to another computer system via a transmission medium or by a transmission wave in the transmission medium. Here, the “transmission medium” for transmitting the program refers to a medium having a function of transmitting information, such as a network (communication network) such as the Internet or a communication line (communication line) such as a telephone line. The program may be for realizing a part of the functions described above. Furthermore, what can implement | achieve the function mentioned above in combination with the program already recorded on the computer system, what is called a difference file (difference program) may be sufficient.

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

  When encoding is performed, it is possible to apply to an application in which it is indispensable to reduce the amount of calculation by setting the CU size first, instead of narrowing down to one after encoding with all CU sizes.

  DESCRIPTION OF SYMBOLS 1 ... Image coding apparatus, 2 ... In-screen encoding part, 3 ... Encoding process size determination part, 4 ... Encoding area complexity calculation part, 5 ... Encoding area complexity determination part, 6 ... Threshold Value setting unit, 7 ... Quadtree division shape determining unit, 8 ... Coding processing unit

Claims (5)

An image encoding method that divides an image into a plurality of rectangles, divides the rectangles into smaller regions, and encodes them.
A feature amount calculating step for calculating a feature amount of a predetermined area of the image;
Based on the feature amount calculated by the feature amount calculation step for the encoding target region, the encoding target region is divided by a quadtree, and each of the divided regions obtained by the division is calculated by the feature amount calculation step. A quadtree dividing step for recursively determining a divided region further divided by a quadtree based on the feature amount;
An image encoding method comprising: an encoding step for performing an encoding process for each of the determined divided regions. In the feature amount calculating step, the complexity of the encoding target region is calculated as the feature amount,
In the quadtree partitioning step, the complexity of the encoding target region is compared with a preset threshold value, and the encoding target region or the split region whose complexity is higher than the threshold value The image encoding method according to claim 1, wherein quadtree partitioning is performed.   The image encoding method according to claim 2, wherein the complexity is calculated based on an average value and a variance value in the divided region. An image encoding device that divides an image into a plurality of rectangles, divides the rectangle into smaller regions, and encodes the image.
Feature amount calculating means for calculating a feature amount of a predetermined area of the image;
Based on the feature amount calculated by the feature amount calculation unit for the encoding target region, the encoding target region is divided by a quadtree, and each divided region obtained by the division is calculated by the feature amount calculation unit. A quadtree dividing means for recursively determining a divided region further divided by a quadtree based on the feature amount;
An image coding apparatus comprising: coding means for performing coding processing on each of the determined divided areas. A computer on an image encoding device that divides an image into a plurality of rectangles, divides the rectangles into smaller regions and encodes them,
A feature amount calculating step for calculating a feature amount of a predetermined area of the image;
Based on the feature amount calculated by the feature amount calculation step for the encoding target region, the encoding target region is divided by a quadtree, and each of the divided regions obtained by the division is calculated by the feature amount calculation step. A quadtree dividing step for recursively determining a divided region further divided by a quadtree based on the feature amount;
An image encoding program that performs an encoding step of performing an encoding process on each of the determined divided areas.

Images