JP2013090330A - Image processing method and device, using image division - Google Patents

Abstract

PROBLEM TO BE SOLVED: To provide an image processing method and a device using image division, capable of further enhancing compression efficiency, with suppressed deterioration of image quality caused by a transmission line error to the minimum, when compressing and decompressing an image signal using a multiplicity of processors.SOLUTION: The image processing method using image division includes the steps of: sectioning an image frame into a multiplicity of sub-blocks having a constant pixel size; configuring a multiplicity of macro blocks by sampling pixels at an identical position in sub-blocks of a constant number; and configuring a multiplicity of slices by sectioning the multiplicity of macro blocks according to the position in the sub-block of the sampled pixel.

Description

  The present invention relates to an image processing method and apparatus using image division, and more particularly to a technique for dividing an image frame in order to process a high-resolution image in parallel using a multiprocessor.

  The screen resolution of mobile devices such as mobile phones, PMPs, and PDAs is gradually increasing. Recently, with the advent of smartphones and tablet PCs, it has become possible to provide high-quality moving image services even on mobile devices. As digital television becomes widespread, the screen resolution is 1920 × 1080 or more, that is, a full HD grade home television has become common, thereby increasing the frame rate. Furthermore, UHDTV (Ultra HDTV) having a 4K or 8K resolution (for example, 7680 × 4320) in the horizontal direction of the screen has attracted attention as a next-generation television.

  An image compression technique that can increase the compression efficiency of an image as the resolution of a display device increases has been developed and used. Note that the amount of computation for image compression and decompression increases as the screen size increases and the frame rate increases.

  Also, as one of the measures to reduce power consumption while processing a large amount of computation, a multi-core processor (Multicore processor) with multiple CPUs on one chip has been developed and widely used in computers and notebook PCs. And has been quickly applied to smartphones and tablet PCs.

  In general, in the conventional image compression method, in order to efficiently compress image data, the original image is divided into a plurality of blocks having a certain size, and one block includes pixels adjacent on the screen. become. As an example, H.C. In the case of the H.264 / AVC standard, an image is divided and processed in units of macroblocks, and the macroblocks are composed of one 16 × 16 luminance (Luminance) sample block and two chrominance (Chrominance) sample blocks. The sample size varies depending on the format of the input image). FIG. 1 is a diagram showing a method of dividing an image into 16 × 16 size macroblocks according to the prior art. As described above, the entire screen is divided into macro blocks having a uniform size, and encoding (encoding) or decoding (decoding) is performed in units of macro blocks.

  In addition, in the conventional image compression standard, in order to prevent degradation of the decoded image due to a transmission path error, an error generated in a certain slice is obtained by using a method of dividing the original image into a plurality of slices and encoding. Will not propagate to other slices. For example, as shown in FIG. 2a, the MPEG-4 standard uses a method of configuring a slice by grouping a plurality of macroblocks in raster scan order. H. In the H.264 / AVC standard, a slice is configured in an arbitrary order regardless of the scan order as shown in FIG. 2b, and consecutive macroblocks are alternately assigned to different slice groups as shown in FIG. 2c. As shown in FIG. 2d, slices can be configured using various methods such as assigning macroblocks belonging to an arbitrary rectangular area to different slice groups.

  Note that the above-described image segmentation method into slices or slice groups can be used to reduce the effects of errors during image transmission and decoding, and many image encoding or decoding processes can be performed. It can also be used to do this using a processor. FIG. 3 is a diagram illustrating a method in which one image is divided into a plurality of slices, which are simultaneously encoded by a plurality of image encoding processors, and then combined as one bit stream.

  However, in the conventional method as described above, since compression information of adjacent slices cannot be used, there is a disadvantage in that the compression efficiency is reduced as compared with a case where the entire screen is encoded as a single slice. When a transmission path error occurs, there is a problem that image quality deterioration occurs in a continuous constant area of the decoded image. Further, when consecutive macroblocks are assigned to different slice groups by the method shown in FIG. 2c or a similar method, image quality degradation due to errors can be distributed over the entire screen, but the compression efficiency further decreases. There is a problem.

  The present invention has been devised to solve the above-described problems, and the object of the present invention is due to transmission path errors when compressing and decompressing image signals using a large number of processors. An object of the present invention is to provide an image processing method and apparatus using image division that can minimize image quality deterioration and further increase the compression efficiency.

  In order to achieve the above object, an image processing method using image division according to an embodiment of the present invention includes a step of dividing an image frame into a number of sub-blocks having a certain pixel size, a certain number of sub-blocks. Sampling a plurality of pixels at the same position in the block to form a plurality of macroblocks, and partitioning the plurality of macroblocks according to positions in the sub-blocks of the sampled pixels to form a plurality of slices. Comprising configuring steps.

  In addition, the method may further include: encoding the multiple slices in parallel using multiple image encoding processors; and generating a bitstream by combining the encoded data.

  An image processing method using image division according to another embodiment of the present invention includes a step of dividing an image frame into a plurality of sub-blocks having a certain pixel size, and an average value of each pixel in the plurality of sub-blocks. And calculating and forming one average slice with pixels having said average value.

  In addition, sampling a pixel at the same position in a certain number of sub-blocks to form a number of macroblocks, and partitioning the number of macroblocks according to the positions of the sampled pixels in the sub-blocks A plurality of slices are configured, and a difference value between a pixel included in each of the plurality of slices and a pixel included in the average slice is calculated to configure the plurality of slices from the difference value. The method may further include the step of:

Furthermore, using a plurality of image encoding processors, the step of encoding the average slice and the plurality of slices composed of the difference values in parallel, and combining the encoded data into a bitstream A step of generating may further be included.
In the encoding step, information regarding the position in the sub-block of the pixels constituting each slice can be written in the slice header.

  An image processing apparatus using image division according to an embodiment of the present invention divides an image frame into a number of sub-blocks having a certain pixel size, and samples pixels at the same position in a certain number of sub-blocks. A plurality of macroblocks, and an image dividing unit that configures a plurality of slices by dividing the plurality of macroblocks according to positions in the sub-blocks of the sampled pixels, and the plurality of slices in parallel. A multiprocessor including a number of image encoding processors for encoding, and a data combining unit that combines the encoded data to generate a bitstream.

The image dividing unit may calculate an average value of each pixel in the plurality of sub-blocks, and may further constitute one average value slice with the pixels having the average value, and is included in each of the plurality of slices. The plurality of slices may be constructed from the difference values by calculating a difference value between the pixels obtained and the pixels included in the average value slice.
The multiple image encoding processors may encode the average slice and the multiple slices configured from the difference values in parallel.

According to the present invention, even when an error occurs in some slices in the data transmission process by dividing and encoding the image so that adjacent pixels are included in different slices, respectively, in the data transmission process. Thus, image quality degradation due to errors is distributed over the entire screen, and it is possible to significantly reduce the noticeable degradation of image quality due to peripheral pixels transmitted without error.
In addition, the compression efficiency can be improved as compared with the case where the same number of slices are divided by the conventional technique.
Furthermore, even when only one of a large number of slices is combined, all the reduced images can be obtained, and further, there is an effect that spatial scalability can be efficiently obtained.

It is a figure which shows the image division method by a prior art. FIG. 3 shows various methods for dividing an original image into a number of slices. FIG. 3 shows various methods for dividing an original image into a number of slices. FIG. 3 shows various methods for dividing an original image into a number of slices. FIG. 3 shows various methods for dividing an original image into a number of slices. It is a figure which shows the image coding method using many processors. 6 is a flowchart illustrating an image processing method using image division according to an embodiment of the present invention. It is a figure for demonstrating in detail the image division | segmentation method which concerns on embodiment of FIG. It is a flowchart which shows the image processing method using the image division which concerns on other embodiment of this invention. 1 is a configuration diagram of an image processing apparatus using image division according to an embodiment of the present invention.

The above-described objects, features, and advantages will be described in detail with reference to the accompanying drawings. As a result, a person having ordinary knowledge in the technical field to which the present invention belongs can easily implement the technical idea of the present invention. In the description of the present invention, if it is determined that there is a possibility that a specific description of a related known technique may unnecessarily disturb the gist of the present invention, a detailed description thereof will be omitted.
Hereinafter, preferred embodiments of the present invention will be described in detail with reference to the accompanying drawings.

  FIG. 4 is a flowchart showing an image processing method using image division according to an embodiment of the present invention, and FIG. 5 is a diagram for explaining the image division method according to the embodiment of FIG. 4 in detail. .

  As shown in FIGS. 4 and 5, the image processing method using image division according to an embodiment of the present invention includes a step of dividing an image frame into a plurality of sub-blocks having a certain pixel size (S401). A plurality of macroblocks are sampled by sampling pixels at the same position in a certain number of sub-blocks (S403), and the plurality of macroblocks are divided according to the positions of the sampled pixels in the sub-blocks. A step of constructing a large number of slices (S405). In addition, the method may further include a step of encoding a large number of slices in parallel using a large number of image encoding processors (S407) and a step of generating a bitstream by combining the encoded data (S409).

  In the present embodiment, it is assumed that one microphone block has a 16 × 16 pixel size. However, it goes without saying that the size of such a macroblock changes depending on the format and resolution of the image.

  In step S401, the image is divided such that M × N pixels of the original image form one sub-block. In this embodiment, M = N = 2, and four adjacent pixel positions in one subpixel are indicated as a, b, c, and d, respectively, as shown in FIG.

  In step S403, pixels at the same position in 16 × 16 sub-blocks are sampled to form a large number of macroblocks. That is, when M: 1 is sampled in the horizontal direction and N: 1 is sampled in the vertical direction, pixels are selected every other pixel in the horizontal and vertical directions from the condition of M = N = 2. Pixels at adjacent positions a, b, c, and d are separated from each other, and are composed of different macroblocks. In this way, the pixels at position a are collected to form macroblocks a0, a1,..., And the pixels at positions b, c, d are collected and macroblocks b0, b1,. .., c0, c1,..., D0, d1,.

  In step S405, a slice is composed of a large number of macroblocks composed of pixels at the same position. That is, the macroblocks a0, a1,... Are collected to configure the slice A, and the slices B, C, and D are configured in the same manner.

  Next, in step S407, slices A, B, C, and D are simultaneously encoded in parallel using a number of image encoding processors, and the data encoded in step S409 is combined to generate one bit stream. Each slice can be transmitted jointly without being constrained in order.

  At this time, information about whether a pixel included in each slice is a pixel corresponding to a certain position in the sub-block in the encoding process is written in the slice header so that the bitstream can be accurately decoded. Can do. Alternatively, information necessary for the slice header can be removed by predefining the values of M and N and predefining the slice transmission order. In the subsequent decoding process, the decoding result of each slice is recombined so that the pixels included in each slice can be displayed at the correct position on the screen according to the information written as described above or in a predefined order. be able to.

  According to the present invention, spatial scalability can be effectively supported with one bit stream. That is, according to the above-described embodiment, each slice has equal spatial information about the image, and even if only one of the slices A, B, C, and D is combined, the entire reduced size is obtained. A screen can be obtained.

  FIG. 6 is a flowchart illustrating an image processing method using image division according to another embodiment of the present invention.

  As shown in FIG. 6, the image processing method according to another embodiment of the present invention includes a step of dividing an image frame into a plurality of sub-blocks having a certain pixel size (S601), A step of calculating an average value of pixels (S603), a step of forming one average value slice with pixels having the calculated average value (S605), and sampling pixels at the same position in a certain number of sub-blocks. Forming a plurality of macroblocks (S607), and dividing the plurality of macroblocks according to the positions in the sub-blocks of the sampled pixels to form a plurality of slices, which are included in each of the plurality of slices. The difference between the averaged slice and the pixels included in the average slice Comprising the step (S609) that constitutes the slice from the difference value. In addition, a step of encoding the average value slice and a number of slices composed of the difference values in parallel using a number of image encoding processors (S611), and the encoded data are combined to generate a bitstream. Step (S613) may be further included.

  This embodiment is an application of the spatial scalability further developed in the embodiments of FIGS. 4 and 5 and having an average value of pixels included in each sub-block as shown in FIG. Another average slice of pixels is generated, and the remaining slices can be composed of the difference values from the pixels included in the average slice to increase the coding efficiency. In this case, only the average value slice can be combined in the decoding process to obtain a small-sized decompressed image, and the remaining slices can be further combined to obtain an image of the same size as the original image. It becomes.

  FIG. 7 is a block diagram showing an image processing apparatus using image division according to an embodiment of the present invention.

  As shown in FIG. 7, the image processing apparatus using image division according to an embodiment of the present invention includes an image dividing unit 701, a multiprocessor 703, and a data combining unit 705. The image dividing unit 701 divides an image frame into a large number of sub-blocks having a certain pixel size, samples pixels at the same position in a certain number of sub-blocks to form a large number of macro-blocks, A macroblock can be partitioned according to the position within a sub-block of sampled pixels to form multiple slices.

  In addition, the image dividing unit 701 can calculate an average value of each pixel in a number of sub-blocks, and can further configure one average value slice with pixels having the calculated average value. A plurality of slices may be configured from the calculated difference values by calculating a difference value between the pixels included in each of the slices and the pixels included in the average value slice.

  The multiprocessor 703 includes multiple image encoding processors P_A, P_B,..., P_X for encoding multiple slices in parallel at the same time. A number of image encoding processors P_A, P_B,..., P_X can write information on the positions in the sub-blocks of the pixels constituting each slice in the slice header at the time of encoding. An example of such a multiprocessor 703 is a multicore processor in which a large number of cores are mounted on one chip and different operations can be performed in parallel.

  The data combining unit 705 generates a single bit stream by combining the data encoded in parallel. Note that the parallel data can be combined in the order determined by the image standard or the like, or in any order.

  More detailed functions and effects of each configuration are the same as those described with reference to FIGS.

  The technical idea of the present invention has been specifically described with the above-described preferred embodiments. However, the above-described embodiments are for explanation, and do not limit the present invention. Further, those skilled in the art of the present invention can implement various modifications within the scope of the technical idea of the present invention.

Claims (9)

Partitioning an image frame into a number of sub-blocks having a constant pixel size;
Sampling a plurality of pixels at the same position in a certain number of sub-blocks to form a plurality of macro-blocks; and partitioning the plurality of macro-blocks according to the positions of the sampled pixels in the sub-blocks And constructing multiple slices;
An image processing method using image division. Encoding the multiple slices in parallel using multiple image encoding processors; and combining the encoded data to generate a bitstream;
The image processing method using image division according to claim 1, further comprising:   The image processing method using image division according to claim 1 or 2, wherein, in the encoding step, information on a position in a sub-block of a pixel constituting each slice is written in a slice header. Partitioning an image frame into a number of sub-blocks having a constant pixel size;
Calculating an average value of each pixel in the multiple sub-blocks;
An image processing method using image segmentation, comprising: forming one average value slice with pixels having the average value; and encoding the average value slice to generate a bitstream. Sampling a plurality of pixels at the same position in a certain number of sub-blocks to form a plurality of macro-blocks; and partitioning the plurality of macro-blocks according to the positions of the sampled pixels in the sub-blocks; Forming a plurality of slices, calculating a difference value between a pixel included in each of the plurality of slices and a pixel included in the average slice, and forming the plurality of slices from the difference value. ;
The image processing method using image division according to claim 4, further comprising: Encoding the average slice and the multiple slices composed of the difference values in parallel using multiple image encoding processors; and combining the encoded data to generate a bitstream Step;
The image processing method using image division according to claim 5, further comprising: The image frame is divided into a large number of sub-blocks having a certain pixel size, and a plurality of macro-blocks are formed by sampling pixels at the same position in a certain number of sub-blocks, and the large number of macro-blocks are sampled. An image dividing unit that forms a large number of slices by dividing the pixel according to the position in the sub-block;
A multiprocessor including a plurality of image encoding processors for encoding the plurality of slices in parallel; and a data combining unit that combines the encoded data to generate a bitstream;
An image processing apparatus using image division.   The image according to claim 7, wherein the image dividing unit calculates an average value of each pixel in the plurality of sub-blocks, and further constitutes one average value slice with pixels having the average value. An image processing apparatus using division.   The image dividing unit calculates a difference value between a pixel included in each of the plurality of slices and a pixel included in the average value slice, and configures the plurality of slices from the difference value. An image processing apparatus using image division according to claim 8.

Images