JP2020058075A - Moving image prediction encoding device, moving image prediction decoding device, moving image prediction encoding method, moving image prediction decoding method, and recording medium - Google Patents

Abstract

To efficiently use an image buffer memory to allow coding efficiency to be improved further.SOLUTION: A moving image prediction decoding method according to the present invention obtains an object image frame size indicating the frame size of an object image, MaxLumaFS indicating a maximum size of a reproduction image that can be stored in a decoded image buffer, and MFSBuffer expressed by the maximum number of reproduction images that can be stored in the decoded image buffer, and if it is determined that a half of MaxLumaFS is equal to or more than the object image frame size, doubles MFSBuffer, and sets maxDPBsize, which is an adaptive maximum number of reproduction images that can be stored in the decoded image buffer, to twice MFSBuffer. The number of reproduction images stored in the decoded image buffer is equal to or less than the number set in maxDPBsize.SELECTED DRAWING: Figure 5

Description

  The present invention relates to a video prediction encoding device, a video prediction decoding device, a video prediction encoding method, a video prediction decoding method, and a recording medium, and in particular, the maximum number of reproduced images that can be stored in a decoded image buffer. The present invention relates to a video prediction encoding device, a video prediction decoding device, a video prediction encoding method, a video prediction decoding method, and a recording medium that manages a video.

  In order to efficiently transmit and store moving image data, a compression encoding technique is used. In the case of a moving image, MPEG1-4 or H.264. 261-H. The H.264 system is widely used.

  In these encoding methods, an image to be encoded is divided into a plurality of blocks before encoding and decoding processes are performed. The following predictive coding method is used to enhance the coding efficiency. In predictive coding in a screen, a predicted signal is generated using an image of a previously reproduced image (decompressed image data compressed in the past) adjacent to the target block in the same screen as the target block. The difference signal subtracted from the block signal is encoded. In predictive coding between screens, a previously reproduced image signal in a screen different from the target block is referred to, a displacement of the signal is searched, a moving signal is compensated for, and a predicted signal is generated. The difference signal obtained by subtracting from the signal is encoded. A previously reproduced image that is referred to for performing motion search / compensation is called a reference image.

  Further, in the bidirectional inter-screen prediction, not only images in the past in the order of display time but also future images displayed after the target image may be referred to in some cases. And it needs to be reproduced in advance). By averaging the predicted signals obtained from past and future images, it is effective in predicting the signal of hidden and newly appearing objects, while reducing the noise contained in both predicted signals Has the effect of doing

  Furthermore, H. In the H.264 inter-picture prediction coding, a prediction signal for a target block refers to a plurality of reference images that have been coded and reproduced in the past, and uses the image signal with the least error as an optimal prediction signal while performing motion search. select. The difference between the pixel signal of the target block and the optimal prediction signal is obtained, subjected to discrete cosine transform, quantized, and entropy-coded. At the same time, it encodes information (referred to as a reference index and a motion vector, respectively) regarding which region of a reference image obtains an optimal prediction signal for the target block.

  By the way, H. In H.264, it is possible to refer to a plurality of reproduced images. These reproduced images are stored in a decoded image buffer (DPB: Decoded picture buffer) as an image buffer memory as a reference image used for prediction. The size of the decoded image buffer (DPB) is defined by a profile and a level, and is defined not as the number of reference images but as a bit amount. That is, the number of images changes according to the frame size of the image even with the same profile and level. For example, in the case of a main (Main) profile at level 3.2, the maximum size (MaxDPBSize: Maximum Decoded Picture Buffer size) of an image buffer for storing a reference image used for prediction is defined as 7680.0 × 1024 [bytes]. In the case of 1280 × 720 4: 2: 0, 5 playback images can be stored in the decoded image buffer (DPB). In the case of 1280 × 1024 4: 2: 0, the playback image can be stored in the decoded image buffer (DPB). Is four. FIG. 1 shows decoding when the maximum number of reproduced images that can be stored according to the frame size of the image is four (FIG. 1A), five (FIG. 1B), and six (FIG. 1C). 4 shows an image arrangement of an image buffer. As described above, by arranging the memory pointer variably in the image buffer memory prepared in advance according to the frame size of the reproduced image, the memory buffer can be used while devising the memory arrangement.

“H.264: Advanced video coding for generic audiovisual services,” Joint Video Team of ITU-T VCEGand ISO / IEC MPEG, ITU-T Rec.H.264 and ISO / IEC 14496-10 (MPEG4-Part 10), November 2007

  Here, in such a definition, H. Since the H.264 decoding device needs to be able to decode an image of any frame size defined by a profile or level in a range supported by the decoding device, a memory used to store a reproduced image in an image buffer memory It is not possible to fix how to place and store. Therefore, it is necessary to change the memory pointer indicating the memory location where each reproduced image is stored according to the frame size of the image to be decoded, which complicates memory control and implementation.

  As a solution to this, a method of fixing the maximum number of reproduced images (max_dec_pic_buffering) that can be stored in the image buffer memory regardless of the actual frame size of the decoded image can be considered. With this solution, the memory arrangement of the decoded image buffer (DPB) is uniquely determined by the maximum frame size as shown in FIG. 2A, so that the memory pointer can be fixed. That is, the memory control becomes easier than the control of the variable memory pointer. However, in this solution, when the frame size of the decoded image is smaller than the maximum frame size, or when the vertical width of the image is half of the frame image like a field image of an interlaced image, FIG. As shown in (1), there is a problem that an unused memory area is generated and the memory is not effectively used. Furthermore, storing more reproduced images loses the possibility of increasing the number of reference images and improving the coding efficiency.

  The present invention solves the above-described problems, and determines the maximum number of reproduced images that can be stored in the image buffer memory according to the frame size of the reproduced image. It is an object of the present invention to provide a moving picture prediction coding apparatus, a moving picture prediction decoding apparatus, a moving picture prediction coding method, a moving picture prediction decoding method, and a recording medium which can also improve the coding efficiency.

  In order to achieve the above object, a video predictive encoding apparatus according to the present invention includes a CPU that is programmable to perform video predictive encoding, and a plurality of CPUs that configure a video by the programmed CPU. A receiving unit that is executable to receive an image and the programmed CPU encodes the received image in one of an intra-screen prediction and an inter-screen prediction to generate compressed image data. An executable encoding unit, a decoding unit capable of decoding the compressed image data by the programmed CPU, and decoding the compressed image data into a reproduced image, and one or more reproduced images by the programmed CPU. A frame memory executable to be stored as a reference image used to encode a subsequent image; and A memory management unit executable to manage the frame memory, comprising: a target image frame size indicating a frame size of the target image; and a reproduced image pre-defined by level information and storable in the frame memory. And a frame memory size (MFSBuffer) represented by a preset maximum number of reproduced images that can be stored in the frame memory. A memory management unit, wherein the memory management unit compares Ｌ times MaxLumaFS with the target image frame size, and determines that 倍 times MaxLumaFS is equal to or larger than the target image frame size. The MFS Buffer is doubled and can be stored in the frame memory. The maximum number of reproduced images (maxDPBsize) may be set to twice the MFSBuffer, and the number of reproduced images stored in the frame memory is equal to or less than the number set in maxDPBsize. .

  In the moving picture prediction encoding apparatus according to the present invention, the programmed CPU causes the memory management unit to compare １ ／ times MaxLumaFS with a target image frame size, and determine that １ ／ times MaxLumaFS is When it is determined that the size is equal to or larger than the target image frame size, the MFSBuffer may be quadrupled, and the maxDPBsize may be set to four times the MFSBuffer.

  In order to achieve the above object, a video predictive decoding device according to the present invention includes a CPU programmed to perform video predictive decoding, and a plurality of compressed images constituting a video being compressed by the CPU. A receiving unit that can be executed to receive encoded data including compressed image data representing a form, wherein a plurality of images are encoded by any of intra-screen prediction and inter-screen prediction. A receiving unit, a decoding unit executable by the CPU to decode the compressed image data, and decompress the decoded image data into a reproduced image; and the CPU uses one or more reproduced images to decode a subsequent image. A decoded picture buffer (DPB: Decoded Picture Buffer) executable to be stored as a reference picture to be stored, and the CPU manages the DPB. And a maximum frame size (predetermined by the level information and indicating the maximum size of the reproduced image that can be stored in the DPB). MaxLumaFS) and a buffer management unit that can be executed to acquire a frame memory size (MFSBuffer) represented by a preset maximum number of reproduction images that can be stored in the DPB. The management unit compares 1/2 of MaxLumaFS with the target image frame size, and when it is determined that 1/2 of MaxLumaFS is equal to or larger than the target image frame size, doubles the MFSBuffer and doubles the MFSBuffer in the DPB. The maximum number of playback images (maxDPBsize) that can be stored in Set to twice the FSBuffer, as is further executable, the number of stored reproduced image in the DPB is a number less set in MaxDPBsize.

  Further, in the moving picture prediction decoding apparatus according to the present invention, the buffer management unit compares 1/4 times MaxLumaFS with the target image frame size, and determines that 1/4 times MaxLumaFS is equal to or larger than the target image frame size. If it is determined, the MFSBuffer may be quadrupled, and the maxDPBsize may be set to be four times MFSBuffer.

  In order to achieve the above object, a video prediction encoding method according to the present invention is a video prediction encoding method executed by a video prediction encoding apparatus, and receives a plurality of images constituting a video. And the received image is encoded by any method of intra-screen prediction or inter-screen prediction, a step of generating compressed image data, a step of decoding the compressed image data, and restoring a reproduced image, Storing, in the frame memory, one or more reproduced images as a reference image used for encoding a subsequent image, a target image frame size indicating a frame size of the target image, and level information in advance. A maximum frame size (MaxLumaFS) indicating the maximum size of a reproduced image that can be stored in the frame memory; Acquiring a frame memory size (MFSBuffer) represented by a preset maximum number of reproduced images that can be stored in the memory, comparing the target image frame size with 1/2 times MaxLumaFS, When it is determined that 1/2 times MaxLumaFS is equal to or larger than the target image frame size, the step of doubling the MFSBuffer and the adaptive maximum number (maxDPBsize) of the reproduced images that can be stored in the frame memory are set to the MFSBuffer. Setting the number of reproduction images stored in the frame memory to be less than or equal to the number set in maxDPBsize.

  Further, in the moving picture prediction encoding method according to the present invention, the step of comparing 1/4 times MaxLumaFS with the target image frame size, and determining that 1/4 times MaxLumaFS is equal to or larger than the target image frame size. In this case, the method may further include a step of quadrupling the MFSBuffer and a step of setting maxDPBsize to four times the MFSBuffer.

  In order to achieve the above object, a moving picture prediction decoding method according to the present invention is a moving picture prediction decoding method executed by a moving picture prediction decoding apparatus, wherein a compressed form of a plurality of pictures constituting a moving picture is provided. Receiving encoded data including compressed image data representing a plurality of images, wherein the plurality of images are encoded by any of intra-screen prediction or inter-screen prediction, and Decoding the compressed form and restoring it to a reproduced image; and storing, in a decoded image buffer (DPB: Decoded Picture Buffer), one or more reproduced images as reference images used for decoding subsequent images. A step, a target image frame size indicating a frame size of the target image, and level information, which are defined in advance in the DPB. Obtaining a maximum frame size (MaxLumaFS) indicating the maximum possible size of a reproduced image and a frame memory size (MFSBuffer) represented by a preset maximum number of reproduced images that can be stored in the DPB; Comparing the half of MaxLumaFS with the target image frame size, and doubling the MFSBuffer when it is determined that the half of MaxLumaFS is equal to or larger than the target image frame size, Setting the adaptive maximum number of reproduced images (maxDPBsize) that can be stored in the DPB to twice the MFSBuffer, wherein the number of reproduced images stored in the DPB is less than or equal to the number set in maxDPBsize. is there.

  Further, in the moving picture prediction decoding method according to the present invention, the step of comparing 1/4 times MaxLumaFS with the target image frame size, and the case where it is determined that 1/4 times MaxLumaFS is equal to or larger than the target image frame size. In addition, the method may further include a step of quadrupling MFSBuffer and a step of setting maxDPBsize to four times MFSBuffer.

  In addition, the moving picture prediction decoding method according to the present invention further includes a step of setting maxDPBsize to twice MFSBuffer and then keeping a memory pointer indicating the storage position of the DPB divided according to MFSBuffer. It may be that.

  In addition, the moving picture prediction decoding method according to the present invention further includes a step of setting maxDPBsize to four times MFSBuffer and then keeping a memory pointer indicating the storage position of the DPB divided according to MFSBuffer. It may be that.

  In order to achieve the above object, a recording medium according to the present invention is a recording medium for storing a moving image prediction decoding program, wherein the moving image prediction decoding program instructs a CPU to compress a plurality of images constituting a moving image. A receiving unit that is capable of receiving encoded data including compressed image data representing a compressed form, wherein a plurality of images are encoded by any of intra-screen prediction and inter-screen prediction. A receiving unit, a decoding unit executable to decode the compressed image data and restore it to a reproduced image, and store one or more reproduced images as reference images used to decode subsequent images A decoded picture buffer (DPB: Decoded Picture Buffer) that can be executed as described above and a buffer management unit that can be executed so as to manage the DPB. The maximum frame size (MaxLumaFS), which is defined in advance by the target image frame size indicating the program size and the level information and indicates the maximum size of the reproduced image that can be stored in the DPB, A buffer management unit that is executable to obtain a frame memory size (MFSBuffer) represented by the set maximum number, and the buffer management unit is configured to execute Ｌ times MaxLumaFS. Is compared with the target image frame size, and when it is determined that 1/2 times MaxLumaFS is equal to or larger than the target image frame size, the MFSBuffer is doubled, and the adaptive maximum number of reproduced images that can be stored in the DPB (MaxDPBsize) is set to twice MFSBuffer , So that a further executable, the number of stored reproduced image in the DPB is a number less set in MaxDPBsize.

  ADVANTAGE OF THE INVENTION According to the moving image prediction encoding apparatus, the moving image prediction decoding apparatus, the moving image prediction encoding method, the moving image prediction decoding method, and the recording medium according to the present invention, it can be stored in the image buffer memory according to the frame size of the reproduced image. By determining the maximum number of reproduced images, the image buffer memory can be efficiently used, and the encoding efficiency can be further improved.

FIG. 3 is a diagram illustrating an image buffer memory, the number of stored images, and a variable memory pointer position. FIG. 4 is a diagram illustrating a problem of a fixed memory pointer in an image buffer memory. 1 is a block diagram illustrating a moving picture prediction encoding device according to an embodiment of the present invention. FIG. 1 is a block diagram illustrating a video prediction decoding device according to an embodiment of the present invention. 5 is a flowchart illustrating a first moving picture prediction encoding / decoding method according to an embodiment of the present invention. FIG. 4 is a schematic diagram for explaining processing of a first moving picture prediction encoding / decoding method according to an embodiment of the present invention. FIG. 2 is a diagram illustrating a hardware configuration of a computer for executing a program recorded on a recording medium. FIG. 11 is a perspective view of a computer for executing a program stored in a recording medium.

  Hereinafter, embodiments of the present invention will be described with reference to FIGS. 3 to 8.

  FIG. 3 is a block diagram showing a moving picture predictive coding apparatus according to an embodiment of the present invention. 301 is an input terminal, 302 is a block divider, 303 is a prediction signal generator, 304 is a frame memory, 305 is a subtractor, 306 is a transformer, 307 is a quantizer, 308 is an inverse quantizer, and 309 is an inverse transform. 310, an adder, 311 an entropy coder, 312 an output terminal, and 314 a frame memory manager (also referred to as a buffer manager). The input terminal 301 corresponds to an input unit. The prediction signal generator 303, the subtractor 305, the transformer 306, the quantizer 307, and the entropy encoder 311 correspond to an encoding unit. The inverse quantizer 308, the inverse transformer 309, and the adder 310 correspond to decoding means. The frame memory 304 corresponds to an image storage unit. The frame memory manager 314 corresponds to a memory control unit.

  The operation of the moving picture predictive coding apparatus configured as described above will be described below. A signal of a moving image including a plurality of images is input to the input terminal 301. An image to be encoded is divided into a plurality of regions by a block divider 302. In the embodiment according to the present invention, the image data is divided into blocks each including 8 × 8 pixels. However, the image data may be divided into other block sizes or shapes. Next, a prediction signal is generated for a region to be subjected to the encoding process (hereinafter, referred to as a target block). In the embodiment according to the present invention, two types of prediction methods are used. That is, inter-screen prediction and intra-screen prediction.

  In the inter prediction, a reproduced image that has been coded in the past and restored is used as a reference image, and motion information that gives a prediction signal with the smallest error with respect to the target block is obtained from the reference image. This process is called motion detection. Further, as occasion demands, the target block may be subdivided, and an inter-screen prediction method may be determined for the subdivided small area. In this case, among the various division methods, the most efficient division method for the entire target block and the respective motion information are determined. In the embodiment according to the present invention, the processing is performed by the prediction signal generator 303, and the target block is input via the line L302 and the reference image is input via the L304. As the reference image, a plurality of previously encoded and restored images are used as reference images. For details, see MPEG-2, 4, H.264, which is a conventional technique. H.264. The motion information and the small area dividing method determined in this way are sent to the entropy encoder 311 via the line L312, encoded, and then transmitted from the output terminal 312. Information (reference index) from which reference image the prediction signal is obtained from among the plurality of reference images is also sent to the entropy encoder 311 via the line L312. In the embodiment according to the present invention, M (M is an integer of 1 or more) reproduced images are stored in the frame memory 304 and used as reference images. The prediction signal generator 303 acquires a reference image signal from the frame memory 304 based on the method of dividing the small area and the reference image and the motion information corresponding to each small area, and generates a prediction signal. The generated inter-picture prediction signal is sent to the subtractor 305 via the line L303.

  In intra prediction, an intra prediction signal is generated using previously reproduced pixel values spatially adjacent to the target block. Specifically, the prediction signal generator 303 obtains already reproduced pixel signals in the same screen from the frame memory 304 and extrapolates these signals to generate an intra-screen prediction signal. Information on the extrapolation method is sent to an entropy encoder 311 via a line L312, encoded, and then sent from an output terminal 312. The intra prediction signal generated in this way is sent to the subtractor 305. A method for generating a prediction signal in a screen in the prediction signal generator 303 is a conventional technique, which is described in H.264. H.264. For the inter-picture prediction signal and the intra-picture prediction signal obtained as described above, the one with the smallest error is selected and sent to the subtractor 305.

  As for the first image, since there is no image before that, all target blocks are processed by intra prediction.

  The subtractor 305 subtracts the prediction signal (via the line L303) from the signal of the target block (via the line L302) to generate a residual signal. This residual signal is subjected to a discrete cosine transform by a transformer 306, and each coefficient thereof is quantized by a quantizer 307. Finally, the transform coefficients quantized by the entropy encoder 311 are encoded, and transmitted from the output terminal 312 together with information on the prediction method.

  In order to perform intra prediction or inter prediction for the subsequent target block, the signal of the compressed target block is inversely processed and restored. That is, the quantized transform coefficient is inversely quantized by the inverse quantizer 308 and then inverse discrete cosine transformed by the inverse transformer 309 to restore the residual signal. The residual signal restored by the adder 310 and the prediction signal sent from the line L303 are added, a signal of the target block is reproduced, and stored in the frame memory 304. Although the converter 306 and the inverse converter 309 are used in the present embodiment, other conversion processing may be used instead of these converters. In some cases, the converter 306 and the inverse converter 309 may not be provided.

  The frame memory 304 is finite and cannot store all reproduced images. Only the reproduced image used for encoding the subsequent image is stored in the frame memory 304. The frame memory 304 is controlled by a frame memory manager 314. The frame memory manager 314 controls so that the oldest one is deleted from the M (here, M is an integer) reproduced images in the frame memory 304 and the latest reproduced image used as a reference image can be stored. I do. The frame size of each image and the maximum frame size defined by the encoding settings (profile, level) are input from the input terminal 313, and the maximum number of reproduced images that can be stored in the frame memory is determined based on this information. , The frame memory manager 314 operates so as not to exceed the maximum number. At the same time, information on the frame size of each image, the maximum frame size defined by the encoding setting, and the maximum number of reproduced images that can be stored in the frame memory are sent to the entropy encoder 311 via the line L314, and are encoded. The image data is output together with the compressed image data. The maximum frame size defined by the frame size and the encoding setting, and the maximum number of reproduced images that can be stored in the frame memory are attached to each image. It may be expressed in the form of a power. If the maximum frame size is determined in advance as the profile or level information, the maximum frame size may be input using such information. In the present embodiment, it is assumed that the value of the frame size is directly converted to binary coding, and the maximum frame size is defined as level information. The control method of the frame memory manager 314 according to the present invention will be described later.

  Next, a video prediction decoding method according to the present invention will be described. FIG. 4 is a block diagram of an image prediction decoding apparatus according to the embodiment of the present invention. 401 is an input terminal, 402 is a data analyzer, 403 is an inverse quantizer, 404 is an inverse transformer, 405 is an adder, 408 is a prediction signal generator, 407 is a frame memory, 407 is an output terminal, and 409 is a frame memory. It is a management device. The input terminal 401 corresponds to an input unit. The inverse quantizer 403 and the inverse transformer 404 correspond to decoding means. Other means may be used as the decoding means. The frame memory 407 corresponds to an image storage. The frame memory manager 409 corresponds to a memory control unit. Further, the inverter 404 may not be provided.

  The operation of the video predictive decoding device configured as described above will be described below. The compressed data compressed and encoded by the method described above is input from the input terminal 401. The compressed data includes information relating to prediction of a target block obtained by dividing an image into a plurality of blocks and generation of a coded residual signal and a prediction signal. Information related to generation of a prediction signal includes information (block size) on block division in the case of inter-screen prediction, and motion information and the above-described reference index. In the case of intra-screen prediction, neighboring reproduced pixels Contains information about the extrapolation method.

  In the data analyzer 402, the residual signal of the target block from the compressed data, information related to the generation of the prediction signal, the quantization parameter, the frame size of the image and the maximum frame size defined by the encoding setting, the frame memory To extract the maximum number of reproduced images that can be stored in. The residual signal of the target block is inversely quantized by the inverse quantizer 403 based on the quantization parameter (via the line L402). The result is subjected to inverse discrete cosine transform by an inverse transformer 404.

  Next, information related to the generation of the prediction signal is sent to the prediction signal generator 408 via the line L406b. The prediction signal generator 408 accesses the frame memory 407 based on information related to generation of a prediction signal, acquires a reference signal from a plurality of reference images, and generates a prediction signal. This prediction signal is sent to the adder 405 via the line L408, added to the restored residual signal, and the target block signal is reproduced, output via the line L405 and stored in the frame memory 407 at the same time.

  The frame memory 407 stores a reproduced image used for decoding and reproducing a subsequent image. The frame memory manager 409 controls the frame memory 407. The frame memory 407 is controlled so as to delete the oldest one of the stored M reproduced images (M is an integer in this case) and store the latest reproduced image used as a reference image. The frame memory manager 409 operates based on information on the frame size of the target image transmitted via the line L406a, the maximum frame size defined by the encoding setting, and the maximum number of reproduced images that can be stored in the frame memory. . The control method of the frame memory manager 409 according to the present invention will be described later.

  Next, the operation of the moving picture prediction encoding method and the moving picture prediction decoding method according to the present invention will be described with reference to FIGS. FIG. 5 is a flowchart illustrating a moving picture prediction encoding / decoding method according to an embodiment of the present invention. The meaning of the variables used in FIG. pic_width is the number of luminance pixels in the horizontal width of the reproduced image, pic_height is the number of luminance pixels in the vertical width of the reproduced image, MaxLumaFS is the maximum image size defined by the level information, and MFSBuffer is the decoded image buffer (DPB) defined as the level information. The memory buffer amount represented by the maximum number of reproducible images that can be stored, maxDPBsize is the memory buffer amount represented by the maximum number of reproducible images that can be stored in the decoded image buffer (DPB), and the max_dec_pic_buffering is the reproduced image used at the time of decoding. , DPBpointer is a variable indicating a pointer indicating a position where a reproduced image is stored in the decoded image buffer.

In FIG. 5, the number of pixels pic_width in the horizontal width of the reproduced image, the number pic_height in the vertical width of the reproduced image, the maximum frame size MaxLumaFS defined as the encoding setting, and the maximum number of reproduced images that can be stored in the decoded image buffer (DPB) The memory buffer amount MFSBuffer represented by is input. Next, the sum of pic_width, the number of pixels in the horizontal width of the input playback image, and pic_height, the number of pixels in the vertical width of the playback image, is compared with half the maximum image size MaxLumaFS (MaxLumaFS / 2) defined as the encoding setting. Is performed (step S502). Here, when the condition is not satisfied (when the value obtained by integrating pic_width and pic_height is larger than the value of MaxLumaFS / 2), maxDPBsize = MFSBuffer is set (step S503). When the condition is satisfied (when the value obtained by integrating pic_width and pic_height is equal to or less than the value of MaxLumaFS / 2), maxDPBsize = 2 * MFSBuffer is set (step S504).
(Formula)
maxDPBsize = 2 * MFSBuffer (when pic_width * pic_height ≦ (MaxLumaFS >> 1))
maxDPBsize = MFSBuffer (other than above)
Note that “>>” in the above formula indicates a right shift operation, and has the same meaning as MaxLumaFS / 2.

Then, the memory buffer amount max_dec_pic_buffering expressed by the maximum storage number of reproduced images used at the time of decoding is set to be equal to or less than maxDPBsize.
(Formula)
max_dec_pic_buffering ≦ maxDPBsize

  When maxDPBsize = MFSBuffer, DPBpointer = 1 is set at the head of the image memory buffer as shown in FIG. 6A, and DPBpointer = 2, 3, 4 is twice the maximum frame size from the position of DPBpointer = 1. The memory pointer is set to a position advanced by three times or four times the memory pointer position.

  In the case of maxDPBsize = 2 * MFSBuffer, in addition to the DPBpointers 1 to 4 already set as shown in FIG. The memory pointer is set to a position advanced by 5/2 times and 7/2 times. (Step S505).

  The set max_dec_pic_buffering and DPBpointer are used for controlling the frame memory in step S508 described later. That is, at the time of decoding, up to max_dec_pic_buffering sheets of reproduced images are stored in the frame memory, and at the time of storage, the reproduced images are controlled to be stored based on the set DPBpointer. In the encoding method, the number of luminance pixels pic_width in the horizontal width of the reproduced image, information on the number of luminance pixels pic_height in the vertical width of the reproduced image, the maximum image size MaxLumaFS, and the maximum number of reproduced images that can be stored in the decoded image buffer (DPB) are represented. The level information including the memory buffer amount MFSBuffer to be provided is externally given. In FIG. 3, it is provided from a control device (not shown) via the input terminal 313.

  On the other hand, in step S506, the image to be processed is encoded and decoded by the method described with reference to FIG. The encoded image data is sent out or stored outside. In step S507, it is determined whether the target image is used as a reference image in the subsequent processing. This determination is determined by the coding type of the image (intra-screen prediction coding, inter-screen prediction coding, bidirectional prediction coding). If it is not used as a reference image, the process proceeds to step S510. If it is used as a reference image, the process proceeds to step S508, where the DPBpointer is controlled at the position shown in FIG. 6, and the image decoded and reproduced in step S509 is stored in the frame memory, and then the process proceeds to step S511. In step S511, if there is a next image, the process proceeds to S506, and if there is no next image, the process ends. In this way, the encoding process is performed up to the last image.

  The above-described processing corresponds to the processing of the entire moving picture encoding method in FIG. 3, and in particular, steps S502, S503, S504, S505, and S508 are performed in the frame memory manager 314 according to the present invention.

  Although FIG. 5 has been described as a moving image encoding method, the present invention can be applied to processing of a moving image decoding method. When performing the decoding process, in step S501, the data (bit stream) of the compression-encoded image is input. From the data, the number of pixels of the width of the reproduced image pic_width, the number of pixels of the vertical width of the reproduced image pic_height, the maximum frame size MaxLumaFS defined as the encoding setting, and the maximum number of reproduced images that can be stored in the decoded image buffer (DPB) The expressed memory buffer amount MFSBuffer and the maximum number of stored reproduction images (memory buffer amount) max_dec_pic_buffering used at the time of decoding are extracted, and the control of steps S502 to S505 and S508 is performed in the same manner as described above. The decoding side checks whether or not the max_dec_pic_buffering restored in step S505 is equal to or less than maxDPBsize. In step S506, a process of decoding the compressed data of the target image and restoring the image is performed. The processing after step S507 is as described above. This processing corresponds to the processing of the moving picture decoding apparatus in FIG. 4, and in particular, steps S502, S503, S504, S505, and S508 are performed in the frame memory manager 409 of the moving picture decoding apparatus according to the present invention.

  FIG. 6 is a schematic diagram for explaining the position of the memory pointer of the frame memory in the moving picture prediction encoding / decoding method according to the embodiment of the present invention. As can be seen by comparing FIGS. 6A and 6B, the positions of the memory pointer numbers 1 to 4 are fixed. If the condition is satisfied in step S503, the control is newly performed as memory pointers 5 to 8. That is, the position of the memory pointer is fixed regardless of the determination in step S503.

(Memory pointer position)
Although the memory pointer numbers of this embodiment are set alternately from 1 to 4 and 5 to 8, the memory pointer numbers may be assigned in the order of 1 to 8.

(Number of memory partitions)
In this embodiment, the number of pixels pic_width of the horizontal width of the reproduced image and the number of pixels pic_height of the vertical width of the reproduced image are compared with half the maximum frame size MaxLumaFS (MaxLumaFS / 2) defined as the encoding setting. May compare pic_width, the number of pixels in the horizontal width of the reproduced image, and pic_height, the number of pixels in the vertical width of the reproduced image, with a maximum frame size MaxLumaFS / 2 ^L (L is an integer of 2 or more) defined as an encoding setting. . At this time, as shown in FIG. 6, the position of the memory pointer indicating the storage position of the reproduced image may be set at 1/2 ^L of MaxLumaFS with the already set pointer fixed.

(Definition of frame size)
In the present embodiment, the frame size is represented by the number of pixels pic_width of the horizontal width of the reproduced image and the number of pixels pic_height of the vertical width of the reproduced image. However, a value indicating the frame size integrated in advance may be used. Further, the frame size may be calculated by other methods.

  In the present embodiment, the maximum image size MaxLumaFS and the memory buffer amount MFSBuffer expressed by the maximum number of reproduced images of the size of MaxLumaFS that can be stored in the decoded image buffer (DPB) are defined as level information. It is not limited to being sent as information. It may be added as information other than the level information.

  In the present embodiment, FSBuffer and maxDPBsize are represented by the maximum number of storable reproduced images, but may be represented by an actual memory amount. In this case, the number of reproduced images may be calculated by dividing the memory amount by the frame size.

  In particular, the present invention determines the maximum number of reproduced images that can be stored in the image buffer memory on the basis of the relationship between the frame size of an image and the maximum frame size defined as a coding condition. When the frame size is smaller than the frame size, there is an effect that a waste of the frame memory is reduced and a memory area for a reference image that can further improve the coding efficiency is secured.

  A moving picture prediction encoding program and a moving picture prediction decoding program according to the present invention for causing a computer to function as the above-described moving picture prediction encoding apparatus and moving picture prediction decoding apparatus are provided by being stored in a recording medium as programs. . Examples of the recording medium include a recording medium such as a floppy (registered trademark) disk, a CD-ROM, a DVD, and a ROM, and a semiconductor memory.

Specifically, the moving image prediction encoding program includes an input module for inputting a plurality of images constituting the moving image, and encoding the image by any of an intra-screen prediction and an inter-screen prediction. Generates compressed image data including information on the size, the maximum number of images that can be stored, and the maximum frame size as an encoding setting, and codes the information on the target image frame size, the maximum number of images that can be stored, and the information on the maximum frame size. An encoding module that encodes together with the encoded data, a decoding module that decodes the compressed image data and restores the reproduced image, and stores one or more of the reproduced images as reference images used to encode subsequent images. An image storage module, and a memory control module for controlling an image buffer memory in the image storage module. Comprising a Lumpur, a memory control module, when the frame size of the input image with respect to the maximum frame size of the image is 1/2 L ^(L is an integer of 1 or more) or less, can be stored in the frame memory This is a moving image prediction encoding program characterized in that the maximum number of reproduced images is determined to be ^2L times.

Similarly, the moving image prediction encoding program encodes a plurality of images constituting a moving image by either intra-screen prediction or inter-screen prediction, and sets a target image frame size and a maximum number of images that can store a reproduced image. And an input module for inputting encoded data indicating compressed image data including information on the maximum frame size as an encoding setting, and information on the compressed image data and the target image frame size, the maximum number of images that can store reproduced images, and the maximum frame size A decoding module that decodes the compressed image data indicating the reproduction image, and restores the reproduced image to information on the target image frame size, the maximum number of images that can be stored and the maximum frame size, and decodes the reproduced image and the subsequent image Image storage module for storing one or more reference images used for Includes a memory control module that controls the module, a memory control module, the frame size is 1/2 L of the input image with respect to the maximum frame size of the image ^(L is an integer of 1 or more) in the case of times or less, A moving image prediction encoding program characterized in that the maximum number of reproduced images that can be stored in a frame memory is determined to be ^2L times.

  FIG. 7 is a diagram showing a hardware configuration of a computer for executing a program recorded on a recording medium, and FIG. 8 is a perspective view of a computer for executing a program stored on a recording medium. The computer includes a DVD player, a set-top box, a mobile phone, and the like that include a CPU and perform processing and control by software.

  As shown in FIG. 7, the computer 30 includes a reading device 12 such as a floppy (registered trademark) disk drive device, a CD-ROM drive device, and a DVD drive device, a working memory (RAM) 14 in which an operating system is resident, and A memory 16 for storing a program stored in the recording medium 10, a display device 18 such as a display, a mouse 20 and a keyboard 22 as input devices, a communication device 24 for transmitting and receiving data and the like, and execution of the program. And a CPU 26 for controlling the CPU. When the recording medium 10 is inserted into the reading device 12, the computer 30 becomes able to access the moving image prediction encoding / decoding program stored in the recording medium 10 from the reading device 12, and the moving image prediction encoding / decoding is performed. The program makes it possible to operate as a moving picture encoding device / decoding device according to the present invention.

  As shown in FIG. 8, the moving picture prediction coding program or the moving picture decoding program may be provided via a network as a computer data signal 40 superimposed on a carrier wave. In this case, the computer 30 can store the moving picture prediction coding program or the moving picture decoding program received by the communication device 24 in the memory 16 and execute the moving picture prediction coding program or the moving picture prediction decoding program. .

  In order to achieve the above object, a moving picture predictive encoding apparatus according to the present invention comprises: an input unit for inputting a plurality of pictures constituting a moving picture; Generates compressed image data including information about the target image frame size, the maximum number of images that can be stored and the maximum frame size as an encoding setting, the target image frame size, the maximum number of images that can be stored, and Encoding means for encoding together with the encoded data relating to the information on the maximum frame size, decoding means for decoding the compressed image data and restoring it to a reproduced image, and a reference used for encoding the reproduced image to a subsequent image Image storage means for storing at least one image as an image, wherein the frame size of the input image is 1 / the maximum frame size of the image. If times below, and determines the maximum number of storable reproduced image to twice in the frame memory.

  In this moving picture predictive coding apparatus, a moving image is stored in an image buffer memory based on a relationship between a frame size attached to each image constituting a moving image or compression-encoded image data and a maximum frame size defined by an encoding setting. The maximum number of possible reproduced images is determined. More specifically, for example, when the frame size of the input image is larger than half the maximum frame size defined in the encoding setting, up to N reproduced images can be stored (N is an integer of 1 or more). When the frame size of the input image is equal to or less than の of the maximum frame size, the amount of memory buffer available for storing the reproduced image is controlled so that a maximum of 2 × N reproduced images can be stored. Accordingly, when the frame size of the decoded reproduced image is equal to or less than half the maximum frame size defined in the encoding setting, the maximum number of storable reproduced images can be increased, and the reference image can be increased. As a result, encoding efficiency can be improved.

  Further, the video predictive encoding apparatus according to the present invention, when the frame size of the input image is 1/4 or less of the maximum frame size of the image, sets the maximum number of reproduced images that can be stored in the frame memory. It may be characterized in that it is determined four times.

  The moving picture prediction decoding apparatus according to the present invention is configured such that a plurality of pictures constituting a moving picture are encoded by either intra-screen prediction or inter-screen prediction, a target image frame size, and a maximum number of images that can be stored in a reproduced image. Input means for inputting encoded data indicating compressed image data including information on the maximum frame size as encoding settings; information on the compressed image data and the target image frame size; the maximum number of images that can store reproduced images and the maximum frame size; Decoding means for decoding the compressed image data indicating the reproduction image, and restoring the reproduced image into information on the target image frame size, the maximum number of images that can be stored and the maximum frame size, and decodes the reproduced image into a subsequent image Image storage means for storing one or more reference images used for the If the frame size of the input image relative's is 1/2 times or less, and determines the maximum number of storable reproduced image to twice in the frame memory.

  In this video predictive decoding device, it is possible to store in the image buffer memory based on the frame size associated with each image constituting the video or the compression-coded image data and the maximum frame size defined by the coding setting. The maximum number of reproduced images is determined. More specifically, for example, when the frame size of the input image is larger than half the maximum frame size defined in the encoding setting, up to N reproduced images can be stored (N is an integer of 1 or more). When the frame size of the input image is equal to or less than 最大 of the maximum frame size, a maximum of 2N reproduced images can be stored. Accordingly, when the frame size of the decoded reproduced image is equal to or less than half the maximum frame size defined in the encoding setting, the maximum number of storable reproduced images can be increased, and the reference image can be increased. As a result, encoding efficiency can be improved.

  Also, the video predictive decoding apparatus according to the present invention, when the frame size of the input image is 1/4 or less of the maximum frame size of the image, sets the maximum number of reproduced images that can be stored in the frame memory to 4 It may be characterized in that it is determined twice.

  Also, the moving picture prediction encoding method according to the present invention is a moving picture prediction encoding method executed by a moving picture prediction encoding apparatus, comprising: an input step of inputting a plurality of pictures constituting a moving picture; Is encoded by any method of intra-screen prediction or inter-screen prediction, and generates compressed image data including information about the target image frame size, the maximum number of images that can be stored as a playback image, and the maximum frame size as an encoding setting, An encoding step of encoding the target image frame size, the maximum number of reproducible images that can be stored, and encoded data relating to information on the maximum frame size; a decoding step of decoding compressed image data and restoring the reproduced image; Image storage means for storing at least one reference image as a reference image used for encoding a subsequent image. And determining that the maximum number of reproduced images that can be stored in the frame memory is doubled when the frame size of the input image is 以下 or less of the maximum frame size of the image. Features.

  Also, the video predictive encoding method according to the present invention, when the frame size of the input image is 1/4 or less of the maximum frame size of the image, the maximum number of reproduced images that can be stored in the frame memory. It may be characterized in that it is determined four times.

  Also, the moving picture prediction decoding method according to the present invention is a moving picture prediction decoding method executed by a moving picture prediction decoding apparatus, and performs intra-screen prediction or inter-screen prediction for a plurality of images constituting a moving image. An input step of inputting encoded data indicating compressed image data which is encoded by any of the above and includes information on a target image frame size, a maximum number of images that can be stored and a maximum frame size as an encoding setting, and compressed image data Decodes compressed image data indicating information about the target image frame size, the maximum number of frames that can store the reproduced image, and the maximum frame size, and reproduces the reproduced image, the target image frame size, the maximum number of frames that can store the reproduced image, and the maximum frame size. Decoding step to restore the reproduced image to information about the An image storage step of storing at least one reference image by an image storage unit, wherein the input image is stored in a frame memory when the frame size of the input image is 以下 or less of the maximum frame size of the image. The maximum number of possible reproduced images is determined to be twice.

  Also, the moving picture prediction decoding method according to the present invention, when the frame size of the input image is 1/4 or less of the maximum frame size of the image, sets the maximum number of reproduced images that can be stored in the frame memory to four. It may be characterized in that it is determined twice.

  Further, the moving picture prediction decoding method according to the present invention is characterized in that when determining the maximum number of reproduced images that can be stored in the frame memory, a memory pointer indicating a storage position of the reproduced image in the memory is fixed. You may do it.

  In this moving image predictive decoding method, the memory pointer indicating the storage position of each reproduced image can be fixed at, for example, the maximum frame size and a half of the maximum frame size. And a simpler and more efficient implementation of memory control becomes possible. Also, when calculating the position of the memory pointer, for example, it is possible to execute the shift operation because the change is limited to, for example, 1/2, and the operation cost is low.

  In addition, the moving image prediction decoding program according to the present invention can encode a plurality of images constituting a moving image by either intra-screen prediction or inter-screen prediction, and store a target image frame size and a reproduced image. An input module for inputting encoded data indicating compressed image data including information on the maximum frame size as the maximum number and encoding settings, the compressed image data and the target image frame size, the maximum number and the maximum frame size capable of storing the reproduced image A decoding module that decodes compressed image data indicating information about the reproduction image and restores the reproduction image and information on the target image frame size, the maximum number of images that can store the reproduction image, and information on the maximum frame size. An image storage module for storing one or more reference images used for decoding; Provided, when the frame size of the input image with respect to the maximum frame size of the image is 1/2 times or less, and determines the maximum number of storable reproduced image to twice in the frame memory.

  The moving picture prediction decoding method and the moving picture prediction decoding program can provide the same effects as those of the above-described moving picture prediction decoding apparatus.

  A moving image prediction decoding method according to the present invention is a moving image prediction decoding method executed by a moving image prediction decoding device, and includes a method of performing either an intra-screen prediction or an inter-screen prediction for a plurality of images constituting a moving image. Inputting encoded data indicating compressed image data including information on the target image frame size, the maximum number of images that can be stored, and the maximum frame size as an encoding setting, and inputting the compressed image data and the target Decodes compressed image data indicating information about the image frame size, the maximum number of frames that can store a reproduced image, and the maximum frame size, and reproduces the reproduced image, the target image frame size, and the information about the maximum number and the maximum frame size that can store the reproduced image. Decoding step to restore the reproduced image to a reference used to decode the subsequent image An image storing step of storing at least one image as an image by an image storing means, wherein the maximum of a reproduced image that can be stored in the frame memory is determined based on a relationship between a maximum frame size of the image and a frame size of the input image. The number of sheets is determined to be equal to, twice, or four times a predetermined value.

  Further, in the moving picture prediction decoding method according to the present invention, when the frame size of the input image is equal to or less than １ ／ times the maximum frame size of the image, the maximum number of reproduced images that can be stored in the frame memory is determined. The value may be determined to be twice the value of.

  Further, in the video predictive decoding method according to the present invention, when the frame size of the input image is 1/4 or less of the maximum frame size of the image, the maximum number of reproduced images that can be stored in the frame memory is determined. May be determined to be four times the value of.

  The moving picture prediction decoding apparatus according to the present invention is configured such that a plurality of pictures constituting a moving picture are encoded by either intra-screen prediction or inter-screen prediction, a target image frame size, and a maximum number of images that can be stored in a reproduced image. Input means for inputting encoded data indicating compressed image data including information on the maximum frame size as encoding settings; information on the compressed image data and the target image frame size; the maximum number of images that can store reproduced images and the maximum frame size; Decoding means for decoding the compressed image data indicating the reproduction image, and restoring the reproduced image into information on the target image frame size, the maximum number of images that can be stored and the maximum frame size, and decodes the reproduced image into a subsequent image Image storage means for storing one or more reference images used for the Based on the relationship between the frame size's and the input image, magnification of a predetermined value the maximum number of storable reproduced image in the frame memory, and determining the double or quadruple.

  In order to achieve the above object, a video predictive decoding method according to the present invention is a video predictive decoding method executed by a video predictive decoding device, wherein a plurality of images constituting a video are displayed on a screen. An input for inputting encoded data indicating compressed image data that is encoded by either prediction or inter-screen prediction and includes information on the target image frame size, the maximum number of reproducible images that can be stored, and the maximum frame size as an encoding setting Step: decoding the compressed image data and the target image frame size, the compressed image data indicating information on the maximum number of frames that can store the reproduced image and the maximum frame size, and storing the reproduced image, the target image frame size, and the reproduced image. A decoding step of restoring the information on the maximum number of frames and the maximum frame size; Image storing step of storing at least one reference image used by the image storage unit by an image storage unit, based on a relationship between a shift operation value of a maximum frame size of the image and a frame size of the input image. The maximum number of reproduced images that can be stored in the memory is determined to be equal, twice, or four times a predetermined value.

  Also, in the moving picture prediction decoding method according to the present invention, when the frame size of the input image is equal to or smaller than a value obtained by shifting the maximum frame size of the image by 1 bit right, the maximum number of reproduced images that can be stored in the frame memory Is determined to be twice the predetermined value, and when the frame size of the input image is equal to or smaller than a value obtained by performing a right shift operation of the maximum frame size of the image by 2 bits, the reproduction image that can be stored in the frame memory is The maximum number may be determined to be four times a predetermined value.

  The moving picture prediction decoding apparatus according to the present invention is configured such that a plurality of pictures constituting a moving picture are encoded by either intra-screen prediction or inter-screen prediction, a target image frame size, and a maximum number of images that can be stored in a reproduced image. Input means for inputting encoded data indicating compressed image data including information on the maximum frame size as encoding settings; information on the compressed image data and the target image frame size; the maximum number of images that can store reproduced images and the maximum frame size; Decoding means for decoding the compressed image data indicating the reproduction image, and restoring the reproduced image into information on the target image frame size, the maximum number of images that can be stored and the maximum frame size, and decodes the reproduced image into a subsequent image Image storage means for storing one or more reference images used for the The maximum number of reproduced images that can be stored in the frame memory is determined to be equal to, twice, or four times a predetermined value based on the relationship between the shift operation value of the shift and the frame size of the input image. Features.

  Reference numeral 301: input terminal, 302: block divider, 303: predicted signal generator, 304: frame memory, 305: subtractor, 306: converter, 307: quantizer, 308: inverse quantizer, 309: inverse transform Adder, 310 adder, 311 entropy coder, 312 output terminal, 313 input terminal, 314 frame memory manager, 401 input terminal, 402 data analyzer, 403 inverse quantizer, 404 … Inverter, 405 adder, 406 output terminal, 407 frame memory, 408 predicted signal generator, 409 frame memory manager

Claims (11)

A CPU programmable to perform video prediction encoding;
A receiving unit executable by the programmed CPU to receive a plurality of images constituting a moving image;
An encoding unit that is executable by the programmed CPU to encode a received image by any of intra-screen prediction and inter-screen prediction, and to generate compressed image data;
A decoding unit capable of decoding compressed image data by the programmed CPU and executing decoding of the compressed image data into a reproduced image;
A frame memory executable by the programmed CPU to store one or more playback images as reference images used to encode subsequent images;
A memory management unit executable by the programmed CPU to manage the frame memory, the memory management unit being defined in advance by a target image frame size indicating frame size of a target image and level information, and stored in the frame memory. A maximum frame size (MaxLumaFS) indicating the maximum size of a storable playback image and a frame memory size (MFSBuffer) represented by a preset maximum number of playback images storable in the frame memory are acquired. A memory management unit, which is executable as
With
The memory management unit,
Compare 1/2 times MaxLumaFS with the target image frame size,
If it is determined that 1/2 times MaxLumaFS is equal to or larger than the target image frame size, MFSBuffer is doubled,
Setting the adaptive maximum number of reproduced images (maxDPBsize) that can be stored in the frame memory to twice the MFSBuffer;
Is even more feasible,
The number of reproduced images stored in the frame memory is equal to or less than the number set in maxDPBsize.
A moving picture prediction encoding device. The programmed CPU causes the memory management unit to:
Compare 1/4 times MaxLumaFS with the target image frame size,
When it is determined that 1/4 of MaxLumaFS is equal to or larger than the target image frame size, MFSBuffer is quadrupled,
set maxDPBsize to 4 times MFSBuffer,
The video prediction encoding apparatus according to claim 1, wherein the apparatus is further executable. A CPU programmed to perform video prediction decoding;
A receiving unit executable by said CPU to receive encoded data including compressed image data representing a compressed form of a plurality of images constituting a moving image, wherein Or a receiving unit encoded by any method of inter-screen prediction,
A decoding unit executable by the CPU to decode compressed image data and restore a reproduced image;
A decoded picture buffer (DPB: Decoded Picture Buffer) executable by the CPU to store one or more playback pictures as reference pictures used to decode subsequent pictures;
A buffer management unit executable by the CPU to manage the DPB, comprising a target image frame size indicating a frame size of a target image, and a reproduced image pre-defined by level information and storable in the DPB. And a frame memory size (MFSBuffer) represented by a preset maximum number of reproduction images that can be stored in the DPB. A buffer management unit;
With
The buffer management unit,
Compare 1/2 times MaxLumaFS with the target image frame size,
If it is determined that 1/2 times MaxLumaFS is equal to or larger than the target image frame size, MFSBuffer is doubled,
Setting the adaptive maximum number of reproduced images (maxDPBsize) that can be stored in the DPB to twice the MFSBuffer;
Is even more feasible,
The number of reproduced images stored in the DPB is equal to or less than the number set in maxDPBsize.
Video prediction decoding device. The buffer management unit,
Compare 1/4 times MaxLumaFS with the target image frame size,
When it is determined that 1/4 of MaxLumaFS is equal to or larger than the target image frame size, MFSBuffer is quadrupled,
set maxDPBsize to 4 times MFSBuffer,
4. The moving image prediction decoding apparatus according to claim 3, wherein the apparatus is further executable. A video prediction encoding method executed by the video prediction encoding device,
Receiving a plurality of images constituting a moving image;
Encoding the received image by any of intra-screen prediction or inter-screen prediction, and generating compressed image data;
Decoding the compressed image data and restoring it to a reproduced image;
Storing in the frame memory one or more reconstructed images as reference images used to encode subsequent images;
A target image frame size that indicates a frame size of the target image, a maximum frame size (MaxLumaFS) that is defined in advance by level information and indicates a maximum size of a reproduced image that can be stored in the frame memory, and stored in the frame memory. Acquiring a frame memory size (MFSBuffer) represented by a preset maximum number of possible playback images;
Comparing 1/2 of MaxLumaFS with the target image frame size;
Doubling MFSBuffer when it is determined that 1/2 times MaxLumaFS is equal to or larger than the target image frame size;
Setting the adaptive maximum number of reproduced images (maxDPBsize) that can be stored in the frame memory to twice MFSBuffer;
Including
The number of reproduced images stored in the frame memory is equal to or less than the number set in maxDPBsize.
Video prediction encoding method. Comparing 1/4 of MaxLumaFS with the target image frame size;
When it is determined that 1/4 of MaxLumaFS is equal to or larger than the target image frame size, quadrupling the MFSBuffer;
setting maxDPBsize to four times MFSBuffer;
The moving picture prediction encoding method according to claim 5, further comprising: A video prediction decoding method executed by the video prediction decoding device,
Receiving encoded data including compressed image data representing a compressed form of a plurality of images constituting the moving image, wherein the plurality of images are encoded by either an intra-screen prediction or an inter-screen prediction. The steps being encoded;
Decoding the compressed form of the plurality of images and restoring the reproduced images;
Storing one or more playback images in a decoded image buffer (DPB) as a reference image used to decode subsequent images;
A target image frame size indicating the frame size of the target image, and a maximum frame size (MaxLumaFS) that is defined in advance by the level information and indicates the maximum size of the reproduced image that can be stored in the DPB; Obtaining a frame memory size (MFSBuffer) represented by a preset maximum number of reproduced images;
Comparing 1/2 of MaxLumaFS with the target image frame size;
Doubling MFSBuffer when it is determined that 1/2 times MaxLumaFS is equal to or larger than the target image frame size;
Setting the adaptive maximum number of reproduced images (maxDPBsize) that can be stored in the DPB to twice the MFSBuffer;
Including
The number of reproduced images stored in the DPB is equal to or less than the number set in maxDPBsize.
Video prediction decoding method. Comparing 1/4 of MaxLumaFS with the target image frame size;
When it is determined that 1/4 of MaxLumaFS is equal to or larger than the target image frame size, quadrupling the MFSBuffer;
setting maxDPBsize to four times MFSBuffer;
The moving picture prediction decoding method according to claim 7, further comprising:   8. The moving picture prediction decoding method according to claim 7, further comprising, after setting maxDPBsize to twice MFSBuffer, keeping a memory pointer indicating a storage position of the DPB divided according to MFSBuffer.   9. The moving picture prediction decoding method according to claim 8, further comprising, after setting maxDPBsize to four times MFSBuffer, keeping a memory pointer indicating a storage position of the DPB divided according to MFSBuffer. A recording medium storing a moving image prediction decoding program, wherein the moving image prediction decoding program causes a CPU to:
A receiving unit capable of receiving encoded data including compressed image data representing a compressed form of a plurality of images constituting a moving image, wherein the plurality of images are intra-screen prediction or inter-screen prediction. A receiving unit, which is encoded by any one of the above methods,
A decoding unit that decodes the compressed image data and is executable to restore the reproduced image;
A decoded picture buffer (DPB) that is executable to store one or more replay pictures as reference pictures used to decode subsequent pictures;
A buffer management unit executable to manage the DPB, comprising: a target image frame size indicating a frame size of a target image; and a maximum size of a reproduced image that is defined in advance by level information and can be stored in the DPB. A buffer management unit executable to acquire a maximum frame size (MaxLumaFS) and a frame memory size (MFSBuffer) represented by a preset maximum number of reproduction images that can be stored in the DPB. ,
To execute
The buffer management unit,
Compare 1/2 times MaxLumaFS with the target image frame size,
If it is determined that 1/2 times MaxLumaFS is equal to or larger than the target image frame size, MFSBuffer is doubled,
Setting the adaptive maximum number of reproduced images (maxDPBsize) that can be stored in the DPB to twice the MFSBuffer;
Is even more feasible,
The number of reproduced images stored in the DPB is equal to or less than the number set in maxDPBsize.
recoding media.

Images