(19)
(11)EP 2 571 271 B1

(12)EUROPEAN PATENT SPECIFICATION

(45)Mention of the grant of the patent:
03.08.2016 Bulletin 2016/31

(21)Application number: 12183156.4

(22)Date of filing:  05.09.2012
(51)International Patent Classification (IPC): 
H04N 19/52(2014.01)
H04N 19/615(2014.01)
H04N 19/523(2014.01)

(54)

Method for coding and reconstructing a pixel block and corresponding devices

Verfahren zur Codierung und Wiederherstellung eines Pixelblocks und zugehörige Verfahren

Procédé de codage et de reconstruction d'un bloc de pixels et dispositifs correspondants


(84)Designated Contracting States:
AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR

(30)Priority: 13.09.2011 FR 1158161

(43)Date of publication of application:
20.03.2013 Bulletin 2013/12

(73)Proprietor: Thomson Licensing
92130 Issy-les-Moulineaux (FR)

(72)Inventors:
  • Thoreau, Dominique
    35576 Cesson-Sévigné (FR)
  • François, Edouard
    35890 Bourg des Comptes (FR)
  • Vieron, Jérôme
    75015 Paris (FR)
  • Martin, Aurélie
    75015 Paris (FR)

(74)Representative: Lorette, Anne 
Technicolor 1, rue Jeanne d'Arc
92130 Issy-les-Moulineaux
92130 Issy-les-Moulineaux (FR)


(56)References cited: : 
EP-A1- 2 346 254
US-A1- 2008 212 676
  
  • SULLIVAN G J ET AL: "Video Compression-From Concepts to the H.264/AVC Standard", PROCEEDINGS OF THE IEEE, IEEE. NEW YORK, US, vol. 93, no. 1, 1 January 2005 (2005-01-01), pages 18-31, XP011123850, ISSN: 0018-9219, DOI: 10.1109/JPROC.2004.839617
  
Note: Within nine months from the publication of the mention of the grant of the European patent, any person may give notice to the European Patent Office of opposition to the European patent granted. Notice of opposition shall be filed in a written reasoned statement. It shall not be deemed to have been filed until the opposition fee has been paid. (Art. 99(1) European Patent Convention).


Description

1. Scope of the invention



[0001] The invention relates to the general domain of image coding. More precisely, the invention relates to a method for coding a pixel block and a method for reconstructing such a block.

2. Prior art



[0002] Video coding devices comprising a motion estimation module capable of estimating motion vectors at a sub-pixel accuracy are known. As an example, the motion estimation module estimates displacements on an interpolated image using interpolation filters at ½, ¼ possibly 1/8 of pixel.

[0003] Motion estimation modules using correlators (for example, phase correlation) or global motion estimations which enable motion vectors to be estimated with a sub-pixel accuracy less than 1/8 of pixel are also known.

[0004] The usual interpolation filters whose size can be large and whose accuracy can be limited to 1/8 of pixel do not allow an accurate motion compensation when the motion vector from which the motion compensation is made has components at a sub-pixel accuracy less than 1/8 of pixel. Such a usual interpolation filter is described in section 8.4.2.2 of ISO/IEC standard 14496-10 to make the interpolation at ½ pixel.

[0005] The document US 2008/212676 describes a local motion estimation. To this aim, an extended-block FFT is calculated for each block. Extending the block for FFT helps to account for the motion of objects that are moving into or out of the block.

[0006] The document from Sullivan et al entitled "Video Compression-From Concepts to the H.264/AVC Standard" explains the basic concepts of video codec design and then explains how these various features have been integrated into international standards.

3. Summary of the invention



[0007] The purpose of the invention is to overcome at least one of the disadvantages of the prior art. For this purpose, the invention relates to a method for coding a block of pixels comprising the following steps:
  • determine a prediction block for the pixel block from a motion vector;
  • calculate a residue between the pixel block and the prediction block; and
  • code the residue
Advantageously, the determination of the prediction block comprises the following steps:
  • determine an intermediate prediction block of size strictly greater than the size of the pixel block;
  • transform the intermediate prediction block into a first block transformed with a first transform; and
  • transform the first transformed block into a second block transformed with a second inverse transform of the first transform whose basis functions are shifted by at least a part of each component of the motion vector, the prediction block being extracted from the second transformed block. According to a first embodiment, the intermediate prediction block is determined by adding at least a pixel line and at least a pixel column in the direction of motion to the block obtained by motion compensation of the pixel block from an intermediate motion vector whose components are the integer parts of the components of the motion vector and the basis functions of the second transform are shifted by the fractional parts of the components of the motion vector.


[0008] According to a first embodiment, the intermediate prediction block is determined by adding at least a pixel line and at least a pixel column in the direction of motion to a block co-located to the block of pixels to be coded in such a way that the size of the intermediate prediction block is strictly greater than the displacement corresponding to the integer parts of the components of the motion vector and the basis functions of the second transform are shifted by all the components of the motion vector.

[0009] According to a particular characteristic of the invention, the size of the intermediate prediction block is a power of 2.

[0010] According to another particular characteristic of the invention, the first transform is a 2D discrete cosine transform.

[0011] The invention also relates to a pixel block comprising the following steps:
  • determine a prediction block for the pixel block from a motion vector;
  • decode a residue for the pixel block; and
  • reconstruct the pixel block from the prediction block and the residue. Advantageously, the determination of the prediction block comprises the following steps:
  • determine an intermediate prediction block of size strictly greater than the size of the pixel block;
  • transform the intermediate prediction block into a first block transformed with a first transform;
  • transform the first transformed block into a second block transformed with a second inverse transform of the first transform whose basis functions are shifted by at least a part of each component of the motion vector, the prediction block being extracted from the second transformed block.


[0012] The invention further relates to a device for coding a pixel block comprising the following means:
  • means to determine a prediction block for the pixel block from a motion vector;
  • means to calculate a residue between the pixel block and the prediction block; and
  • means to code the residue.


[0013] Advantageously, the prediction block determination means comprise the following means:
  • means to determine an intermediate prediction block of size strictly greater than the size of the pixel block;
  • means to transform the intermediate prediction block into a first block transformed with a first transform;
  • means to transform the first transformed block into a second block transformed with a second inverse transform of the first transform whose basis functions are shifted by at least a part of each component of the motion vector; and
  • means to extract the prediction block of the second transformed block.


[0014] The invention also relates to a device for decoding a stream with a view to the reconstruction of a pixel block comprising the following means:
  • means to determine a prediction block for the pixel block from a motion vector;
  • means to decode a residue for the pixel block from the stream;
  • means to reconstruct the pixel block from the prediction block and the residue;


[0015] Advantageously, the prediction block determination means comprise the following means:
  • means to determine an intermediate prediction block of size strictly greater than the size of the pixel block;
  • means to transform the intermediate prediction block into a first block transformed with a first transform;
  • means to transform the first transformed block into a second block transformed with a second inverse transform of the first transform whose basis functions are shifted by at least a part of each component of the motion vector; and
  • means to extract the prediction block of the second transformed block.

4. List of figures



[0016] The invention will be better understood and illustrated by means of non-restrictive embodiments and advantageous implementations, with reference to the accompanying drawings, wherein:
  • figure 1 shows a coding method according to the invention;
  • figures 2 and 3 show a block to be coded Bc and an intermediate prediction block b';
  • figure 4 shows a transformed block B2 and a prediction block Bp;
  • figure 5 shows a reconstruction method according to the invention;
  • figure 6 illustrates a coding device according to the invention; and
  • figure 7 shows a decoding device according to the invention.

5. Detailed description of the invention



[0017] Figure 1 shows, according to the invention, a method for coding a pixel block Bc of an image Ic, Bc is a block of size MxM with M integer. The pixel block Bc belongs to an image Ic of an image sequence. During a step 10, a prediction block Bp is determined for the pixel block to be coded Bc from a motion vector Vp of components (Vx, Vy) with Vx=dx+dxrs and Vy=dy+dyrs, where (dx, dy) are the integer parts of the components and (dxrs, dyrs) are the fractional parts of the components. For example, if Vx=2.28 then dx=2 and dxrs=0.28 and if Vx=-3.73 then dx=-3 and dxrs=-0.73. This motion vector Vp associated with the block Bc comes from for example a motion estimation by phase correlation or even an global motion estimation. Vp indicates the displacement of the block Bc between the current image Ic and the reference image Ir. The invention is in no way limited by the method used to obtain the motion vector Vp. The step for determining 10 the prediction block Bp, also known as motion compensation step, comprises, according to a first embodiment represented on figure 2, a step 110 for determining an intermediate prediction block b' in a reference image Ir from an intermediate motion vector whose components are the integer parts (dx, dy) of the components of the motion vector Vp as illustrated on figure 2. The intermediate prediction block b' comprises block b which is obtained by motion compensation of block Bc from the intermediate motion vector of components (dx, dy). b' is of size NxN, with N strictly greater than M. In a first variant, N=M+1. More precisely, the block b' is obtained by adding to block b at least 1 pixel line and 1 pixel column on 2 of its sides in the direction of displacement/movement as illustrated in figure 2. The direction of motion is given by the motion vector Vp. Thus, in figure 2, the sub-pixel displacement takes place downwards and to the right in the block b'. Consequently, the block b' is a block increased by 1 pixel line on top and 1 pixel column on the left in relation to block b.

[0018] For example, if the block Bc to be coded and therefore the associated prediction block Bp is a block of size 8x8, the intermediate prediction block b' is a block of size 9x9 obtained by adding to block b 1 pixel line and 1 pixel column in the direction of motion as illustrated in figure 2.

[0019] According to a second embodiment variant, the intermediate prediction block b' is obtained by adding to block b as many pixel lines and columns as necessary on 2 of its sides in the direction of displacement/movement so that N is a power of 2. This variant enables fast transform algorithms to be used. The intermediate prediction block b' is then transformed during a step 120 into a first transformed block B1 with a first transform T of dimension N. T is, for example, a separable DCT ("discrete cosine transform") transform whose basis functions c(i,j) are defined as follows:

and



[0020] Consequently,



[0021] The first transformed block B1 is transformed, during a step 130, into a second transformed block B2 with a second inverse transform of the first transform and whose basis functions are shifted by the fractional parts (dxrs, dyrs) of the components of the motion vector Vp. The second transform is, for example, an inverse DCT transform whose basis functions are defined as follows:





[0022] The basis functions are therefore shifted by the fractional parts (dxrs, dyrs) of the components of the motion vector in the case described with reference to figure 2.

[0023] Consequently,



[0024] During a step 140, the prediction block Bp is obtained by extracting from the second transformed block B2 the part corresponding to block b. In the particular case of figure 2, the block Bp is obtained by deleting from B2 the first pixel line and the first pixel column.

[0025] The intermediate prediction block b' is therefore obtained from block b by adding pixel lines and columns so that the sub-pixel displacements of the image signal inside the intermediate prediction block b' via the inverse transforms (3) and (4) are carried out from the pixels of the intermediate prediction block b' to the pixels of block b (i.e. in the direction opposite to motion).

[0026] During a step 12, a residue or residual block is calculated between the pixel block Bc and the prediction block Bp. The residue is generally calculated by differentiating pixel by pixel between the pixel block to be coded Bc and the prediction block Bp determined at step 10. This difference is possibly weighted by taking account of a luminosity variation model.

[0027] During a step 14, the residue is coded in a coded data stream. This step generally comprises the transformation, the quantization and the entropy coding of the residue. These steps are well known to those skilled in the art of video coders and are not further described. This step 14 possibly comprises the coding of the motion vector Vp associated with the pixel block to be coded Bc. According to one variant, the motion vector is not coded. For example, the motion vector Vp associated with the pixel block to be coded Bc is determined on the coder and decoder side by the same method of the template matching type.

[0028] According to a second embodiment, the intermediate prediction block b' determined at step 110 is obtained from block b co-located to block Bc in the image Ir by enlarging it, i.e. by adding one or more pixel lines and columns, in the direction of motion up to a size NxN strictly greater than the displacement corresponding to the integer parts (dx,dy) of the vector Vp, i.e. N>dx and N>dy as illustrated in figure 3. This size is for example a power of 2 in order to enable fast transform algorithms to be used.

[0029] The intermediate prediction block b' is then transformed during a step 120 into a first transformed block B1 with a first transform T. T is, for example, a separable DCT ("discrete cosine transform"). Consequently,



[0030] The first transformed block B1 is transformed, during a step 130, into a second transformed block B2 with a second inverse transform of the first transform whose basis functions are shifted by the integer and fractional parts of the components of the motion vector.

[0031] The second transform is, for example, an inverse DCT transform whose basis functions are defined as follows:





[0032] The basis functions are therefore shifted by the components of the motion vector Vp in the case described with reference to figure 3.

[0033] Consequently,



[0034] During a step 140, the prediction block Bp is obtained by extracting from the second transformed block B2 the part corresponding to the co-located block b. In figure 4, the block Bp is hatched.

[0035] One of the advantages of the coding method according to the invention is that it uses a smaller support than the supports of usual interpolation filters. So, for ¼ pixel interpolation, the usual interpolation filters have a length equal to 6 coefficients. The use of such interpolation filters poses problems on the image edges and necessitates the use of padding techniques. The method according to the invention makes it possible to extricate oneself from this problem. Indeed, on the image edges, the first embodiment with N=M+1 is preferentially used.

[0036] Figure 5 represents according to the invention a method for reconstructing a pixel block Bc from a stream F of coded data.

[0037] The steps identical to the steps of the coding method are identified in figure 5 using the same numerical references and are not further described.

[0038] During a step 10, a prediction block is determined for the pixel block to be coded from a motion vector Vp of components (Vx, Vy) with Vx=dx+dxrs and Vy=dy+dyrs. This vector comes from for example the decoding of a part of a stream F of coded data. According to one variant, the vector Vp is determined by template matching. Step 10 comprises in particular steps 110, 120, 130 and 140 to determine a prediction block Bp. These steps are identical to those described with reference to figure 1 for the coding method. The embodiment variants described for the coding method are also applicable to the decoding method.

[0039] During a step 22, a residue is decoded for the block to be reconstructed Bc from the stream F. This step generally comprises the entropy decoding of at least a part of the stream F, the reverse quantization and the inverse transform. These steps are well known to those skilled in the art of video coders and are not further described. These are the inverse steps to those carried out at step 14 of the coding method.

[0040] During a step 24, the block Bc is reconstructed from the residue and the prediction block Bp. The block Bc is generally reconstructed by adding up pixel by pixel the residue and the prediction block determined at step 10. This sum is possibly weighted by taking account of a luminosity variation model.

[0041] The invention further relates to a coding device 12 described with reference to figure 6 and a decoding device 13 described with reference to figure 7. In this figure, the modules shown are functional units that may or may not correspond to physically distinguishable units. For example, these modules or some of them can be grouped together in a single component or circuit, or constitute functions of the same software. On the contrary, some modules may be composed of separate physical entities.

[0042] The coding device 12 receives at input images I belonging to a sequence of images. Each picture is divided into blocks of pixels with each of which at least one item of picture data, is associated, e.g. of luminance and/or of chrominance. The coding device 12 notably implements a coding with temporal prediction. Only the modules of the coding device 12 relating to coding by temporal prediction or INTER coding are represented in figure 6. Other modules not represented and known to those skilled in the art of video coders implement the INTRA coding with or without spatial prediction. The coding device 12 notably comprises a calculation module ADD1 capable of subtracting pixel by pixel from a current block Bc a prediction block Bp to generate a residue or residual block noted res. It further comprises a module TQ capable of transforming then quantizing the residual block res into quantized data. The transform T is for example a DCT. The coding device 12 further comprises an entropy coding module COD able to code the quantized data into a stream F of coded data. It further comprises a module ITQ carrying out the reverse operation of module TQ. The module ITQ carries out a reverse quantization followed by an inverse transform. The module ITQ is connected to a calculation module ADD2 able to add pixel by pixel the block of data from the module ITQ and the prediction block Bp to generate a block of reconstructed image data that is stored in a memory MEM.

[0043] The coding device 12 also comprises a motion estimation module ME able to estimate at least one motion vector Vp between the block Bc and a reference image Ir stored in the memory MEM, this image having previously been coded then reconstructed. According to one variant, the motion estimation can be made between the current block Bc and the source image corresponding to Ir, in which case the memory MEM is not connected to the motion estimation module ME. According to a method well known to those skilled in the art, the motion estimation module searches in the reference image Ir, respectively in the corresponding source image, for a motion vector so as to minimise an error calculated between the current block Bc and a block in the reference image Ir, respectively in the corresponding source image, identified using said motion vector. According to one variant, the motion vector is determined by phase correlation or global motion estimation. The motion data are transmitted by the motion estimation module ME to a decision module DECISION able to select a coding mode for the block Bc in a predefined set of coding modes. The chosen coding mode is for example the one that minimizes a bitrate-distortion type criterion. However, the invention is not restricted to this selection method and the mode chosen can be selected according to another criterion for example an a priori type criterion. The coding mode selected by the decision module DECISION as well as the motion data, for example the motion vector or vectors in the case of the temporal prediction mode or INTER mode are transmitted to a prediction module PRED. The motion vector or vectors and the selected coding mode are moreover transmitted to the entropy coding module COD to be coded in the stream F. If a prediction mode INTER is retained by the decision module DECISION the prediction module PRED then determines in the reference image Ir previously reconstructed and stored in the memory MEM, the prediction block Bp from the motion vector determined by the motion estimation module ME. If a prediction mode INTRA is retained by the decision module DECISION, the prediction module PRED determines in the current image, among the blocks previously coded and stored in the memory MEM, the prediction block Bp.

[0044] The prediction module PRED is able to determine the prediction block Bp according to steps 110, 120, 130 and 140 of the coding method described with reference to figure 1.

[0045] The decoding device 13 is described with reference to figure 7. The decoding device 13 receives at input a stream F of coded data representative of a sequence of images. The stream F is for example transmitted by a coding device 12. The decoding device 13 comprises an entropy decoding module DEC able to generate decoded data, for example coding modes and decoded data relating to the content of the images. The decoding device 13 further comprises a motion data reconstruction module. According to a first embodiment, the motion data reconstruction module is the entropic decoding module DEC that decodes a part of the stream F representative of motion vectors.

[0046] According to a variant not shown in figure 7, the motion data reconstruction module is a motion estimation module. This solution for reconstructing motion data by the decoding device 13 is known as "template matching".

[0047] The decoded data relating to the content of the pictures is then sent to a module ITQ capable of carrying out a reverse quantization followed by an inverse transformation. The module ITQ is identical to the module ITQ of the coding device 12 having generated the coded stream F. The module ITQ is connected to a calculation module ADD able to add pixel by pixel the block from the module ITQ and a prediction block Bp to generate a block of reconstructed image data that is stored in a memory MEM. The decoding device 13 also comprises a prediction module PRED identical to the prediction module PRED of the coding device 12. If a prediction mode INTER is decoded, the prediction module PRED determines in a reference image Ir previously reconstructed and stored in the memory MEM, the prediction block Bp from the motion vector Vp decoded for the current block Bc by the entropy decoding module DEC. If a prediction mode INTRA is decoded, the prediction module PRED determines in the current image among the blocks previously reconstructed and stored in the memory MEM, the prediction block Bp.

[0048] The prediction module PRED is able to determine the prediction block Bp according to steps 110, 120, 130 and 140 of the reconstruction method described with reference to figure 5.

[0049] The coding and decoding devices according to the invention are for example implemented in various forms of hardware, software, firmware, special purpose processors, or a combination thereof. Preferably, the present principles may be implemented as a combination of hardware and software. Moreover, the software is preferably implemented as an application program tangibly embodied on a program storage device. The application program may be uploaded to, and executed by, a machine comprising any suitable architecture. Preferably, the machine is implemented on a computer platform having hardware such as one or more central processing units (CPU), a random access memory (RAM), and input/output (I/O) interface(s). The computer platform also includes an operating system and microinstruction code. The various processes and functions described herein may either be part of the microinstruction code or part of the application program (or a combination thereof) that is executed via the operating system. In addition, various other peripheral devices may be connected to the computer platform such as an additional data storage device and a printing device.

[0050] According to variants, the coding and decoding devices according to the invention are implemented according to a purely hardware realisation, for example in the form of a dedicated component (for example in an ASIC (Application Specific Integrated Circuit) or FPGA (Field-Programmable Gate Array) or VLSI (Very Large Scale Integration) or of several electronic components integrated into a device or even in a form of a mix of hardware elements and software elements.

[0051] Obviously, the invention is not limited to the embodiments mentioned above.

[0052] In particular, those skilled in the art may apply any variant to the stated embodiments and combine them to benefit from their various advantages. In particular, the invention described with the DCT can be applied to other separable or non separable transforms.

[0053] In addition, the invention can be applied to any form of blocks, i.e. not rectangular. As previously mentioned, the invention can be applied to other transforms, like shape adaptive transforms of the SADCT type. The SADCT is notably described in the following documents:

Kaup A., Panis S., On the Performance of the Shape Adaptive DCT in Object-based coding of motion compensated difference Images; 1997

Stasinski R., Konrad J., Reduced-complexity shape-adaptive dct for region-based image coding, USA; 1998



[0054] The invention described for a pixel block can be applied to several blocks of an image or even to several blocks of a sequence of several images.


Claims

1. Method for coding a pixel block comprising:

- determining (10) a prediction block for said pixel block from a motion vector for which each component comprises an integer part and a fractional part;

- calculating (12) a residue between said pixel block and said prediction block;

- coding (14) said residue,

said coding method being characterized in that the determination of said prediction block comprises:

- motion compensating (110) said pixel block with an intermediate motion vector whose components are the integer parts of the components of said motion vector;

- determining (110) an intermediate prediction block of size strictly greater than the size of said pixel block by adding at least one line of adjacent pixels and at least one column of adjacent pixels in the direction of motion to said motion compensated pixel block, said direction of motion being given by said motion vector;

- transforming (120) into a frequency domain said intermediate prediction block into a first block transformed with a first transform; and

- transforming (130) said first transformed block into a second block transformed with a second transform inverse of said first transform whose basis functions are shifted by the fractional parts of the components of said motion vector, said prediction block being obtained by deleting said at least one line of adjacent pixels and said at least one column of adjacent pixels (140) from said second transformed block.


 
2. Coding method according to claim 1, in which the size of the intermediate prediction block is a power of 2.
 
3. Coding method according to claim 1 or 2, in which said first transform is a 2D discrete cosine transform.
 
4. Method for reconstructing a pixel block comprising:

- determining (10) a prediction block for said pixel block from a motion vector for which each component comprises an integer part and a fractional part;

- decoding (22) a residue for said pixel block;

- reconstructing (24) said pixel block from said prediction block and said residue;

said reconstruction method being characterized in that the determination of said prediction block comprises the following steps:

- motion compensating (110) said pixel block with an intermediate motion vector whose components are the integer parts of the components of said motion vector;

- determining (110) an intermediate prediction block of size strictly greater than the size of said pixel block by adding at least one line of adjacent pixels and at least one column of adjacent pixels in the direction of motion to said motion compensated pixel block, said direction of motion being given by said motion vector;

- transforming (120) into a frequency domain said intermediate prediction block into a first block transformed with a first transform;

- transforming (130) said first transformed block into a second block transformed with a second transform inverse of said first transform whose basis functions are shifted by the fractional parts of the components of said motion vector, said prediction block being obtained by deleting said at least one line of adjacent pixels and said at least one column of adjacent pixels (140) from said second transformed block.


 
5. Method for reconstructing a pixel block according to claim 4, in which the size of the intermediate prediction block is a power of 2.
 
6. Method for reconstructing a pixel block according to claim 4 or 5 , in which said first transform is a 2D discrete cosine transform.
 
7. Device for coding a pixel block comprising:

- means to determine a prediction block for said pixel block from a motion vector for which each component comprises an integer part and a fractional part;

- means to calculate a residue between said pixel block and said prediction block;

- means to code said residue;

said coding device being characterized in that the determination means of said prediction block comprise:

- means for motion compensating said pixel block with an intermediate motion vector whose components are the integer parts of the components of said motion vector;

- means to determine an intermediate prediction block of size strictly greater than the size of said pixel block by adding at least one line of adjacent pixels and at least one column of adjacent pixels in the direction of motion to said motion compensated pixel block, said direction of motion being given by said motion vector;

- means to transform into a frequency domain said intermediate prediction block into a first block transformed with a first transform;

- means to transform said first transformed block into a second block transformed with a second transform inverse of said first transform whose basis functions are shifted by the fractional parts of the components of said motion vector; and

- means to extract said prediction block from said second transformed block by deleting said at least one line of adjacent pixels and said at least one column of adjacent pixels from said second transformed block.


 
8. Coding device according to claim 7, wherein the device is adapted to execute the steps of the coding method according to any of claims 1 to 3.
 
9. Device for decoding a stream comprising:

- means to determine a prediction block for a pixel block from a motion vector for which each component comprises an integer part and a fractional part;

- means to decode a residue for said pixel block from said stream;

- means to reconstruct said pixel block from said prediction block and said residue;

said decoding device being characterized in that the determination means of said prediction block comprise:

- means for motion compensating said pixel block with an intermediate motion vector whose components are the integer parts of the components of said motion vector;

- means to determine an intermediate prediction block of size strictly greater than the size of said pixel block by adding at least one line of adjacent pixels and at least one column of adjacent pixels in the direction of motion to said motion compensated pixel block, said direction of motion being given by said motion vector;

- means to transform into a frequency domain said intermediate prediction block into a first block transformed with a first transform;

- means to transform said first transformed block into a second block transformed with a second transform inverse of said first transform whose basis functions are shifted by the fractional parts of the components of said motion vector; and

- means to extract said prediction block from said second transformed block by deleting said at least one line of adjacent pixels and said at least one column of adjacent pixels from said second transformed block.


 
10. Decoding device according to claim 9, wherein the device is adapted to execute the steps of the reconstruction method according to any of claims 4 to 6.
 


Ansprüche

1. Verfahren zum Codieren eines Pixelblocks, wobei das Verfahren umfasst:

- Bestimmen (10) eines Vorhersageblocks für den Pixelblock aus einem Bewegungsvektor, für den jede Komponente einen ganzzahligen Anteil und einen gebrochenen Anteil umfasst;

- Berechnen (12) eines Rests zwischen dem Pixelblock und dem Vorhersageblock;

- Codieren (14) des Rests,

wobei das Codierungsverfahren dadurch gekennzeichnet ist, dass die Bestimmung des Vorhersageblocks umfasst:

- Bewegungskompensieren (110) des Pixelblocks mit einem Zwischenbewegungsvektor, dessen Komponenten die ganzzahligen Anteile der Komponenten des Bewegungsvektors sind;

- Bestimmen (110) eines Zwischenvorhersageblocks mit einer Größe, die streng größer als die Größe des Pixelblocks ist, durch Hinzufügen mindestens einer Zeile angrenzender Pixel und mindestens einer Spalte angrenzender Pixel in der Bewegungsrichtung zu dem bewegungskompensierten Pixelblock, wobei die Bewegungsrichtung durch den Bewegungsvektor gegeben ist;

- Transformieren (120) des Zwischenvorhersageblocks in einen ersten Block, der mit einer ersten Transformation transformiert wird, in einen Frequenzbereich; und

- Transformieren (130) des ersten transformierten Blocks in einen zweiten Block, der mit einer zweiten Transformation, die invers zu der ersten Transformation ist, deren Basisfunktionen um die gebrochenen Anteile der Komponenten des Bewegungsvektors verschoben sind, transformiert wird, wobei der Vorhersageblock durch Löschen der mindestens einen Zeile angrenzender Pixel und der mindestens einen Spalte angrenzender Pixel (140) aus dem zweiten transformierten Block erhalten wird.


 
2. Codierungsverfahren nach Anspruch 1, in dem die Größe des Zwischenvorhersageblocks eine Potenz 2 ist.
 
3. Codierungsverfahren nach Anspruch 1 oder 2, in dem die erste Transformation eine diskrete 2D-Kosinustransformation ist.
 
4. Verfahren zum Wiederherstellen eines Pixelblocks, wobei das Verfahren umfasst:

- Bestimmen (10) eines Vorhersageblocks für den Pixelblock aus einem Bewegungsvektor, für den jede Komponente einen ganzzahligen Anteil und einen gebrochenen Anteil umfasst;

- Decodieren (22) eines Rests für den Pixelblock;

- Wiederherstellen (24) des Pixelblocks aus dem Vorhersageblock und aus dem Rest;

wobei das Verfahren zum Wiederherstellen dadurch gekennzeichnet ist, dass die Bestimmung des Vorhersageblocks die folgenden Schritte umfasst:

- Bewegungskompensieren (110) des Pixelblocks mit einem Zwischenbewegungsvektor, dessen Komponenten die ganzzahligen Anteile der Komponenten des Bewegungsvektors sind;

- Bestimmen (110) eines Zwischenvorhersageblocks mit einer Größe, die streng größer als die Größe des Pixelblocks ist, durch Hinzufügen mindestens einer Zeile angrenzender Pixel und mindestens einer Spalte angrenzender Pixel in der Bewegungsrichtung zu dem bewegungskompensierten Pixelblock, wobei die Bewegungsrichtung durch den Bewegungsvektor gegeben ist;

- Transformieren (120) des Zwischenvorhersageblocks in einen ersten Block, der mit einer ersten Transformation transformiert wird, in einen Frequenzbereich;

- Transformieren (130) des ersten transformierten Blocks in einen zweiten Block, der mit einer zweiten Transformation, die invers zu der ersten Transformation ist, deren Basisfunktionen um die gebrochenen Anteile der Komponenten des Bewegungsvektors verschoben sind, transformiert wird, wobei der Vorhersageblock durch Löschen der mindestens einen Zeile angrenzender Pixel und der mindestens einen Spalte angrenzender Pixel (140) aus dem zweiten transformierten Block erhalten wird.


 
5. Verfahren zum Wiederherstellen eines Pixelblocks nach Anspruch 4, in dem die Größe des Zwischenvorhersageblocks eine Potenz von 2 ist.
 
6. Verfahren zum Wiederherstellen eines Pixelblocks nach Anspruch 4 oder 5, in dem die erste Transformation eine diskrete 2D-Kosinustransformation ist.
 
7. Vorrichtung zum Codieren eines Pixelblocks, wobei die Vorrichtung umfasst:

- Mittel zum Bestimmen eines Vorhersageblocks für den Pixelblock aus einem Bewegungsvektor, für den jede Komponente einen ganzzahligen Anteil und einen gebrochenen Anteil umfasst;

- Mittel zum Berechnen eines Rests zwischen dem Pixelblock und dem Vorhersageblock;

- Mittel zum Codieren des Rests;

wobei die Codierungsvorrichtung dadurch gekennzeichnet ist, dass die Mittel zum Bestimmen des Vorhersageblocks umfassen:

- Mittel zum Bewegungskompensieren des Pixelblocks mit einem Zwischenbewegungsvektor, dessen Komponenten die ganzzahligen Anteile der Komponenten des Bewegungsvektors sind;

- Mittel zum Bestimmen eines Zwischenvorhersageblocks mit einer Größe, die streng größer als die Größe des Pixelblocks ist, durch Hinzufügen mindestens einer Zeile angrenzender Pixel und mindestens einer Spalte angrenzender Pixel in der Bewegungsrichtung zu dem bewegungskompensierten Pixelblock, wobei die Bewegungsrichtung durch den Bewegungsvektor gegeben ist;

- Mittel zum Transformieren des Zwischenvorhersageblocks in einen ersten Block, der mit einer ersten Transformation transformiert wird, in einen Frequenzbereich; und

- Mittel zum Transformieren des ersten transformierten Blocks in einen zweiten Block, der mit einer zweiten Transformation, die invers zu der ersten Transformation ist, deren Basisfunktionen um die gebrochenen Anteile der Komponenten des Bewegungsvektors verschoben sind, transformiert wird; und

- Mittel zum Extrahieren des Vorhersageblocks aus dem zweiten transformierten Block durch Löschen der mindestens einen Zeile angrenzender Pixel und der mindestens einen Spalte angrenzender Pixel aus dem zweiten transformierten Block.


 
8. Codierungsvorrichtung nach Anspruch 7, wobei die Vorrichtung dafür ausgelegt ist, die Schritte des Verfahrens zum Codieren nach einem der Ansprüche 1 bis 3 auszuführen.
 
9. Vorrichtung zum Decodieren eines Datenstroms, wobei die Vorrichtung umfasst:

- Mittel zum Bestimmen eines Vorhersageblocks für einen Pixelblock aus einem Bewegungsvektor, für den jede Komponente einen ganzzahligen Anteil und einen gebrochenen Anteil umfasst;

- Mittel zum Decodieren eines Rests für den Pixelblock aus dem Datenstrom;

- Mittel zum Wiederherstellen des Pixelblocks aus dem Vorhersageblock und aus dem Rest;

wobei die Decodierungsvorrichtung dadurch gekennzeichnet ist, dass die Mittel zum Bestimmen des Vorhersageblocks umfassen:

- Mittel zum Bewegungskompensieren des Pixelblocks mit einem Zwischenbewegungsvektor, dessen Komponenten die ganzzahligen Anteile der Komponenten des Bewegungsvektors sind;

- Mittel zum Bestimmen eines Zwischenvorhersageblocks mit einer Größe, die streng größer als die Größe des Pixelblocks ist, durch Hinzufügen mindestens einer Zeile angrenzender Pixel und mindestens einer Spalte angrenzender Pixel in der Bewegungsrichtung zu dem bewegungskompensierten Pixelblock, wobei die Bewegungsrichtung durch den Bewegungsvektor gegeben ist;

- Mittel zum Transformieren des Zwischenvorhersageblocks in einen ersten Block, der mit einer ersten Transformation transformiert wird, in einen Frequenzbereich;

- Mittel zum Transformieren des ersten transformierten Blocks in einen zweiten Block, der mit einer zweiten Transformation, die invers zu der ersten Transformation ist, deren Basisfunktionen um die gebrochenen Anteile der Komponenten des Bewegungsvektors verschoben sind, transformiert wird; und

- Mittel zum Extrahieren des Vorhersageblocks aus dem zweiten transformierten Block durch Löschen der mindestens einen Zeile angrenzender Pixel und der mindestens einen Spalte angrenzender Pixel aus dem zweiten transformierten Block.


 
10. Decodierungsvorrichtung nach Anspruch 9, wobei die Vorrichtung dafür ausgelegt ist, die Schritte des Verfahrens zum Wiederherstellen nach einem der Ansprüche 4 bis 6 auszuführen.
 


Revendications

1. Procédé de codage d'un bloc de pixels, comprenant :

- la détermination (10) d'un bloc de prédiction pour ledit bloc de pixels à partir d'un vecteur de mouvement 5 pour lequel chaque composant comprend une partie entière et une partie fractionnelle ;

- le calcul (12) d'un résidu entre ledit bloc de pixels et ledit bloc de prédiction ;

- le codage (14) dudit résidu,

ledit procédé de codage étant caractérisé en ce que la détermination dudit bloc de prédiction comprend :

- un mouvement compensant (110) ledit bloc de pixels avec un vecteur de mouvement intermédiaire dont les composants sont les parties entières des composants dudit vecteur de mouvement ;

- la détermination (110) d'un bloc de prédiction intermédiaire de taille strictement supérieure à la taille dudit bloc de pixels par l'ajout d'au moins une ligne de pixels adjacents et d'au moins une colonne de pixels adjacents dans la direction du mouvement dudit bloc de pixels compensé par le mouvement, ladite direction du mouvement étant donnée par ledit vecteur de mouvement ;

- la transformation (120) en un domaine de fréquence dudit bloc de prédiction intermédiaire dans un premier bloc transformé avec une première transformation ; et

- la transformation (130) dudit premier bloc transformé en un second bloc transformé avec une seconde transformation inverse de ladite première transformation dont les fonctions de base sont modifiées par les parties fractionnelles des composants dudit vecteur de mouvement, ledit bloc de prédiction étant obtenu par la suppression desdites lignes de pixels adjacents et desdites colonnes de pixels adjacents (140) dudit second bloc transformé.


 
2. Procédé de codage selon la revendication 1, dans lequel la taille du bloc de prédiction intermédiaire est une puissance de 2.
 
3. Procédé de codage selon la revendication 1 ou 2, dans lequel ladite première transformation est une transformation en cosinus discrète 2D.
 
4. Procédé de reconstruction d'un bloc de pixels, comprenant :

- la détermination (10) d'un bloc de prédiction pour ledit bloc de pixels à partir d'un vecteur de mouvement pour lequel chaque composant comprend une partie entière et une partie fractionnelle ;

- le décodage (22) d'un résidu pou ledit bloc de pixels ;

- la reconstruction (24) dudit bloc de pixels à partir dudit bloc de prédiction et dudit résidu ;

ledit procédé de reconstruction étant caractérisé en ce que la détermination dudit bloc de prédiction comprend les étapes suivantes :

- un mouvement compensant (110) ledit bloc de pixels avec un vecteur de mouvement intermédiaire dont les composants sont les parties entières des composants dudit vecteur de mouvement ;

- la détermination (110) d'un bloc de prédiction intermédiaire de taille strictement supérieure à la taille dudit bloc de pixels par l'ajout d'au moins une ligne de pixels adjacents et d'au moins une colonne de pixels adjacents dans la direction du mouvement dudit bloc de pixels compensé par le mouvement, ladite direction du mouvement étant donnée par ledit vecteur de mouvement ;

- la transformation (120) en un domaine de fréquence dudit bloc de prédiction intermédiaire dans un premier bloc transformé avec une première transformation ;

- la transformation (130) dudit premier bloc transformé en un second bloc transformé avec une seconde transformation inverse de ladite première transformation dont les fonctions de base sont modifiées par les parties fractionnelles des composants dudit vecteur de mouvement, ledit bloc de prédiction étant obtenu par la suppression desdites lignes de pixels adjacents et desdites colonnes de pixels adjacents (140) dudit second bloc transformé.


 
5. Procédé de reconstruction d'un bloc de pixels selon la revendication 4, dans lequel la taille du bloc de prédiction intermédiaire est une puissance de 2.
 
6. Procédé de reconstruction d'un bloc de pixels selon la revendication 4 ou 5, dans lequel ladite première transformation est une transformation en cosinus discrète 2D.
 
7. Dispositif de codage d'un bloc de pixels, comprenant :

- un moyen permettant de déterminer un bloc de prédiction pour ledit bloc de pixels à partir d'un vecteur de mouvement pour lequel chaque composant comprend une partie entière et une partie fractionnelle ;

- un moyen permettant de calculer un résidu entre ledit bloc de pixels et ledit bloc de prédiction ;

- un moyen permettant de coder ledit résidu ;

ledit dispositif de codage étant caractérisé en ce que la détermination dudit bloc de prédiction 5 comprend :

- un moyen permettant de compenser le mouvement dudit bloc de pixels avec un vecteur de mouvement intermédiaire dont les composants sont des parties entières des composants dudit vecteur de mouvement ;

- un moyen permettant de déterminer un bloc de prédiction intermédiaire de taille strictement supérieure à la taille dudit bloc de pixels par l'ajout d'au moins une ligne de pixels adjacents et d'au moins une colonne de pixels adjacents dans la direction du mouvement dudit bloc de pixels compensé par le mouvement, ladite direction du mouvement étant donnée par ledit vecteur de mouvement ;

- un moyen permettant de transformer en un domaine de fréquence ledit bloc de prédiction intermédiaire en un premier bloc transformé avec une première transformation ;

- un moyen permettant de transformer ledit premier bloc transformé en un second bloc transformé avec une seconde transformation inverse de ladite première transformation dont les fonctions de base sont modifiées par les parties fractionnelles des composants dudit vecteur de mouvement ; et

- un moyen permettant d'extraire ledit bloc de prédiction dudit second bloc transformé en supprimant lesdites lignes de pixels adjacents et lesdites colonnes de pixels adjacents dudit second bloc transformé.


 
8. Dispositif de codage selon la revendication 7, dans lequel le dispositif est adapté pour exécuter les étapes du procédé de codage conformément à l'une des revendications 1 à 3.
 
9. Dispositif permettant de décoder un flux comprenant :

- un moyen permettant de déterminer un bloc de prédiction pour un bloc de pixels à partir d'un vecteur de mouvement pour lequel chaque composant comprend une partie entière et une partie fractionnelle ;

- un moyen permettant de décoder un résidu pour ledit bloc de pixels à partir dudit flux ;

- un moyen permettant de reconstruire ledit bloc de pixels dudit bloc de prédiction et dudit résidu ;

ledit dispositif de décodage étant caractérisé en ce que la détermination dudit bloc de prédiction comprend :

- un moyen permettant de compenser le mouvement dudit bloc de pixels avec un vecteur de mouvement intermédiaire dont les composants sont les parties entières des composants dudit vecteur de mouvement ;

- un moyen permettant de déterminer un bloc de prédiction intermédiaire de taille strictement supérieure à la taille dudit bloc de pixels par l'ajout d'au moins une ligne de pixels adjacents et d'au moins une colonne de pixels adjacents dans la direction du mouvement dudit bloc de pixels compensé par le mouvement, ladite direction du mouvement étant donnée par ledit vecteur de mouvement ;

- un moyen permettant de transformer en un domaine de fréquence ledit bloc de prédiction intermédiaire en un premier bloc transformé avec une première transformation ;

- un moyen permettant de transformer ledit premier bloc transformé en un second bloc transformé avec une seconde transformation inverse de ladite première transformation dont les fonctions de base sont modifiées par les parties fractionnelles des composants dudit vecteur de mouvement ; et

- un moyen permettant d'extraire ledit bloc de prédiction dudit second bloc transformé en supprimant lesdites lignes de pixels adjacents et lesdites colonnes de pixels adjacents dudit second bloc transformé.


 
10. Dispositif de décodage selon la revendication 9, dans lequel le dispositif est adapté pour exécuter les étapes du procédé de reconstruction selon l'une des revendications 4 à 6.
 




Drawing




















Cited references

REFERENCES CITED IN THE DESCRIPTION



This list of references cited by the applicant is for the reader's convenience only. It does not form part of the European patent document. Even though great care has been taken in compiling the references, errors or omissions cannot be excluded and the EPO disclaims all liability in this regard.

Patent documents cited in the description




Non-patent literature cited in the description