CN103826131B - Picture decoding method and picture decoding apparatus - Google Patents

Abstract

A kind of picture decoding method is it is characterised in that possess：The step selecting at least one motion reference block from the block of pixels that the decoding with movable information finishes；The step selecting at least one available block from described motion reference block, described available block is the block of pixels of the candidate with the movable information being applied to decoder object block, and has mutually different movable information；The code table set in advance with reference to the quantity according to described available block, is decoded to the coded data being inputted, thus the step obtaining the selection information for determining selection block；According to described selection information, the step selecting a selection block from described available block；The step generating the prognostic chart picture of described decoder object block using the movable information of described selection block；The step forecast error of described decoder object block being decoded according to described coded data；And the step decoding image is obtained according to described prognostic chart picture and described forecast error.

Description

Picture decoding method and picture decoding apparatus

The present invention be submit on April 23rd, 2013 Application No. " 201310142052.8 ", invention entitled " image solution The divisional application of the application of code method and picture decoding apparatus ", its original female case is to enter Chinese state on October 08th, 2012 Family the stage, national applications number be " 201080066017.7 ", invention entitled " method for encoding images and picture decoding side The application of method ".

Technical field

The present invention relates to the coding for moving image and rest image and coding/decoding method.

Background technology

In recent years, in ITU-T and ISO/IEC, all as ITU-T Rec.H.264 and ISO/IEC14496-10（With Under, referred to as H.264）And suggested the dynamic image encoding method increasing substantially code efficiency.In h .264, prediction process, Conversion process and entropy code are processed with rectangle block unit（For example, 16 × 16 pixel block units, 8 × 8 pixel block units etc.）Enter OK.In prediction is processed, for the rectangular block of coded object（Coded object block）, carry out with reference to the encoded frame finishing（Ginseng According to frame）And carry out the motion compensation of the prediction of time orientation.In such motion compensation, need to the fortune including motion vector Dynamic information is encoded and is sent to decoding side, and this motion vector is as coded object block and the block in the reference of reference frame in institute Between displacement information spatially vector.In addition, in the case of carrying out motion compensation using multiple reference frame, it is right to need Movable information is encoded together with reference to frame number.Therefore, have with movable information and with reference to the relevant encoding amount of frame number Shi Zeng great.

One as the method obtaining motion vector in motion compensated prediction, with good grounds distributes to encoded finishing The motion vector of block derive and distribute to the motion vector of coded object block, and generated pre- according to the motion vector derived The Direct Model of altimetric image（With reference to patent documentation 1 and patent documentation 2）.In Direct Model, because not entering to motion vector Row coding, it is possible to reduce the encoding amount of movable information.H.264/AVC, Direct Model is for example adopted.

Patent documentation

Patent documentation 1:Japanese Patent No. 4020789

Patent documentation 2:Rice state patent the 7233621st

Content of the invention

In Direct Model, using the median of the motion vector of the block being finished according to the coding adjoining with coded object block The motion vector to predict generation coded object block for the method for the fixation of calculating motion vector.Therefore, motion vector computation from Low by spending.

In order to improve the degree of freedom calculating motion vector, propose to have to finish block from multiple codings and select one and to coding The method of object block assigned motion vector.In the method it is necessary to always send the selection information determining selected piece, so that Decoding side can determine the block that selected coding finishes.Therefore, when from multiple codings finish in block select one and determine should When this distributes to the motion vector of coded object block, there is a problem of increased and select information-related encoding amount.

The present invention is to make to solve the above problems it is therefore intended that providing code efficiency high method for encoding images And picture decoding method.

The method for encoding images of one embodiment of the present invention possesses following steps:Finish from the coding with movable information Block of pixels in select the 1st step of at least one motion reference block;Select at least one sharp from above-mentioned motion reference block With the second step of block, this available block is the block of pixels of the candidate with the movable information being applied to coded object block, and has Mutually different movable information;The third step of a selection block is selected from above-mentioned available block;Fortune using above-mentioned selection block Dynamic information generates the 4th step of the prognostic chart picture of above-mentioned coded object block;To the prediction between above-mentioned prognostic chart picture and original image The 5th step that error is encoded;And with reference to the code table set in advance according to the quantity of above-mentioned available block, to determination The 6th step that the selection information of above-mentioned selection block is encoded.

The picture decoding method of another embodiment of the present invention possesses following steps:Complete from the decoding with movable information The 1st step of at least one motion reference block is selected in complete block of pixels;Select at least one can from above-mentioned motion reference block Using the second step of block, this available block is the block of pixels of the candidate with the movable information being applied to decoder object block, and has There is mutually different movable information;The code table set in advance with reference to the quantity according to above-mentioned available block, to inputted Coded data is decoded, thus obtaining the third step of the selection information for determining selection block;According to above-mentioned selection information, The 4th step of a selection block is selected from above-mentioned available block;Movable information using above-mentioned selection block generates above-mentioned decoding 5th step of the prognostic chart picture of object block;Decode the 6th step of the forecast error of above-mentioned decoder object block according to above-mentioned coded data Suddenly;And the 7th step of decoding image is obtained according to above-mentioned prognostic chart picture and above-mentioned forecast error.

In accordance with the invention it is possible to raising code efficiency.

Brief description

Fig. 1 is the block diagram of the structure of picture coding device roughly illustrating the 1st embodiment.

Fig. 2A is to illustrate that the process unit of the coding in picture decoding portion shown in Fig. 1 is the figure of of the size of microlith.

Fig. 2 B is the process unit of the coding in picture decoding portion shown in Fig. 1 is the figure of another of size of microlith.

Fig. 3 is the figure illustrating the picture coding portion shown in Fig. 1 to the order that the block of pixels of coded object frame in is encoded.

Fig. 4 is the figure of of the movable information frame illustrating that the movable information memorizer shown in Fig. 1 is kept.

Fig. 5 is the flow chart illustrating to process of the order of received image signal of Fig. 1.

Fig. 6 A is the figure of of the interaction prediction process performed by the dynamic compensating unit illustrating Fig. 1.

Fig. 6 B is the figure of another of the interaction prediction process performed by the dynamic compensating unit illustrating Fig. 1.

Fig. 7 A is to illustrate that interaction prediction processes the figure of of the used size of motion compensation block.

Fig. 7 B is to illustrate that interaction prediction processes the figure of another of the used size of motion compensation block.

Fig. 7 C is to illustrate that interaction prediction processes the figure of another other example of the used size of motion compensation block.

Fig. 7 D is to illustrate that interaction prediction processes the figure of another of the used size of motion compensation block.

Fig. 8 A is the figure of of the configuration illustrating direction in space and time orientation motion reference block.

Fig. 8 B is the figure of another of the configuration illustrating direction in space motion reference block.

Fig. 8 C is the figure of the relative position illustrating direction in space motion reference block with respect to the coded object block shown in Fig. 8 B.

Fig. 8 D is the figure of another of the configuration illustrating time orientation motion reference block.

Fig. 8 E is the figure of another other example of the configuration illustrating time orientation motion reference block.

Fig. 8 F is the figure of another other example of the configuration illustrating time orientation motion reference block.

Fig. 9 is to illustrate that the available block obtaining section of Fig. 1 selects to utilize of the method for block from motion reference block Flow chart.

Figure 10 is of the available block selecting the motion reference block shown in from Fig. 8 for the method illustrating according to Fig. 9 Figure.

Figure 11 is the figure of of the available block message of available block obtaining section output illustrating Fig. 1.

Figure 12 A is illustrate the same sex determination of movable information between the block that carried out by the available block obtaining section of Fig. 1 one The figure of example.

Figure 12 B is illustrate the same sex determination of movable information between the block that carried out by the available block obtaining section of Fig. 1 another The figure of one.

Figure 12 C is illustrate the same sex determination of movable information between the block that carried out by the available block obtaining section of Fig. 1 another The figure of one other examples.

Figure 12 D is illustrate the same sex determination of movable information between the block that carried out by the available block obtaining section of Fig. 1 another The figure of one.

Figure 12 E is illustrate the same sex determination of movable information between the block that carried out by the available block obtaining section of Fig. 1 another The figure of one other examples.

Figure 12 F is illustrate the same sex determination of movable information between the block that carried out by the available block obtaining section of Fig. 1 another The figure of one.

Figure 13 is the block diagram of the structure of the prediction section roughly illustrating Fig. 1.

Figure 14 is the figure of the group of movable information illustrating that the time orientation movable information obtaining section of Figure 13 is exported.

Figure 15 is that explanation can be utilized a few pixels precision in the motion compensation process based on the dynamic compensating unit of Figure 13 Interpolation processing explanatory diagram.

Figure 16 is the flow chart of of the action of the prediction section illustrating Figure 13.

Figure 17 is to illustrate that the dynamic compensating unit of Figure 13 is right to encoding by the copying motion information of time orientation motion reference block Figure as the situation of block.

Figure 18 is the block diagram of the structure of the variable length code portion roughly illustrating Fig. 1.

Figure 19 is that basis can utilize block message generative grammar（syntax）Example figure.

Figure 20 is the figure of the example of 2 values illustrating selection block Message Syntax corresponding with available block message.

Figure 21 is the proportional zoom that movable information is described（scaling）Explanatory diagram.

Figure 22 is the figure of the syntactic constructs according to embodiment.

Figure 23 A is the figure of of the microlith layer grammer according to the 1st embodiment.

Figure 23 B is the figure of another of the microlith layer grammer according to the 1st embodiment.

Figure 24 A be illustrate with H.264 in B section when the corresponding code table of mb_type and mb_type figure.

Figure 24 B is the figure of of the code table illustrating embodiment.

Figure 24 C be illustrate with H.264 in P section when the corresponding code table of mb_type and mb_type figure.

Figure 24 D is the figure of another of the code table illustrating embodiment.

Figure 25 A is to illustrate according to embodiment and of the corresponding code table of mb_type and mb_type in B section The figure of example.

Figure 25 B is the another of the corresponding code table of mb_type and mb_type in illustrating to cut into slices with P according to embodiment The figure of one.

Figure 26 is the block diagram of the structure of picture coding device roughly illustrating the 2nd embodiment.

Figure 27 is the block diagram of the structure of the prediction section roughly illustrating Figure 26.

Figure 28 is the block diagram of the structure of the 2nd prediction section roughly illustrating Figure 27.

Figure 29 is the block diagram of the structure of the variable length code portion roughly illustrating Figure 26.

Figure 30 A is the figure of that illustrates microlith layer grammer according to the 2nd embodiment.

Figure 30 B is the figure of another that illustrates microlith layer grammer according to the 2nd embodiment.

Figure 31 is the block diagram of the picture decoding apparatus roughly illustrating the 3rd embodiment.

Figure 32 is the block diagram illustrating in greater detail the coding row lsb decoder shown in Figure 31.

Figure 33 is the block diagram illustrating in greater detail the prediction section shown in Figure 31.

Figure 34 is the block diagram of the picture decoding apparatus roughly illustrating the 4th embodiment.

Figure 35 is the block diagram illustrating in greater detail the coding row lsb decoder shown in Figure 33.

Figure 36 is the block diagram illustrating in greater detail the prediction section shown in Figure 33.

The explanation of reference

10:Received image signal;11:Prediction image signal;12:Prediction error image signal;13:Quantization transformation series Number;14:Coded data;15:Decoding predictive error signal;16:Local decoder picture signal;17:With reference to picture signal;18:Fortune Dynamic information;20:Bit stream;21:Movable information;25,26:Information frame;30:Available block message;31:Select block message;32:In advance Survey handover information;33:Conversion coefficient information;34:Predictive error signal;35:Prediction image signal;36:Decoded image signal; 37:With reference to picture signal;38:Movable information;39:Reference movement information;40:Movable information;50:Encoding control information;51:Instead Feedforward information;60:Available block message;61:Select block message;62:Prediction handover information;70:Decoding control information;71:Control Information;80:Coded data;100:Picture coding portion;101:Prediction section;102:Subtractor;103:Conversion quantization portion;104: Variable length code portion;105:Inverse guantization (IQ) inverse transformation portion;106:Adder;107:Frame memory;108:Information storage unit; 109:Available block obtaining section;110:Direction in space movable information obtaining section;111:Time orientation movable information obtaining section;112: Information switching switch;113:Dynamic compensating unit;114:Parameter coding portion;115:Transform coefficients encoding portion;116:Selection block encodes Portion;117:Multiplexing portion;118:Movable information selector;120:Output state;150:Coding-control portion;200:Image is compiled Code portion;201:Prediction section;202:2nd prediction section;203:Forecasting Methodology switching switch;204:Variable length code portion;205:Fortune Dynamic information acquiring section;216:Selection block encoding section;217:Movable information encoding section;300:Picture decoding portion;301:Coding row solution Code portion;301:Coding row lsb decoder;302:Inverse guantization (IQ) inverse transformation portion;303:Adder;304:Frame memory;305:Prediction Portion;306:Information storage unit;307:Available block obtaining section;308:Output state;310:Direction in space movable information obtains Portion;311:Time orientation movable information obtaining section;312:Movable information switching switch;313:Dynamic compensating unit;314:Information is selected Select portion;320:Separated part;321:Parameter lsb decoder;322:Conversion coefficient lsb decoder;323:Selection block lsb decoder;350:Decoding control Portion processed;400:Picture decoding portion;401:Coding row lsb decoder;405:Prediction section;410:2nd prediction section;411:Forecasting Methodology is cut Change switch;423:Selection block lsb decoder;424:Information lsb decoder;901:High-level syntax;902:Sequential parameter group grammer;903: Graphic parameter group grammer;904:Slice-level grammer;905:Section head grammer;906:Slice of data grammer;907:Microlith level grammer; 908:Microlith layer grammer;909:Microlith predicts grammer

Specific embodiment

Hereinafter, as needed, referring to the drawings to the picture coding of embodiments of the present invention and the method for picture decoding And device illustrates.In addition, in the following embodiments, the part with regard to giving same numbering is set to carry out same The part of action, eliminates the explanation of repetition.

（1st embodiment）

Fig. 1 roughly illustrates the structure of the picture coding device of the 1st embodiment of the present invention.As shown in figure 1, this image Code device possesses picture coding portion 100, coding-control portion 150 and output state 120.This picture coding device both may be used To realize by hardware such as LSI chips, or can also be set to realize by making computer execute image encoding program.

In picture coding portion 100, for example, inputted as moving image or quiet using the pixel block unit having split original image The only original image of image（Received image signal）10.As described in detail afterwards, picture coding portion 100 is to received image signal 10 are compressed encoding and generate coded data 14.The coded data 14 being generated is temporarily stored to output state 120, is compiling The output timing of code control unit 150 management, to storage system (not shown)（Storage medium）Or Transmission system（Communication line）Send Go out.

Coding-control portion 150 controls the feedback control, quantization control, predictive mode control and the entropy that produce encoding amount to compile Code controls whole coded treatment in such picture coding portion 100.Specifically, coding-control is believed by coding-control portion 150 Breath 50 is supplied to picture coding portion 100, suitably receives feedback information 51 from picture coding portion 100.Encoding control information 50 wraps Include information of forecasting, movable information 18 and quantum-chemical descriptors information etc..Information of forecasting includes prediction mode information and block size Information.Movable information 18 includes motion vector, with reference to frame number and prediction direction（One direction prediction, twocouese prediction）.Amount Sonization parameter information includes quantization width（Quantization step size）Deng quantum-chemical descriptors and quantized matrix.Feedback information 51 include the generation encoding amount based on picture coding portion 100, for example, use when determining quantum-chemical descriptors.

Picture coding portion 100 is to split what original image obtained block of pixels（For example, microlith, sub-block, 1 pixel etc.）For list Position, encodes to received image signal 10.Therefore, received image signal 10 with split the pixel block unit of original image by according to Secondary it is input to picture coding portion 100.In the present embodiment, the process unit of coding is set to microlith, will believe with input picture Number 10 corresponding, block of pixels as coded object（Microlith）Simply referred to as coded object block.In addition, coded object block will be included The picture frame of picture frame, i.e. coded object is referred to as coded object frame.

Such coded object block, for example, both can be 16 × 16 block of pixels as shown in Fig. 2A it is also possible to Fig. 2 B institute Show such 64 × 64 block of pixels.In addition, coded object block can also be 32 × 32 block of pixels, 8 × 8 block of pixels etc..In addition, it is micro- The shape of block is not limited to the example of the square shape as shown in Fig. 2A and Fig. 2 B it is also possible to be set to the arbitrary shapes such as rectangular shape Shape.In addition, above-mentioned process unit is not limited to block of pixels as microlith or frame or field.

In addition, the coded treatment for each block of pixels of coded object frame in can execute in any order.In this reality Apply in mode, for the purpose of simplifying the description, as shown in figure 3, being set to the block of pixels to bottom right for the block of pixels from the upper left of coded object frame Line by line, i.e. according to raster scan order, coded treatment is executed to block of pixels.

Picture coding portion 100 shown in Fig. 1 possesses：Prediction section 101, subtractor 102, conversion quantization portion 103, variable Length coding portion 104, inverse guantization (IQ) inverse transformation portion 105, adder 106, frame memory 107, movable information memorizer 108 with And available block obtaining section 109.

In picture coding portion 100, received image signal 10 is imported into prediction section 101 and subtractor 102.Subtractor 102 receives input picture signals 10, and receive prediction image signal 11 from prediction section 101 described later.Subtractor 102 calculates defeated Enter the difference of picture signal 10 and prediction image signal 11, generate prediction error image signal 12.

Conversion quantization portion 103 receives prediction error image signal 12 from subtractor 102, to the prediction-error image receiving As signal 12 implements conversion process, generate conversion coefficient.Conversion process, for example, is discrete cosine transform（DCT:Discrete Cosine Transform）Deng orthogonal transformation.In another embodiment, conversion quantization portion 103 can also substitute discrete Cosine transform and using wavelet transformation and isolated component parsing etc. method generate conversion coefficient.In addition, conversion quantization Portion 103 carries out quantization according to the quantum-chemical descriptors being provided by coding-control portion 150 to the conversion coefficient being generated.By quantization Conversion coefficient afterwards（Conversion coefficient information）13 export to variable length code portion 104 and inverse guantization (IQ) inverse transformation portion 105.

Inverse guantization (IQ) inverse transformation portion 105 according to the quantum-chemical descriptors being provided by coding-control portion 150, i.e. with converted quantity Sub-ization portion 103 identical quantum-chemical descriptors, carry out inverse guantization (IQ) to the conversion coefficient 13 after quantization.Then, inverse guantization (IQ) Inverse transformation portion 105 implements inverse transformation to the conversion coefficient after inverse guantization (IQ), generates decoding predictive error signal 15.Based on inverse quantum The inversion process changing inverse transformation portion 105 is consistent with the inversion process of the conversion process based on conversion quantization portion 103. For example, inversion process is inverse discrete cosine transform（IDCT:Inverse Discrete Cosine Transform）Or it is inverse little Wave conversion etc..

Adder 106 receives decoding predictive error signal 15 from inverse guantization (IQ) inverse transformation portion 105, in addition, from prediction section 101 reception prediction image signals 11.Adder 106 will decode predictive error signal 15 and be added with prediction image signal 11 and generate Local decoder picture signal 16.The local decoder picture signal 16 being generated is saved as reference picture picture in frame memory 107 Signal 17.The reference picture signal 17 that frame memory 107 is preserved, when encoding to coded object block thereafter, by pre- Survey portion 101 reads and reference.

Prediction section 101 receives with reference to picture signal 17 from frame memory 107, and from available block obtaining section described later 109 receptions can utilize block message 30.In addition, prediction section 101 receives reference movement information from movable information memorizer 108 described later 19.Prediction section 101 generates coded object block according to reference to picture signal 17, reference movement information 19 and available block message 30 Prediction image signal 11, movable information 18 and select block message 31.Specifically, prediction section 101 possesses：According to available Block message 30 and reference movement information 19 generate movable information 18 and the movable information selector 118 of selection block message 31； And the dynamic compensating unit 113 of prediction image signal 11 is generated according to movable information 18.Prediction image signal 11 is fed to subtraction Device 102 and adder 106.Movable information 18 is stored in movable information memorizer 108, for right for coding thereafter Prediction as block is processed.In addition, selecting block message 31 to be fed to variable length code portion 104.With regard to prediction section 101 rear detailed Explanation.

In movable information memorizer 108 interim preserve movable information 18 and as reference movement information 19.Fig. 4 illustrates to move One of the structure of information storage unit 108.As shown in figure 4, movable information memorizer 108 maintains reference movement letter with frame unit Breath 19, reference movement information 19 is formed with movable information frame 25.The movable information 18 relevant with the block that coding finishes is by defeated successively Enter movable information memorizer 108, as a result, movable information memorizer 108 keeps scramble time different multiple movable informations Frame 25.

Reference movement information 19 is with fixed block unit（For example, 4 × 4 pixel block unit）It is maintained at movable information frame 25 Interior.The pixel that motion vector block 28 shown in Fig. 4 represents with coded object block, the formed objects such as block and selection block can be utilized Block, for example, is 16 × 16 block of pixels.In motion vector block 28, for example, it is assigned motion vector for every 4 × 4 block of pixels.By profit Processed with the interaction prediction of motion vector block and be referred to as motion vector block prediction process.When generating movable information 18, motion letter The reference movement information 19 that breath memorizer 108 is kept is read by prediction section 101.The motion that aftermentioned such available block has Information 18 refers to the reference movement information 19 that the region that the available block in movable information memorizer 108 is located at keeps.

In addition, movable information memorizer 108 is not limited to keep the example of reference movement information 19 with 4 × 4 pixel block units, Reference movement information 19 can also be kept with other pixel block units.For example, relevant with reference movement information 19 block of pixels list Position both can be 1 pixel or 2 × 2 block of pixels.In addition, the shape of the block of pixels relevant with reference movement information 19 is not It is limited to the example of square shape, arbitrary shape can be set to.

The available block obtaining section 109 of Fig. 1 obtains reference movement information 19 from movable information memorizer 108, according to acquirement Reference movement information 19, be selected to from encoded multiple pieces completing prediction section 101 prediction process in utilize Available block.Selected available block can be given prediction section 101 using block message 30 by conduct and variable-length is compiled Code portion 104.To become for selecting to be referred to as motion reference block using the block that the coding of the candidate of block finishes.With regard to motion reference Block and the system of selection of available block, after be described in detail.

Variable length code portion 104, in addition to conversion coefficient information 13, also receives from prediction section 101 and selects block message 31, receive the coding parameter of information of forecasting and quantum-chemical descriptors etc. from coding-control portion 150, from available block obtaining section 109 Reception can utilize block message 30.Variable length code portion 104 to the conversion coefficient 13 after quantization, select block message 31, can profit Carry out entropy code with block message 30 and coding parameter（For example, fixed-length coding, Huffman encoding or arithmetic coding etc.）, generate and compile Code data 14.Coding parameter includes selecting block message 31 and information of forecasting, and include the information relevant with conversion coefficient and The all of parameter required in decoding such as relevant information of quantization.The coded data 14 being generated is temporarily stored in output Buffer 120, and it is fed to storage system (not shown) or Transmission system.

Fig. 5 illustrates the processing routine of received image signal 10.As shown in figure 5, first, prediction is generated by prediction section 101 Picture signal 11（Step S501）.In the generation of the prediction image signal 11 of step S501, by available block described later One available block is chosen as selection block, and using the movable information selecting block message 31, selection block to have and reference picture As signal 17 makes prediction image signal 11.Prediction image signal 11 and received image signal 10 are calculated by subtractor 102 Difference, generates prediction error image signal 12（Step S502）.

Then, by converting quantization portion 103, orthogonal transformation and quantization are implemented to prediction error image signal 12, Generate conversion coefficient information 13（Step S503）.Conversion coefficient information 13 and selection block message 31 are fed to variable length code Portion 104, is carried out variable length code, and generates coded data 14（Step S504）.In addition, in step S504, according to choosing Select block message 31 and carry out switch code table, so that having the entry with the quantity equal amount of available block in code table and right Block message 31 is selected to carry out variable length code.The bit stream 20 of coded data is fed to storage system system (not shown) or passes Defeated path.

Conversion coefficient information 13 generated in step S503 carries out inverse quantum by inverse guantization (IQ) inverse transformation portion 105 Change, and be carried out inversion process, become decoding predictive error signal 15（Step S505）.Predictive error signal 15 will be decoded add With reference in picture signal 17, become local decoder picture signal 16 to used in step S501（Step S506）, and conduct It is stored in frame memory 107 with reference to picture signal（Step S507）.

Then, each structure in above-mentioned picture coding portion 100 is described in detail.

The picture coding portion 100 of Fig. 1 has prepared multiple predictive modes in advance, the prediction image signal 11 of each predictive mode Generation method and motion compensation block size are mutually different.The method generating prediction image signal 11 as prediction section 101, tool For body, if substantially dividing, have and use and coded object frame（Or field）Relevant reference picture signal 17 generates prediction The interior prediction of image（Infra-frame prediction）, with using the reference frame that finishes with more than one coding（Referential field）Relevant reference Picture signal 17 generates the interaction prediction of prognostic chart picture（Inter prediction）.Prediction section 101 optionally switches interior prediction and friendship Inter prediction, generates the prediction image signal 11 of coded object block.

Fig. 6 A illustrates of interaction prediction being carried out by dynamic compensating unit 113.In interaction prediction, as shown in Figure 6A, According to as the block of reference frame in before encoded 1 frame completing and the block with coded object block same position（Also referred to as predict Block）23, the block 24 of the position using with the motion vector 18a according to included by movable information 18 and spatially displacement is relevant Reference picture signal 17, generate prediction image signal 11.That is, in the generation of prediction image signal 11, using coded object The position of block（Coordinate）, and have with by the block 24 with reference to frame in that motion vector 18a included by movable information 18 determines The reference picture signal 17 closed.In interaction prediction, a few pixels precision can be carried out（For example, 1/2 pixel precision or 1/4 picture Plain precision）Motion compensation, by reference to picture signal 17 be filtered process, generate inter polated pixel value.For example, exist H.264, in, luminance signal can be proceeded to the interpolation processing till 1/4 pixel precision.In the motion carrying out 1/4 pixel precision In the case of compensation, the quantity of information of movable information 18 becomes 4 times of integer-pel precision.

In addition, in interaction prediction, the example of the reference frame before being not limited to using 1 frame as shown in Fig. 6 A, such as Fig. 6 B Shown, it is possible to use arbitrarily to encode the reference frame finishing.Relevant in the multiple reference frame maintaining different from time location With reference in the case of picture signal 17, represent which time location prediction image signal is generated with reference to picture signal 17 according to 11 information to illustrate by referring to frame number.It is contained in movable information 18 with reference to frame number.Can be with region with reference to frame number Unit（Figure, block unit etc.）Changed.That is, different reference frame can be used for each block of pixels.As one, when In the case of employing the reference frame before 1 frame that finishes of coding in prediction, this region be set to 0 with reference to frame number, when pre- In the case of employing the reference frame before 2 frames that finish of coding in survey, this region be set to 1 with reference to frame number.As it Its example, when the reference picture signal 17 maintaining 1 frame amount in frame memory 107（The quantity of reference frame is 1）In the case of, reference frame Number is always set at 0.

In addition, in interaction prediction, can select to be suitable to the block size of coded object block among multiple motion compensation blocks. I.e. it is also possible to coded object block is divided into multiple small pixel blocks, and carry out motion compensation for each small pixel block.Fig. 7 A is extremely Fig. 7 C illustrates the size of the motion compensation block of microlith unit, and Fig. 7 D illustrates sub-block（Block of pixels below 8 × 8 pixels）The fortune of unit The size of dynamic compensation block.As shown in Figure 7 A, in the case that coded object block is 64 × 64 pixels, as motion compensation block, energy Enough select 64 × 64 block of pixels, 64 × 32 block of pixels, 32 × 64 block of pixels or 32 × 32 block of pixels etc..In addition, as shown in Figure 7 B, In the case that coded object block is 32 × 32 pixels, 32 × 32 block of pixels, 32 × 16 pictures can be selected as motion compensation block Plain block, 16 × 32 block of pixels or 16 × 16 block of pixels etc..In addition, as seen in figure 7 c, when coded object block is 16 × 16 pixels In the case of, motion compensation block can be set as 16 × 16 block of pixels, 16 × 8 block of pixels, 8 × 16 block of pixels or 8 × 8 pixels Block etc..In addition, as illustrated in fig. 7d, in the case that coded object block is 8 × 8 pixels, motion compensation block can select 8 × 8 pictures Plain block, 8 × 4 block of pixels, 4 × 8 block of pixels or 4 × 4 block of pixels etc..

As described above, the small pixel block with reference to frame in used in interaction prediction（For example, 4 × 4 block of pixels）There is fortune Dynamic information 18, it is possible to according to the property of the local of received image signal 10 come shape using optimal motion compensation block with And motion vector.In addition, the microlith of Fig. 7 A to Fig. 7 D and sub- microlith can arbitrarily combine.When coded object block is Fig. 7 A institute In the case of showing such 64 × 64 block of pixels, each for segmentation 4 32 × 32 block of pixels obtaining of 64 × 64 block of pixels Individual come to select each block size shown in Fig. 7 B such that it is able to periodically utilize 64 × 64～16 × 16 pixels block.Similarly, In the case of can choosing the block size shown in Fig. 7 D, can periodically utilize 64 × 64～4 × 4 block size.

Then, reference picture 8A to Fig. 8 F is illustrating motion reference block.

Motion reference block is the side of the mutual agreement according to the picture coding device of Fig. 1 and picture decoding apparatus described later The region that finishes of coding from coded object frame and with reference to frame in for the method（Block）Middle selection.Fig. 8 A illustrates according to coded object block Position and select one of the configuration of motion reference block.In the example of Fig. 8 A, from coded object frame and with reference to frame in 9 motion reference block A～D and TA～TE of regional choice of finishing of coding.Specifically, from coded object frame, select The 4 block A adjoining with left, the upper, upper right of coded object block, upper left, B, C, D, as motion reference block, select from reference frame With the block TA of coded object block same position and the right side with this block TA, under, left and upper 4 adjacent block of pixels TB, TC, TD, TE are as motion reference block.In the present embodiment, the motion reference block selecting from coded object frame is referred to as space side To motion reference block, the motion reference block selecting from reference frame is referred to as time orientation motion reference block.Give each of Fig. 8 A The coding p of motion reference block represents the index of motion reference block.This index is by the motion reference according to time orientation, direction in space The order of block and number, but be not so limited, as long as index does not repeat it is not necessary to must be according to this order.For example, time orientation with And the motion reference block of direction in space can also order pell-mell number.

In addition, direction in space motion reference block is not limited to the example shown in Fig. 8 A it is also possible to as shown in Figure 8 B It is pixel a adjoining with coded object block, the block belonging to b, c, d（For example, microlith or sub- microlith etc.）.In this case, from Top left pixel e in coded object block is to each pixel a, the relative position of b, c, d（Dx, dy）It is set to shown in Fig. 8 C.? This, in the example shown in Fig. 8 A and Fig. 8 B, microlith is shown as N × N block of pixels.

In addition, as in fig. 8d it is also possible to by the whole block A1～A4 adjacent with coded object block, B1, B2, C, D select It is selected as direction in space motion reference block.In the example of Fig. 8 D, the quantity of direction in space motion reference block is 8.

In addition, time orientation motion reference block both can with a part of overlapping of each piece of TA～TE shown in Fig. 8 E it is also possible to Each piece of TA～TE configured separate as shown in Figure 8 F.In Fig. 8 E, time orientation motion reference block TA and TB is illustrated with oblique line Overlapping part.In addition, time orientation motion reference block is not limited to must be position corresponding with coded object block （Collocate position）Block and be located at block about example it is also possible to configuration is in the optional position with reference to frame in Block.For example, it is possible to will be had by referring to the block that position and the arbitrary coding adjacent with coded object block of block finish Movable information 18 determine, be chosen as central block with reference to the block of frame in（For example, block TA）, and by this central block and about Block be chosen as time orientation motion reference block.In addition, time orientation reference block can also configure from central block unequal interval.

Above-mentioned arbitrary in the case of, if direction in space of having reached an agreement in advance in code device and decoding apparatus is with timely Between the quantity of direction motion reference block and position, then the quantity of motion reference block and position can arbitrarily be arranged.In addition, The size of motion reference block not have to be with coded object block formed objects.For example, as in fig. 8d, motion reference block is big Little both can bigger than coded object block it is also possible to less than coded object block.In addition, motion reference block is not limited to square shape, Can be set as the arbitrary shapes such as rectangular shape.In addition, motion reference block can also be set as any size.

In addition, motion reference block and available block can also only configure a certain in time orientation and direction in space Side.Alternatively, it is also possible to cut into slices according to P, B cut into slices these section species come setup time direction motion reference block and can profit With block, and can also the motion reference block in configuration space direction and available block.

The method that Fig. 9 illustrates to can be utilized block obtaining section 109 to select to utilize block from motion reference block.Available block is energy Enough blocks to coded object block application movable information, and there is mutually different movable information.The reference of available block obtaining section 109 Reference movement information 19, according to the method shown in Fig. 9, judges whether motion reference block is each that block can be utilized, and export can profit With block message 30.

As shown in figure 9, first, select the motion reference block that index p is 0（S800）.In the explanation of Fig. 9, imagination according to Index p is from 0 to M-1（M represents the quantity of motion reference block.）Sequential processing motion reference block situation.In addition, setting for rope The available determination processing drawing the motion reference block from 0 to p-1 for the p terminates and becomes the fortune determining whether available object The index of dynamic reference block illustrates for p.

Whether available block obtaining section 109 judges whether motion reference block p has movable information 18, be assigned at least One motion vector（S801）.In the case that motion reference block p does not have motion vector, i.e. time orientation motion reference block p Be do not have movable information I section in block or the whole small pixel block in time orientation motion reference block p by interior pre- In the case that survey encodes, enter step S805.In step S805, motion reference block p is judged as not can be utilized block.

In the case of being judged to that in step S801 motion reference block p has movable information, enter step S802.Can be utilized Block obtaining section 109 selectes the motion reference block q having been chosen to that block can be utilized（Available block q）.Here, q is less than p Value.Then, available block obtaining section 109 is carried out to the movable information 18 of motion reference block p and the movable information 18 of available block q Relatively, determine whether that there is same movable information（S803）.When the movable information 18 being judged to motion reference block p be selected as In the case of movable information 18 identical of the motion reference block q of available block, enter step S805, be judged to motion reference block p For block not can be utilized.

Meet q when being directed to<Whole available block q of p, is judged to the motion letter of motion reference block p in step S803 In the case that the movable information 18 of breath 18 and available block q differs, enter step S804.In step S804, available block Motion reference block p is judged to block can be utilized by obtaining section 109.

If motion reference block p is judged as block can be utilized or block not can be utilized, available block obtaining section 109 judgement is No performing for whole motion reference blocks can be utilized judgement（S806）.It is not performed the motion ginseng that judgement can be utilized existing In the case of block, for example, in p<In the case of M-1, enter step S807.Then, available block obtaining section 109 will index p Plus 1（Step S807）, and again execution step S801 to step S806.When being judged in step S806 for whole motions Reference block performs when judgement can be utilized, and terminates to utilize determination processing.

By executing above-mentioned available determination processing, judge that each motion reference block is block to can be utilized or not can be utilized Block.Available block obtaining section 109 generates the available block message 30 including the information relevant with available block.Thus, by from Select to utilize block in motion reference block, the quantity of information relevant with available block message 30 reduces, and as a result, can reduce volume The amount of code data 14.

Figure 10 is shown for that the motion reference block shown in Fig. 8 A performs the result of available determination processing.In figure In 10, it is judged to 2 direction in space motion reference blocks（P=0,1）And 2 time orientation motion reference blocks（P=5,8）For can profit With block.Figure 11 illustrates of the available block message 30 relevant with the example of Figure 10.As shown in figure 11, available block message 30 Including the index of motion reference block, utilizability and motion reference block title.In the example of Figure 11, index p=0,1,5,8 For block can be utilized, available block number is 4.Prediction section 101 selects an optimal available block conduct from these available blocks Selection block, and export the information relevant with selection block（Select block message）31.Block message 31 is selected to include utilizing the quantity of block And the index value of selected available block.For example, in the case that the quantity of available block is 4, using maximum entry Code table for 4, selects block message 31 to encode by variable length code portion 104 to corresponding.

In addition, in step S801 of Fig. 9, when at least one of block in time orientation motion reference block p is interior In the case of the block of predictive coding, motion reference block p can also be judged to not can be utilized block by available block obtaining section 109. I.e. it is also possible to be set to only in the case that the whole block in time orientation motion reference block p is encoded with interaction prediction, Enter step S802.

Figure 12 A to Figure 12 E is shown in the comparison of movable information 18 of step S803, is judged to the fortune of motion reference block p The movable information 18 identical example of dynamic information 18 and available block q.Have many with oblique line respectively shown in Figure 12 A to Figure 12 E Individual block and 2 blocks whitewashed.In Figure 12 A to Figure 12 E, for the purpose of simplifying the description it is assumed that not considering to compare this 2 with the block of oblique line The movable information 18 of the block whitewashed.One of 2 blocks whitewashed is motion reference block p, and another is to be judged to can be utilized Motion reference block q（Available block q）.Especially, unless otherwise specified, any one of 2 white blocks can be Motion reference block p.

Figure 12 A illustrates that the both sides of motion reference block p and available block q are the example of the block of direction in space.Figure 12 A's In example, the movable information 18 of such as fruit block A and B is identical, then be judged to that movable information 18 is identical.Now it is not necessary to block A and The size of B is identical.

Figure 12 B illustrates that of motion reference block p and available block q is the block A of direction in space, and another is the time The example of the block TB in direction.In Figure 12 B, there is a block with movable information in the block TB of time orientation.If the time The movable information 18 of the block TB in direction is identical with the movable information 18 of the block A of direction in space, then be judged to that movable information 18 is identical. Now it is not necessary to the size of block A and TB is identical.

Figure 12 C illustrates that of motion reference block p and available block q is the block A of direction in space, and another is time side To block TB another example.Figure 12 C illustrates for the block TB of time orientation to be divided into multiple fritters, and exists and multiple have motion The fritter situation of information 18.In the example of Figure 12 C, whole block with movable information 18 has identical movable information 18, if this movable information 18 is identical with the movable information 18 of the block A of direction in space, it is judged to that movable information 18 is identical.This When it is not necessary to the size of block A and TB is identical.

Figure 12 D illustrates that motion reference block p and available block q is the example of the block of time orientation.In this case, As the movable information 18 of fruit block TB and TE, then it is judged to that movable information 18 is identical.

Figure 12 E illustrates that motion reference block p and available block q is another example of the block of time orientation.Figure 12 E illustrates Block TB and TE of time orientation is divided into multiple fritters respectively, and the multiple fritters with movable information 18 are individually present Situation.In this case, for each the fritter comparing motion information 18 in block, if for whole fritter movable informations 18 is identical, then be judged to that the movable information 18 of block TB is identical with the movable information 18 of block TE.

Figure 12 F illustrates that motion reference block p and available block q is another example of the block of time orientation.Figure 12 F illustrates The block TE of time orientation is divided in multiple fritters and block TE has multiple fritters with movable information 18.? Whole movable information 18 of block TE be identical movable information 18 and with movable information 18 identical of block TD in the case of, It is judged to that block TD is identical with the movable information 18 of TE.

Thus, in step S803, the movable information 18 of motion reference block p and the movable information 18 of available block q are judged Whether identical.In the example of Figure 12 A to Figure 12 F, the quantity of available block q being set to be compared with motion reference block p is 1 It is illustrated, but it is also possible to movable information 18 by motion reference block p in the case that the quantity of available block q is more than 2 It is compared with the movable information 18 of each available block q.In addition, in the case of the proportional zoom stated after application, ratio contracts Movable information 18 after putting becomes the movable information 18 of described above.

In addition, the movable information identical of the movable information of motion reference block p and available block q judges to be not limited to motion letter Cease the on all four situation of each motion vector being comprised.For example, as long as the norm of the difference of 2 motion vectors（norm）In regulation In the range of, the movable information that can be seen as motion reference block p is substantially the same with the movable information of available block q.

Figure 13 illustrates the more detailed structure of prediction section 101.As described above, the input of this prediction section 101 can utilize block message 30th, reference movement information 19 and reference picture signal 17, output prediction image signal 11, movable information 18 and selection block letter Breath 31.As shown in figure 13, movable information selector 118 possesses direction in space movable information obtaining section 110, time orientation motion letter Breath obtaining section 111 and movable information switching switch 112.

In direction in space movable information obtaining section 110 input can using block message 30 and with direction in space motion reference The relevant reference movement information 19 of block.Direction in space movable information obtaining section 110 output include positioned at direction in space respectively can profit The movable information being had with block and the movable information 18A of the index value of available block.When as can be using block message 30 In the case of information shown in input Figure 11, direction in space movable information obtaining section 110 generates 2 movable informations and exports 18A, respectively Movable information exports the movable information 19 that 18A includes to have using block and this available block.

In time orientation movable information obtaining section 111 input can using block message 30 and with time orientation motion reference block Relevant reference movement information 19.Time orientation movable information obtaining section 111, output utilization can using block message 30 determine can By the use of the movable information 19 that has of time orientation motion reference block and available block index value and as movable information 18B.Time orientation motion reference block is divided into multiple small pixel blocks, and each small pixel block has movable information 19.As Figure 14 institute Show, the movable information 18B that time orientation movable information obtaining section 111 is exported includes being had using each small pixel block in block The group of some movable informations 19.In the case that movable information 18B includes the group of movable information 19, can be with partition encoding object The small pixel block unit that block obtains executes motion compensated prediction to coded object block.When as figure can be inputted using block message 30 In the case of information shown in 11, time orientation movable information obtaining section 111 generates 2 movable informations and exports 18B, each motion letter The group of the movable information 19 that breath output inclusion can be had using block and this available block.

In addition, time orientation movable information obtaining section 111 can also obtain the movable information 19 that each pixel fritter has The meansigma methodss of the motion vector being comprised or representative value, and using the meansigma methodss of motion vector or representative value as movable information 18B And export.

The movable information switching switch 112 of Figure 13 is according to from direction in space movable information obtaining section 110 and time orientation Movable information 18A and 18B of movable information obtaining section 111 output, selects one suitably block alternatively block to can be utilized, and Will movable information 18 corresponding with selection block（Or the group of movable information 18）Export dynamic compensating unit 113.In addition, movable information The switching switch 112 output selection block message 31 relevant with selection block.Block message 31 is selected to include indexing p or motion reference block Title etc., is also often simply referred to as selection information.Block message 31 is selected to be not limited to index the title of p and motion reference block, As long as can determine the position of selection block or arbitrary information.

Movable information switches switch 112, such as be minimum by the coding cost derived by the cost formula shown in following formula 1 Available block be chosen as selection block.

[number 1]

J=D+ λ × R (1)

Here, J presentation code cost, D illustrate that received image signal 10 and with reference between picture signal 17 square The coding distortion of error sum.In addition, R represents the encoding amount estimated by virtual encoder, λ represents by quantization width isotactic Fixed Lagrange（Lagrange）Inderminate coefficient.Both with substituted 1, and only can be calculated using encoding amount R or coding distortion D Coding cost J, it is possible to use the value that encoding amount R or coding distortion D is approximately obtained to make the cost function of formula 1. In addition, coding distortion D be not limited to square error and or forecast error absolute value and（SAD:sums of absolute difference：Difference absolute value and）.Encoding amount R is it is also possible to only use the coding relevant with movable information 18 Amount.In addition, being not limited to the available block being minimum by coding cost to be chosen as the example of selection block it is also possible to cost tool will be encoded An available block having the value in the range of certain of more than the value of minimum is chosen as selection block.

The movable information that dynamic compensating unit 113 has according to the selected selection block of movable information selector 118（Or fortune The group of dynamic information）, derive the position taking out the block of pixels with reference to picture signal 17 as prediction image signal 11.When to motion In the case that compensation section 113 have input the group of movable information, dynamic compensating unit 113 will take out ginseng as prediction image signal 11 Block of pixels according to picture signal 17 is divided into small pixel block（For example, 4 × 4 block of pixels）, and, each to these small pixel blocks The corresponding movable information 18 of individual application, thus obtain prediction image signal 11 according to reference to picture signal 17.Obtain prognostic chart picture The position of the block of signal 11, for example as shown in Figure 6A, becomes the motion vector 18a being comprised corresponding to movable information 18 and from little The position in direction in space displacement for the block of pixels.

Motion compensation process for coded object block can be using the process same with motion compensation process H.264. Here, as one, illustrating the interpolating method of 1/4 pixel precision.In the interpolation of 1/4 pixel precision, work as motion vector Each component be 4 multiple when, motion vector indicate integer pixel positions.In the case of in addition, motion vector instruction with The corresponding predicted position of interpolation position of fraction precision.

[number 2]

x_pos=x+(mv_x/4) (2)

Y_pos=y+ (mv_y/4)

Here, x and y illustrates that the beginning location of prediction object block（For example, left upper apex）Vertical and level The index in direction, x_pos and y_pos represents the corresponding predicted position with reference to picture signal 17.（Mv_x, mv_y）Represent tool There is the motion vector of 1/4 pixel precision.Then, the location of pixels obtaining for segmentation, by referring to the correspondence of picture signal 17 Location of pixels fill up or interpolation processing generate prediction pixel.Figure 15 illustrates of prediction pixel generation H.264.In Figure 15 In with the square shown in the Latin alphabet of daimonji（Square with oblique line）Represent the pixel of integer position, with grid lines table The square showing illustrates the inter polated pixel of 1/2 location of pixels.In addition, being illustrated and 1/4 location of pixels with whitewashing the square representing Corresponding inter polated pixel.For example, in fig .15,1/2 pixel corresponding with the position of the Latin alphabet b, h is calculated by following formula 3 Interpolation processing.

[number 3]

B=(E-5 × F+20 × G+20 × H-5 × I+J+16) ＞ ＞ 5 (3)

H=(A-5 × C+20 × G+20 × M-5 × R+T+16) ＞ ＞ 5

Here, the Latin alphabet shown in formula 3 and following formula 4（For example, b, h, C1 etc.）Represent in figure 16 and impart phase The pixel value of the same pixel of the Latin alphabet.In addition, ">>" represent right shift operation, ">>5 " be equivalent to divided by 32.That is, 1/2 picture The inter polated pixel of plain position uses 6 tap FIR（Finite Impulse Response：Finite impulse response）Wave filter（Tap Coefficient:（1, -5,20,20, -5,1）/32）To calculate.

In addition, the interpolation processing of 1/4 pixel corresponding with the position of the Latin alphabet a, d is counted by following formula 4 in Figure 15 Calculate.

[number 4]

A=(G+b+1) ＞ ＞ 1 (4)

D=(G+h+1) ＞ ＞ 1

Thus, the inter polated pixel of 1/4 location of pixels uses the averaging filter of 2 taps（Tap coefficient:（1/2,1/2）） Calculate.The interpolation processing of 1/2 pixel corresponding with the Latin alphabet j of the centre being present in 4 integer pixel positions is using vertical Two directions of direction 6 tap and horizontal direction 6 tap generate.Same side is also utilized to the location of pixels beyond illustrated Method generates inter polated pixel value.

In addition, interpolation processing is not limited to formula 3 and the example of formula 4, it is possible to use other interpolated coefficients generate.Separately Outward, interpolated coefficients both can use the value of the fixation providing from coding-control portion 150, or it is also possible to according to above-mentioned coding Cost and be directed to each frame optimization interpolated coefficients, and using optimize after interpolated coefficients generate.

In addition, in the present embodiment, to being microlith with motion reference block（For example, 16 × 16 block of pixels）The motion of unit Vector block prediction processes relevant process and is described, but be not limited to microlith it is also possible to 16 × 8 pixel block units, 8 × 16 Pixel block unit, 8 × 8 pixel block units, 8 × 4 pixel block units, 4 × 8 pixel block units or 4 × 4 pixel block units are executing Prediction is processed.In this case, the information relevant with motion vector block is derived with pixel block unit.Alternatively, it is also possible to 32 The unit that × 32 pixel block units, 32 × 16 pixel block units, 64 × 64 pixel block units etc. are more than 16 × 16 block of pixels is carried out Above-mentioned prediction is processed.

Using the reference movement vector in motion vector block as the motion vector of the small pixel block in coded object block During substitution, both can substitute into（A）The negative value of reference movement vector（Reversion vector）, or,（B）Can also substitute into employ with little The weighted mean of reference movement vector or central authorities that block corresponding reference movement vector is adjoined with this reference movement vector Value, maximum, minima.

Figure 16 roughly illustrates the action of prediction section 101.As shown in figure 16, first, obtain and include time orientation with reference to fortune The reference frame of motion block（Motion reference frame）（Step S1501）.With regard to motion reference frame, typically with coded object frame time away from From minimum reference frame, it is reference frame in the past in time.For example, motion reference frame is the tight above quilt in coded object frame The frame encoding.In other examples, can also obtain in movable information memorizer 108 as motion reference frame and preserve motion Some reference frame of information 18.Then, direction in space movable information obtaining section 110 and time orientation movable information obtaining section 111 obtain the available block message 30 from the output of available block obtaining section 109 respectively（Step S1502）.Then, movable information is cut Change switch 112, for example, select an alternatively block from available block according to formula 1（Step S1503）.Then, motion compensation The copying motion information that selected selection block is had by portion 113 is to coded object block（Step S1504）.Now, work as selection In the case that block is direction in space reference block, as shown in figure 17, the movable information 18 that this selection block has is copied into coding Reference block.In addition, in the case that selection block is time orientation reference block, the group of movable information 18 that this selection block has with Positional information is copied into coded object block together.Then, using the movable information 18 being copied by dynamic compensating unit 113 or motion The group execution motion compensation of information 18, and export prediction image signal 11 and movable information used in motion compensated prediction 18.

Figure 18 illustrates the more detailed structure of variable length code portion 104.As shown in figure 18, variable length code portion 104 Possess parameter coding portion 114, transform coefficients encoding portion 115, selection block encoding section 116 and multiplexing portion 117.Parameter coding Portion 114 except to conversion coefficient information 13 and select in addition to block message 31 encodes, also to prediction mode information, block size The parameter that the decoding of information, quantum-chemical descriptors information etc. is required is encoded, and generates coded data 14A.Transform coefficients encoding Portion 115 encodes to conversion coefficient information 13, generates coded data 14B.In addition, selection block encoding section 116 is with reference to available Block message 30, to selecting block message 31 to encode, generates coded data 14C.

As shown in figure 19, when available block message 30 includes indexing and available with index corresponding motion reference block In the case of property, exclude not available motion reference block from multiple motion reference blocks set in advance, only will be available Motion reference block is transformed to grammer（stds_idx）.In Figure 19,5 motion reference blocks in 9 motion reference blocks can not profit With so for the 4 motion reference blocks eliminating after this 5 motion reference blocks, starting to be sequentially allocated grammer stds_ from 0 idx.In this example embodiment, the to be encoded block message that selects is not to select from 9, but select from 4 available blocks, So the encoding amount of distribution（Bin number）Fifty-fifty decrease.

Figure 20 is 2 value informations representing grammer stds_idx and grammer stds_idx（bin）One of code table.As Shown in Figure 18, the quantity of available motion reference block is fewer, and the average bin number that the coding of grammer stds_idx is required more subtracts Few.For example, in the case that the quantity of available block is 4, grammer stds_idx can be to represent less than or equal to 3 bits.Grammer 2 value informations of stds_idx（bin）Both identical bin number can be become with the stds_idx whole for each available block number Mode 2 value it is also possible to carry out 2 values according to the 2 value methods determining by prior learning.Alternatively, it is also possible to prepare Multiple 2 value methods, and switch over for each coded object block suitability.

Entropy code can be applied in these encoding section 114,115,116（Such as fixed-length coding, Huffman encoding or arithmetic Coding etc.）, coded data 14A, 14B being generated, 14C are multiplexed and exported by the portions that multiplexes 117.

In the present embodiment, imagination is using the frame being encoded than 1 frame before coded object frame reference as reference frame Example is illustrated but it is also possible to the motion vector in the reference movement information 19 that had using selection block and reference frame Number, carries out proportional zoom to motion vector（scaling）（Or standardization）, and to coded object block application reference movement information 19.

Process with regard to this proportional zoom, specifically illustrate with reference to Figure 21.Tc presentation code object frame shown in Figure 21 Time gap and motion reference frame between（POC（Represent the number of DISPLAY ORDER）Distance）, calculated by following formula 5.Figure Tr [i] shown in 21 represents the time gap between motion reference frame and the frame i of selection block institute reference, is counted by following formula 6 Calculate.

[number 5]

Tc=Clip (- 128,127, DiffPicOrderCnt (curPOC, colPOC)) (5)

Tr [i]=Clip (- 128,127, DiffPicOrderCnt (colPOC, refPOC)) (6)

Here, the POC of curPOC presentation code object frame（Picture Order Count：Sequential counting）, colPOC table Show the POC that the POC of motion reference frame, refPOC represent the frame i of selection block institute reference.In addition, Clip（Min, max, target） It is following CLIP function：Target is output min during the value less than min, and target is greater than exporting max during the value of max, except this Target is exported in the case of in addition.In addition, DiffPicOrderCnt（X, y）It is the function of the difference calculating 2 POC.

If setting the motion vector of selection block as MVr=（MVr_x, MVr_y）, to coded object block application motion vector For MV=（MV_x, MV_y）, then motion vector MV is calculated by following formula 7.

[number 6]

MV_x=(MVr_x × tc+Abs (tr [i]/2))/tr [i] (7)

MV_y=(MVr_y × tc+Abs (tr [i]/2))/tr [i]

Here, Abs（x）Represent the function of the absolute value taking out x.Thus, in the proportional zoom of motion vector, distribute to The motion vector MVr of selection block is transformed to the motion vector MV between coded object frame and motion the 1st reference frame.

In addition, following illustrate another example relevant with the proportional zoom of motion vector.

First, for each section or each frame, according to following formula 8, with regard to the whole of motion reference frame can be obtained Time gap tr obtains scaling coefficient（DistScaleFactor[i]）.The quantity of scaling coefficient and selection block are joined According to frame quantity, i.e., the quantity of reference frame equal.

[number 7]

Tx=(16384+Abs (tr [i]/2))/tr [i] (8)

DistScaleFactor [i]=Clip (- 1024,1023, (tc × tx+32)) ＞ ＞ 6

Calculating with regard to the tx shown in formula 8 is it is also possible to tabular in advance.

In the proportional zoom for each coded object block, by using following formula 9, only pass through multiplication, addition, displacement Computing just can calculate motion vector MV.

[number 8]

MV_x=(DistScaleFactor [i] × MVr_x+128) ＞ ＞ 8 (9)

MV_y=(DistScaleFactor [i] × Mvr_y+128) ＞ ＞ 8

In the case of implementing such proportional zoom process, the process of prediction section 101 and available block obtaining section 109 Movable information 18 all after application percentage scaling.In the case of implementing proportional zoom process, the reference of coded object block institute Reference frame becomes motion reference frame.

Figure 22 illustrates the grammatical structure in picture coding portion 100.As shown in figure 22, grammer mainly includes 3 parts, that is, High-level syntax 901, slice-level grammer 904 and microlith level grammer 907.High-level syntax 901 maintains the above upper layer of section Syntactic information.Slice-level grammer 904 is directed to the information required for each section keeps, and microlith level grammer 907 is directed to Fig. 7 A extremely Each microlith shown in Fig. 7 D maintains required data.

Each several part includes more detailed grammer.High-level syntax 901 includes sequential parameter group grammer 902 and graphic parameter group The sequences such as grammer 903 and the grammer of figure level.Slice-level grammer 904 includes cut into slices head grammer 905 and slice of data grammer 906 etc..In addition, microlith level grammer 907 includes microlith layer grammer 908 and microlith prediction grammer 909 etc..

Figure 23 A and Figure 23 B illustrates the example of microlith layer grammer.Available_ shown in Figure 23 A and Figure 23 B Block_num represents the quantity that can utilize block, in the case of being more than 1 value when this value, needs to select the coding of block message.Separately Outward, stds_idx illustrate select block message, using aforesaid and available block number the corresponding code table of quantity to stds_idx Encoded.

Grammer when Figure 23 A encodes to selection block message after being shown in mb_type.Represented by mb_type Pattern is the size determining or the pattern determining（TARGET_MODE）In the case of, and in available_block_ Num be more than 1 value in the case of, stds_idx is encoded.For example, when the movable information of selection block becomes available Block size is 64 × 64 pixels, 32 × 32 pixels, in the case of 16 × 16 pixels, or in the case of Direct Model, to stds_ Idx is encoded.

Grammer when Figure 23 B encodes to selection block message before being shown in mb_type.Work as available_block_ Num be more than 1 value in the case of, stds_idx is encoded.In addition, if available_block_num is 0, then Carry out H.264 representative conventional motion compensation, so encoding to mb_type.

In addition, the unspecified grammer in the present invention that in the ranks can also insert of the table shown in Figure 23 A and Figure 23 B is wanted Element is it is also possible to include relevant with condition difference in addition description.Or it is also possible to syntax table is split, merges into Multiple tables.Additionally, it is not necessary to must be using same term it is also possible to arbitrarily be changed according to the mode of utilization.And then, in this microlith layer Each grammer segment describing in grammer can also change, so that described later microlith data syntax is expressly recited.

In addition, the information of mb_type can be cut by using the information of stds_idx.Figure 24 A be with H.264 in B The corresponding code table of mb_type and mb_type during section.N shown in Figure 24 A is to represent 16,32,64 grade coded object blocks Size value, M is the value of the half of N.Therefore, in the case that mb_type is 4～21, coded object block is shown as rectangular Shape block.In addition, the L0 of Figure 24 A, L1, Bi represent that one direction is predicted respectively（Only List0 direction）, one direction prediction（Only List1 side To）, twocouese prediction.In the case that coded object block is rectangular blocks, mb_type is rectangular for 2 in coded object block Shape block each, carried out L0, the information of the prediction of some in L1, Bi including expression.In addition, B_Sub represents for 4 points Cut each above-mentioned process of execution of the block of pixels of microlith.For example, in the case that coded object block is 64 × 64 pixel microliths, Coded object block, each for 4 segmentation 4 32 × 32 block of pixels obtaining of this microlith, distribution mb_type goes forward side by side further Row coding.

Here, when the selection block that stds_idx represents is Spatial Left（The picture adjoining with the left side of coded object block Plain block）In the case of, the movable information of the block of pixels that the left side with coded object block is adjoined is set to the motion of coded object block Information, so stds_idx has and the mb_type=4 using Figure 24 A, growing crosswise shown in 6,8,10,12,14,16,18,20 Rectangular blocks are to coded object block perform prediction identical implication.In addition, the selection block shown in stds_idx is SpatialUp's In the case of, the movable information that the upside with coded object block is adjoined is set to the movable information of coded object block, so stds_ Idx has and the mb_type=5 using Figure 24 A, and the rectangular blocks execution of the lengthwise shown in 7,9,11,13,15,17,19,21 is pre- Survey identical implication.Therefore, by using stds_idx, the mb_ reducing Figure 24 A as shown in Figure 24 B can be made The code table on the hurdle of type=4～21.Similarly, with regard to shown in Figure 24 C H.264 in P section when mb_type and The corresponding code table of mb_type is it is also possible to make the code table of the quantity reducing mb_type as shown in Figure 24 D.

Encoded alternatively, it is also possible to be included in the information of stds_idx in the information of mb_type.Figure 25 A illustrate by The information of stds_idx is included in the code table during information of mb_type, illustrates mb_type and mb_type pair with B section One of the code table answered.The B_STDS_X of Figure 25 A（X=0,1,2）The pattern suitable with stds_idx is shown, adds and can be utilized The B_STDS_X of the amount of block number（In Figure 25 A, available block number is 3）.Similarly, Figure 25 B illustrates the mb_ relevant with P section Another example of type.The explanation of Figure 25 B is identical with B section, so omitting.

The order of mb_type and 2 value methods（Binization）It is not limited to the example shown in Figure 25 A and Figure 25 B, also may be used With in other sequences and 2 value methods encode to mb_type.B_STDS_X and P_STDS_X does not need continuously, Can also configure between each mb_type.In addition, 2 value methods（Binization）Can also based on the selection frequency learning in advance Design.

In the present embodiment, though collect multiple microliths and carry out motion compensated prediction extension microlith in also can answer Use the present invention.In addition, in the present embodiment, the scanning sequency with regard to coding can be arbitrary order.For example, to row scanning Or Z scanning etc. also can apply the present invention.

As described above, the picture coding device of present embodiment selects to utilize block, root from multiple motion reference blocks Generate the information for determining the motion reference block to the application of coded object block according to the quantity of selected available block, and to this Information is encoded.Therefore, the picture coding device according to present embodiment, even if cut down the volume relevant with motion vector information Code amount, also can carry out motion compensation with the small pixel block unit thinner than coded object block, it is possible to realizing high coding effect Rate.

（2nd embodiment）

Figure 26 illustrates the picture coding device involved by the 2nd embodiment of the present invention.In the 2nd embodiment, mainly The part different from the 1st embodiment and action are illustrated.As shown in figure 26, in the picture coding portion of present embodiment In 200, the structure of prediction section 201 and variable length code portion 204 is different from the 1st embodiment.As shown in figure 27, prediction section 201 possess the 1st prediction section 101 and the 2nd prediction section 202, optionally switch this 1st and the 2nd prediction section 101,202 and give birth to Become prediction image signal 11.1st prediction section 101 has the prediction section 101 with the 1st embodiment（Fig. 1）Identical structure, according to The movable information 18 being had using selection block carries out the prediction mode of motion compensation（1st prediction mode）Generate prognostic chart picture letter Numbers 11.2nd prediction section 202 according to coded object block using a motion vector carry out motion compensation, as H.264 Prediction mode（2nd prediction mode）Generate prediction image signal 11.2nd prediction section 202 using from received image signal 10 with And the reference picture signal 17 of frame memory generates prediction image signal 11B.

Figure 28 roughly illustrates the structure of the 2nd prediction section 202.As shown in figure 28, the 2nd prediction section 202 has：Movable information Obtaining section 205, generates movable information 21 using received image signal 10 and with reference to picture signal 17；And dynamic compensating unit 113（Fig. 1）, generate prediction image signal 11A using with reference to picture signal 17 and movable information 21.This movable information obtaining section 205 according to received image signal 10 and with reference to picture signal 17, for example, are obtained by Block- matching and distribute to coded object The motion vector of block.As coupling metewand, using for each accumulation received image signal 10 with mate after It is worth obtained from the difference of interpolating image.

In addition, movable information obtaining section 205 can also use the difference to prediction image signal 11 and received image signal 10 The value dividing to enter line translation to obtain, to determine optimal motion vector.Alternatively, it is also possible to consider size, the Yi Jiyun of motion vector Dynamic vector and the encoding amount with reference to frame number, or, determine optimal motion vector using formula 1.Matching process both can root To execute it is also possible to be directed to each pixel precision periodically according to from the outside hunting zone information providing of picture coding device Execution.Alternatively, it is also possible to not scan for processing, and using the movable information being provided by coding-control portion 150 as movable information The output 21 of obtaining section 205.

The prediction section 101 of Figure 27 is further equipped with Forecasting Methodology switching switch 203, selects and exports from the 1st prediction section 101 prediction image signal 11A and the one party of the prediction image signal 11B from the 2nd prediction section 202.For example, prediction side Method switching switch 203 is directed to each prediction image signal 11A and 11B, using received image signal 10, for example, asks according to formula 1 Go out to encode cost, select the one party in prediction image signal 11A and 11B in the way of encoding cost and be less, and as pre- Altimetric image signal 11 and export.In addition, Forecasting Methodology switching switch 203 also with movable information 18 and select block message 31 1 Rise, output represents exported prediction image signal 11 is which output from the 1st prediction section 101 and the 2nd prediction section 202 Prediction image signal prediction handover information 32.The movable information 18 being exported is encoded by variable length code portion 204, afterwards It is multiplexed as coded data 14.

Figure 29 roughly illustrates the structure of variable length code portion 204.Variable length code portion 204 shown in Figure 29 except Outside the structure of variable length code portion 104 shown in Figure 18, it is also equipped with movable information encoding section 217.In addition, the selection of Figure 29 Block encoding section 216 is different from the selection block encoding section 116 of Figure 18, and prediction handover information 32 is encoded and generated coded data 14D.In the case that the 1st prediction section 101 performs prediction process, selection block encoding section 216 is further to available block message 30 and select block message 31 encoded.Available block message 30 after coded and selection block message 31 are contained in coding Data 14D.In the case that the 2nd prediction section 202 performs prediction process, movable information encoding section 217 is entered to movable information 18 Row coding, and generate coded data 14E.Selection block encoding section 216 and movable information encoding section 217 are respectively according to prediction switching Information 32 performs prediction come which side to judge in the 1st prediction section 101 and the 2nd prediction section 202 and processes, this prediction switching letter Breath 32 represents whether prognostic chart picture is to be generated by using the motion compensated prediction of the movable information of selection block.

Multiplexing portion 117 is from parameter coding portion 114, transform coefficients encoding portion 115, selection block encoding section 216 and fortune Dynamic information encoding section receives coded data 14A, 14B, 14D, 14E, and to coded data 14A being received, 14B, 14D, 14E are carried out Multiplexing.

Figure 30 A and Figure 30 B each illustrates the example of the microlith layer grammer of present embodiment.Shown in Figure 30 A Available_block_num represents the quantity that can utilize block, in the case of being more than 1 value when it, selection block encoding section 216 Encode to selecting block message 31.In addition, whether stds_flag is to represent in motion compensated prediction by the motion of selection block The labelling that information employs as the movable information of coded object block, i.e. be to represent that Forecasting Methodology switching switch 203 have selected Which labelling in 1st prediction section 101 and the 2nd prediction section 202.When the quantity of available block is more than 1 and stds_ In the case that flag is 1, represents and employ the movable information that selection block has in motion compensated prediction.In addition, working as stds_ In the case that flag is 0, the movable information that do not had using selection block, and directly to movable information 18 in the same manner as H.264 Information encoded, or the difference value predicted is encoded.In addition, stds_idx represents selection block message, and can As described above using the corresponding code table of block number.

Grammer when Figure 30 A encodes to selection block message after being shown in mb_type.Only determining mb_type In the case of the size of pattern representing or the pattern determining, stds_flag and stds_idx is encoded.For example, It is in the case that available block size is 64 × 64,32 × 32,16 × 16 or Direct Model in the movable information of selection block In the case of, stds_flag and stds_idx is encoded.

Grammer when Figure 30 B encodes to selection block message before being shown in mb_type.For example when stds_flag is 1 In the case of it is not necessary to encode to mb_type.In the case that stds_flag is 0, mb_type is encoded.

As previously discussed, the picture coding device of the 2nd embodiment optionally switches the 1st prediction of the 1st embodiment Portion 101 and using H.264 wait prediction mode the 2nd prediction section 202, and received image signal is compressed coding so that Coding cost reduces.Therefore, enter than the picture coding device of the 1st embodiment in the picture coding device of the 2nd embodiment One step improves code efficiency.

（3rd embodiment）

Figure 31 roughly illustrates the picture decoding apparatus of the 3rd embodiment.As shown in figure 31, this picture decoding apparatus possesses Picture decoding portion 300, decoding control section 350 and output state 308.Picture decoding portion 300 is controlled by decoding control section 350 System.The picture decoding apparatus of the 3rd embodiment are corresponding with the picture coding device of the 1st embodiment.Namely be based on the image of Figure 31 The coded treatment that the decoding process of decoding apparatus is processed with the picture coding based on Fig. 1 has the relation of complementation.The image of Figure 31 Decoding apparatus both can be realized by hardware such as LSI chips, or can also execute image decoding program by making computer To realize.

The picture decoding apparatus of Figure 31 possess coding row lsb decoder 301, inverse guantization (IQ) inverse transformation portion 302, adder 303rd, frame memory 304, prediction section 305, movable information memorizer 306 and available block obtaining section 307.In picture decoding In portion 300, it is input to coding row lsb decoder 301 from the coded data 80 of storage system (not shown) system or Transmission system.Should Coded data 80, for example, corresponding with the coded data 14 sent from the picture coding device of Fig. 1 with the state being multiplexed.

In the present embodiment, using the block of pixels as decoder object（For example, microlith）Simply referred to as decoder object block.Separately Outward, the picture frame including decoder object block is referred to as decoder object frame.

In coding row lsb decoder 301, for every 1 frame or every 1 field, the solution based on syntactic analysiss is carried out according to grammer Read.Specifically, coding row lsb decoder 301 carries out length-changeable decoding to the coding row of each grammer successively, and to inclusion conversion Coefficient information 33, select having with decoder object block of the information of forecastings such as block message 61, block size information and prediction mode information Coding parameter closing etc. is decoded.

In the present embodiment, decoding parametric includes conversion coefficient 33, selects block message 61 and information of forecasting, and wraps Include required all of parameter during the decoding such as the information relevant with conversion coefficient information relevant with quantization.With information of forecasting, The relevant information of conversion coefficient and the information relevant with quantization are transfused to decoding control section 350 as control information 71. The decoding control information 70 of the parameter including required for the decoding such as information of forecasting and quantum-chemical descriptors is carried by decoding control section 350 The each several part in supply picture decoding portion 300.

In addition, as illustrated afterwards, coding row lsb decoder 301 is decoded to coded data 80, obtains simultaneously Information of forecasting and selection block message 61.Inclusion motion vector and the movable information 38 with reference to frame number can not also be solved Code.

Inverse guantization (IQ) inverse transformation portion 302 is sent to by the conversion coefficient 33 that coding row lsb decoder 301 is understood.Arranged by coding Various information, i.e. quantum-chemical descriptors and the quantized matrix relevant with quantization that lsb decoder 301 is understood is provided to decoding Control unit 350, is downloaded to inverse guantization (IQ) inverse transformation portion 302 in inverse guantization (IQ).Inverse guantization (IQ) inverse transformation portion 302 according to The information relevant with quantization downloaded, carries out inverse guantization (IQ) to conversion coefficient 33, then implements inversion process（For example, Inverse discrete cosine transform etc.）, obtain predictive error signal 34.The inverse transformation in the inverse guantization (IQ) inverse transformation portion 302 based on Figure 31 Process is the inverse transformation of the conversion process in the conversion quantization portion based on Fig. 1.For example, by picture coding device（Fig. 1） In the case of implementing wavelet transformation, inverse guantization (IQ) inverse transformation portion 302 executes corresponding inverse guantization (IQ) and inverse wavelet transform.

Adder 303 is input to by the predictive error signal 34 that inverse guantization (IQ) inverse transformation portion 302 has restored.Addition Predictive error signal 34 is added by device 303 with the prediction image signal 35 generating in prediction section 305 described later, generates decoding figure As signal 36.The decoded image signal 36 being generated is exported from picture decoding portion 300, and is temporarily stored to output state 308, afterwards, the output timing according to decoding control section 350 management is exported.In addition, this decoded image signal 36 is deposited in frame It is saved as in reservoir 304 with reference to picture signal 37.Read reference according to each frame or each field from frame memory 304 successively Picture signal 37, and input to prediction section 305.

Available block obtaining section 307 receives reference movement information 39 from movable information memorizer 306 described later, and export can Using block message 60.The action of available block obtaining section 307 and the available block obtaining section 109 illustrating in the 1st embodiment （Fig. 1）Identical.

Movable information memorizer 306 receives movable information 38 from prediction section 305, and is temporarily saved as reference movement information 39.The movable information 38 exporting from prediction section 305 is temporarily saved as reference movement information 39 by movable information memorizer 306.Fig. 4 Illustrate one of movable information memorizer 306.Movable information memorizer 306 maintains scramble time different multiple movable informations Frame 26.Decode the group of the movable information 38 finishing or movable information 38, be stored in as reference movement information 39 and decode Time corresponding movable information frame 26.In movable information frame 26, for example, preserve reference movement information with 4 × 4 pixel block units 39.The reference movement information 39 that movable information memorizer 306 is kept, by prediction section 305 in the motion generating decoder object block Read and reference during information 38.

Then, the motion reference block of present embodiment and available block are described.Motion reference block is according to by aforesaid The candidate blocks that picture coding device and picture decoding apparatus method set in advance select from the decoded region finishing. Fig. 8 A illustrates relevant with block can be utilized.In fig. 8 a, it is configured with 4 motion reference blocks and the ginseng of decoder object frame in This total 9 motion reference block of 5 motion reference blocks according to frame in.Motion reference block A, B, the C of the decoder object frame in of Fig. 8 A, D be with respect to decoder object block left, on, the block that adjoins of upper right, upper left.In the present embodiment, will be from inclusion decoder object The motion reference block selecting in the decoder object frame of block is referred to as direction in space motion reference block.In addition, the motion ginseng with reference to frame in According to block TA with reference to frame in and decoder object block same position block of pixels, the pixel that will connect with this motion reference block TA Block TB, TC, TD, TE are chosen as motion reference block.The motion reference block selecting from the block of pixels with reference to frame in is referred to as the time Direction motion reference block.In addition, the frame being located at time orientation motion reference block is referred to as motion reference frame.

Direction in space motion reference block is not limited to the example shown in Fig. 8 A, as shown in Figure 8 B it is also possible to will be with decoder object Pixel a that block adjoins, the block of pixels belonging to b, c, d is chosen as direction in space motion reference block.In this case, pixel a, b, C, d are with respect to the relative position of the top left pixel in decoder object block（Dx, dy）As shown in Figure 8 C.

In addition, as in fig. 8d it is also possible to by whole block of pixels A1 adjoining with decoder object block～A4, B1, B2, C, D is chosen as direction in space motion reference block.In Fig. 8 D, the quantity of direction in space motion reference block is 8.

In addition, as illustrated in fig. 8e, time orientation motion reference block TA～TE both can mutually partly overlapping it is also possible to such as It is separated from each other shown in Fig. 8 F.In addition, time orientation motion reference block need not necessarily be block and the position of Collocate position In block about, as long as motion can be then the block of pixels of optional position with reference to frame in.For example, it is also possible to utilize and decoding The movable information of the decoded block finishing that object block adjoins, the reference block indicated by motion vector that movable information is comprised It is chosen as the center of motion reference block（For example, block TA）.In addition, the reference block of time orientation may not be and equally spaced joins Put.

In the method selecting motion reference block as described above, if picture decoding apparatus and picture decoding apparatus Both sides have the information relevant with the quantity of direction in space and time orientation motion reference block and position, then motion reference block Can select from arbitrary quantity and position.In addition, the size of motion reference block need not necessarily be and decoder object block Identical size.For example as in fig. 8d, the size of motion reference block both can be with the size of ratio decoder object block greatly it is also possible to compare Decoder object block slight greatly, can be arbitrary size.In addition, the shape of motion reference block is not limited to square shape, It can be rectangular shape.

Then, available block is illustrated.Available block is the block of pixels selecting from motion reference block, is can be right The block of pixels of decoder object block application movable information.Available block has mutually different movable information.For such as Fig. 8 A institute Show such decoder object frame and the total 9 motion reference blocks with reference to frame in, sentenced by executing the available block shown in Fig. 9 Determine process to select to utilize block.Figure 10 illustrates to execute the result that the available block determination processing shown in Fig. 9 obtains.In Figure 10 In, the block of pixels with oblique line represents block not can be utilized, the block whitewashed represents and can utilize block.I.e., it is judged to move from direction in space Select 2 in reference block, select 4 conducts of 2 total selections can utilize block from time orientation motion reference block.Prediction section Movable information selector 314 in 305 according to the selection block message 61 receiving from selection block lsb decoder 323, from configuration in the time And in these available blocks of direction in space, select an optimal available block alternatively block.

Then, available block obtaining section 307 is illustrated.Available block obtaining section 307 has and the 1st embodiment Available block obtaining section 109 identical function, obtains reference movement information 39 from movable information memorizer 306, for each fortune Dynamic reference block output represents and not can be utilized can the information of block to can be utilized block message 60 using block or.

Flow chart explanation with reference to Fig. 9 can utilize the action of block obtaining section 307.First, available block obtaining section 307 judges Motion reference block（Index p）Whether there is movable information（Step S801）.That is, step S801 determines whether motion reference block At least one small pixel block in p has movable information.In the case of being judged to that motion reference block p does not have movable information, I.e. time orientation motion reference block is the block in the I section do not have movable information, or, complete in time orientation motion reference block In the case that the small pixel block in portion is decoded by interior prediction, enter step S805.In step S805, this motion reference block p is sentenced It is set to and block not can be utilized.

In the case of being judged to that in step S801 motion reference block p has movable information, available block obtaining section 307 Select the motion reference block q being judged to that block can be utilized（It is referred to as and block q can be utilized）（Step S802）.Here, q is less than p's Value.Then, available block obtaining section 307 compares the movable information of this motion reference block p and available block q for whole q Movable information, judges whether motion reference block p has and available block q identical movable information（S803）.As motion reference block p Have with available block q identical motion vector in the case of, enter step S805, in step S805, take by using block Obtain portion 307 and this motion reference block p is judged to not can be utilized block.Motion reference block p has different from whole available block q In the case of movable information, in step S804, can profit by can be utilized block obtaining section 307 to be judged to this motion reference block p With block.

By the available block determination processing above-mentioned to whole motion reference block execution, sentence for each motion reference block Surely it is block to can be utilized or block not can be utilized, and generate and can utilize block message 60.Figure 11 illustrates can be utilized one of block message 60. As shown in figure 11, available block message 60 includes index p and the utilizability of motion reference block.In fig. 11, available block Information 60 illustrates to be chosen as can be utilized block by the motion reference block that index p is 0,1,5 and 8, and the quantity of available block is 4.

In addition, at least one of block in step S801 of Fig. 9 it is also possible in time orientation motion reference block p In the case of being the block by intraprediction encoding, motion reference block p is judged to not can be utilized block by available block obtaining section 307. Can also be set to, only in the case that the whole block in time orientation motion reference block p is encoded with interaction prediction, enter Enter step S802.

Figure 12 A to Figure 12 E is shown in the comparison of movable information 38 of step S803, by the motion letter of motion reference block p The movable information 38 of breath 38 and available block q is judged to identical example.Figure 12 A to Figure 12 E is each shown with many with oblique line Individual block and 2 blocks whitewashed.In Figure 12 A to Figure 12 E, for the purpose of simplifying the description it is assumed that not considering the block with oblique line, and compare this 2 The situation of the movable information 38 of the individual block whitewashed.One side of 2 blocks whitewashed is motion reference block p, and the opposing party is to have judged For available motion reference block q（Available block q）.Unless otherwise specified, any one in 2 white blocks can It is motion reference block p.

Figure 12 A illustrates that the both sides of motion reference block p and available block q are the example of the block of direction in space.Figure 12 A's In example, the movable information 38 of such as fruit block A and B is identical, then be judged to that movable information 38 is identical.Now, block A's and B is big Little do not need identical.

Figure 12 B illustrates that a side of motion reference block p and available block q is the block A of direction in space, and the opposing party is time side To block TB example.In Figure 12 B, there is a block with movable information in the block TB of time orientation.If time side To block TB movable information 38 identical with the movable information 38 of the block A of direction in space, then be judged to that movable information 38 is identical.This When, the size of block A and TB does not need identical.

Figure 12 C illustrates that a side of motion reference block p and available block q is the block A of direction in space, and the opposing party is time side To block TB another example.Figure 12 C illustrates that the block TB of time orientation is divided into multiple fritters, and exist multiple have motion letter The situation of the fritter of breath 38.In the example of Figure 12 C, whole block with movable information 38 has identical movable information 38, if this movable information 38 is identical with the movable information 38 of the block A of direction in space, it is judged to that movable information 38 is identical.This When, the size of block A and TB does not need identical.

Figure 12 D illustrates that motion reference block p and available block q is the example of the block of time orientation.In this case, Movable information 38 as fruit block TB and TE is identical, then be judged to that movable information 38 is identical.

Figure 12 E illustrates that motion reference block p and available block q is another example of the block of time orientation.Figure 12 E shows Go out and block TB and TE of time orientation is divided into multiple fritters respectively, and be individually present and multiple there is the little of movable information 38 The situation of block.In this case, for each the fritter comparing motion information 38 in block, if for whole fritter motions Information 38 is identical, then be judged to that the movable information 38 of block TB is identical with the movable information 38 of block TE.

Figure 12 F illustrates that motion reference block p and available block q is another example of the block of time orientation.Figure 12 F illustrates The block TE of time orientation is divided into multiple fritters, in block TE, there are multiple fritters with movable information 38.In block Whole movable information 38 of TE is identical movable information 38 and the movable information 38 identical situation having with block TD Under, it is judged to that block TD is identical with the movable information 38 of TE.

Thus, in step S803, the movable information 38 of motion reference block p and the movable information 38 of available block q are judged Whether identical.In the example of Figure 12 A to Figure 12 F, if the quantity of the available block q being compared with motion reference block p is entered for 1 Gone explanation, but in the case that the quantity of available block q is more than 2 it is also possible to by the movable information 38 of motion reference block p with The movable information 38 of each available block q is compared.In addition, in the case of the proportional zoom stated after application, proportional zoom Movable information 38 afterwards becomes the movable information 38 of described above.

In addition, the movable information identical of the movable information of motion reference block p and available block q judges to be not limited to motion letter Cease the on all four situation of each motion vector being comprised.As long as example, it is also possible to the norm of the difference of 2 motion vectors is in regulation In the range of be considered as the movable information of motion reference block p and the movable information of available block q is substantially the same.

Figure 32 is the block diagram illustrating in greater detail coding row lsb decoder 301.As shown in figure 32, coding row lsb decoder 301 has Have：The separated part 320 that coded data 80 is separated into syntactical unit, the conversion coefficient lsb decoder that conversion coefficient is decoded 322nd, to the selection block lsb decoder 323 selecting block message to be decoded and to relevant with prediction block sizes and quantization The parameter lsb decoder 321 that parameter etc. is decoded.

Parameter lsb decoder 321 receives the coded number including the parameter relevant with prediction block sizes and quantization from separated part According to 80A, coded data 80A is decoded and generates with control information 71.Conversion coefficient lsb decoder 322 receives from separated part 320 Coded conversion coefficient 80B, is decoded to the conversion coefficient 80B of this coding, obtains conversion coefficient information 33.Selection block solution Code portion 323 input coded data 80C relevant with selection block and available block message 60, output selects block message 61.As figure Shown in 11, the available block message 60 being inputted illustrates utilizability for each motion reference block.

Then, with reference to Figure 33, describe prediction section 305 in detail.

As shown in figure 33, prediction section 305 has movable information selector 314 and dynamic compensating unit 313, and movable information selects Select portion 314 there is direction in space movable information obtaining section 310, time orientation movable information obtaining section 311 and movable information and cut Change switch 312.Prediction section 305 substantially has and the prediction section 101 identical structure of explanation and work(in the 1st embodiment Energy.

Prediction section 305 input can be using block message 60, selection block message 61, reference movement information 39 and reference picture picture letter Numbers 37, output prediction image signal 35 and movable information 38.Direction in space movable information obtaining section 310 and time orientation fortune Dynamic information acquiring section 311 is respectively provided with the direction in space movable information obtaining section 110 illustrating in the 1st embodiment with timely Between direction movable information obtaining section 111 identical function.Direction in space movable information obtaining section 310 use can utilize block message 60 And reference movement information 39, generate the motion letter including the movable information of each available block and index positioned at direction in space Breath 38A.Time orientation movable information obtaining section 311 use can utilize block message 60 and reference movement information 39, generate and include The movable information of each available block and the movable information of index positioned at time orientation（Or the group of movable information）38B.

In movable information switching switch 312, according to selecting block message 61, from from direction in space movable information obtaining section 310 movable information 38A and the movable information from time orientation movable information obtaining section 311（Or the group of movable information） Select one in 38B, obtain movable information 38.Selected movable information 38 is sent to dynamic compensating unit 313 and motion letter Breath memorizer 306.Dynamic compensating unit 313 is mended according to the motion of selected movable information 38 and explanation in the 1st embodiment Repay portion 113 and similarly carry out motion compensated prediction, generate prediction image signal 35.

Phase in the proportional zoom function of the motion vector of dynamic compensating unit 313, with explanation in the 1st embodiment Same, therefore omit the description.

Figure 22 illustrates the grammatical structure in picture decoding portion 300.As shown in figure 22, grammer mainly includes 3 parts, that is, high Level grammer 901, slice-level grammer 904 and microlith level grammer 907.High-level syntax 901 keeps the language of the above upper layer of section Method information.Slice-level grammer 904 is directed to the information required for each section keeps, and microlith level grammer 907 is directed to Fig. 7 A to Fig. 7 D Each shown microlith keeps required data.

Each several part includes more detailed grammer.High-level syntax 901 includes sequential parameter group grammer 902 and graphic parameter group The sequences such as grammer 903 and the grammer of figure level.Slice-level grammer 904 includes cut into slices head grammer 905 and slice of data grammer 906 etc..In addition, microlith level grammer 907 includes microlith layer grammer 908 and microlith prediction grammer 909 etc..

Figure 23 A and Figure 23 B illustrates the example of microlith layer grammer.Available_ shown in Figure 23 A and Figure 23 B Block_num represents the quantity that can utilize block, in the case of being more than 1 value when it, needs to select the decoding of block message.Separately Outward, stds_idx illustrates to select block message, using code table corresponding with aforesaid available block number, stds_idx is compiled Code.

Grammer when Figure 23 A is decoded to selection block message after being shown in mb_type.Pre- shown in as mb_type Survey pattern is the size determining or the pattern determining（TARGET_MODE）In the case of, and in available_ Block_num be more than 1 value in the case of, stds_idx is decoded.For example, the movable information of selection block become can profit Be that block size is 64 × 64 pixels, stds_idx is encoded in the case of 32 × 32 pixels, 16 × 16 pixels, or In the case of Direct Model, stds_idx is encoded.

Grammer when Figure 23 B is decoded to selection block message before being shown in mb_type.Work as available_block_ Num be more than 1 value in the case of, stds_idx is decoded.In addition, if available_block_num is 0, then Carry out H.264 representative conventional motion compensation, so encoding to mb_type.

In the ranks being likely to of table shown in Figure 23 A and Figure 23 B illustrates unspecified grammatical feature in the present invention, also may be used To include the description relevant with condition difference in addition.Or it is also possible to syntax table is split, merges into multiple tables.Separately Outer it is not necessary to must be using identical term it is also possible to arbitrarily be changed according to the mode of utilization.In addition, in this microlith layer Each grammer segment describing in grammer can also be changed to be explicitly recited in microlith data syntax described later.

As described above, the picture decoding apparatus of present embodiment are to by the picture coding dress of aforesaid 1st embodiment Put the image encoding to be decoded.Therefore, the picture decoding of present embodiment can reproduce according to smaller coded data Height is as the decoding image of matter.

（4th embodiment）

Figure 34 roughly illustrates the picture decoding apparatus of the 4th embodiment.As shown in figure 34, picture decoding apparatus possess figure As lsb decoder 400, decoding control section 350 and output state 308.The picture decoding apparatus of the 4th embodiment and the 2nd are in fact The picture coding device applying mode corresponds to.In the 4th embodiment, mainly to the part different from the 3rd embodiment and dynamic Illustrate.As shown in figure 34, in the picture decoding portion 400 of present embodiment, row lsb decoder 401 and prediction section are encoded 405 is different from the 3rd embodiment.

The prediction section 405 of present embodiment optionally switches following two prediction mode, and generates prediction image signal 35, this two kinds of prediction mode include：Carry out the prediction mode of motion compensation using the movable information that selection block has（1st is pre- Survey mode）；As H.264, carry out the prediction mode of motion compensation using a motion vector for decoder object block（The 2 prediction mode）.

Figure 35 is the block diagram illustrating in greater detail coding row lsb decoder 401.Coding row lsb decoder 401 shown in Figure 35 is in figure It is further equipped with movable information lsb decoder 424 in the structure of coding row lsb decoder 301 shown in 32.In addition, the choosing shown in Figure 35 Select block lsb decoder 423 different from the selection block lsb decoder 323 shown in Figure 32, coded data 80C relevant with selection block is carried out Decoding, obtains predicting handover information 62.Prediction handover information 62 represents that the prediction section 101 in the picture coding device of Fig. 1 uses Which in 1st and the 2nd prediction mode.When prediction handover information 62 represents that prediction section 101 employs the 1st prediction mode In the case of, that is, in the case of utilizing the 1st prediction mode that decoder object block is encoded, selection block lsb decoder 423 is to coded data Selection block message in 80C is decoded, and obtains selecting block message 61.When prediction handover information 62 represents that prediction section 101 uses In the case of 2nd prediction mode, that is, utilize the 2nd prediction mode to decoder object block coding in the case of, selection block lsb decoder 423 to not selecting block message decoding, and movable information lsb decoder 424 decodes to coded movable information 80D, obtains movable information 40.

Figure 36 is the block diagram illustrating in greater detail prediction section 405.Prediction section 405 shown in Figure 34 possesses the 1st prediction section 305th, the 2nd prediction section 410 and Forecasting Methodology switching switch 411.2nd prediction section 410 decodes using by coding row lsb decoder 401 Movable information 40 and with reference to picture signal 37, carry out the motion compensated prediction same with the dynamic compensating unit 313 of Figure 33, Generate prediction image signal 35B.1st prediction section 305 is identical with the prediction section 305 of explanation in the 3rd embodiment, generates prediction Picture signal 35B.In addition, Forecasting Methodology switching switch 411 is according to prediction handover information 62, from pre- from the 2nd prediction section 410 Altimetric image signal 35B and select one party, and as prediction section in the prediction image signal 35A of the 1st prediction section 305 405 prediction image signal 35 and export.Meanwhile, Forecasting Methodology switching switch 411 by selected in the 1st prediction section 305 or In person the 2nd prediction section 410, used movable information is sent to movable information memorizer 306 as movable information 38.

Then, with regard to the grammatical structure relevant with present embodiment, mainly the point different from the 3rd embodiment is said Bright.

Figure 30 A and Figure 30 B is shown respectively the example of the microlith layer grammer of present embodiment.Shown in Figure 30 A Available_block_num represents the quantity that can utilize block, in the case of being more than 1 value when it, selection block lsb decoder 423 Selection block message in coded data 80C is decoded.In addition, whether stds_flag is to represent in motion compensated prediction Labelling that the movable information of selection block is employed as the movable information of decoder object block, represent that Forecasting Methodology switching is opened Close 411 and have selected which side the labelling in the 1st prediction section 305 and the 2nd prediction section 410.When the quantity of available block is more than In the case that 1 and stds_flag is 1, represents and employ the movable information that selection block has in motion compensated prediction.Separately Outward, in the case that stds_flag is 0, the movable information that do not had using selection block, and directly right in the same manner as H.264 The information of movable information is encoded or the difference value predicted is encoded.In addition, stds_idx represents that selection block is believed Breath, code table corresponding with available block number is as previously mentioned.

Grammer when Figure 30 A is decoded to selection block message after being shown in mb_type.Only shown in mb_type In the case that predictive mode is the block size determining or the pattern determining, stds_flag and stds_idx is solved Code.For example, in the case that block size is 64 × 64,32 × 32,16 × 16, or in the case of Direct Model, to stds_ Flag and stds_idx is decoded.

Grammer when Figure 30 B is decoded to selection block message before being shown in mb_type.For example when stds_flag is 1 In the case of it is not necessary to mb_type decode.In the case that stds_flag is 0, mb_type is decoded.

As previously discussed, the picture decoding apparatus of present embodiment are to the picture coding by aforesaid 2nd embodiment The image of device code is decoded.Therefore, the picture decoding of present embodiment can according to smaller coded data again Existing height is as the decoding image of matter.

In addition, the invention is not restricted to above-mentioned embodiment itself, in implementation phase, energy in the range of without departing from its main idea Enough structural element is deformed and is embodied.In addition, by structural elements multiple disclosed in above-mentioned embodiment suitable group Close, various inventions can be formed.For example, it is also possible to delete several structures the entire infrastructure key element shown in from embodiment will Element.Alternatively, it is also possible to be combined as the structural element in different embodiments.

As this, even if as described below the 1st to the 4th above-mentioned embodiment is carried out deforming also can obtain equally Effect.

（1）In the 1st to the 4th embodiment, it is divided into the rectangular blocks such as 16 × 16 block of pixels and presses to by process object frame Example when being encoded or decode according to the order of the block of pixels to bottom right for the block of pixels from picture upper left as shown in Figure 4 enters Go explanation, but coding or decoding order have been not limited to this example.For example, coding or decoding order both can be from lower right to The order of upper left or from upper right order to left down.In addition, coding or decoding order both can be the central authorities from picture Portion helically the order to periphery or the periphery from picture to the order of central part.

（2）In the 1st to the 4th embodiment, to be defined to the chrominance signal of regardless of luminance signal and colour difference signal In case of component, son is illustrated.But, both can be using different predictions for luminance signal and colour difference signal Process, or can also be processed using identical prediction.In the case of being processed using different predictions, will be for colour difference signal The Forecasting Methodology selecting, is encoded/decoded using the method same with luminance signal.

Further it is evident that, implement various modifications in the scope without departing from idea of the invention and similarly can implement.

Industrial applicability

The image coding/decoding method of the present invention can improve code efficiency, so having industrial applicability.

Claims (2)

1. a kind of picture decoding method is it is characterised in that have：

1st step, receives the input of the pattern information related to the predictive mode of decoder object block；

Second step, illustrates the information specifying correspondingly with described pattern information, the pixel finishing from the decoding with movable information Multiple motion reference blocks are selected in block；

Third step, selects at least one available block from the plurality of motion reference block, and at least one available block described is There is the block of pixels of the candidate of movable information being applied to described decoder object block, and there is mutually different movable information；

4th step, the code regulation set in advance based on the quantity according to described available block, to the coded data being inputted It is decoded, thus obtaining the selection information for determining selection block；

5th step, according to described selection information, selects a selection block from described available block；

6th step, generates the prognostic chart picture of described decoder object block using the movable information of described selection block；

7th step, is decoded to the forecast error of described decoder object block according to described coded data；And

8th step, obtains decoding image according to described prognostic chart picture and described forecast error,

The plurality of motion reference block include with described decoder object block identical frame multiple pieces and with described decoder object Multiple pieces of the different frame of block,

In described third step, described motion reference block have movable information and this movable information with corresponding to by In the case of being judged to that the movable information of available motion reference block is inconsistent, it is judged to block can be utilized.

2. a kind of picture decoding apparatus are it is characterised in that possess：

Input unit, receives the input of the pattern information related to the predictive mode of decoder object block；

Available block obtaining section, illustrates the information specifying correspondingly with described pattern information, complete from the decoding with movable information Select multiple motion reference blocks in complete block of pixels, select at least one available block, institute from the plurality of motion reference block Stating at least one available block is the block of pixels with the candidate of movable information being applied to described decoder object block, and has phase Mutually different movable information；

1st lsb decoder, the code regulation set in advance based on the quantity according to described available block, to the coded number being inputted According to being decoded, thus obtaining the selection information for determining selection block；

Selector, according to described selection information, selects a selection block from described available block；

Prediction section, generates the prognostic chart picture of described decoder object block using the movable information of described selection block；

2nd lsb decoder, is decoded to the forecast error of described decoder object block according to described coded data；And

Adder, obtains decoding image according to described prognostic chart picture and described forecast error,

The plurality of motion reference block include with described decoder object block identical frame multiple pieces and with described decoder object Multiple pieces of the different frame of block,

In described available block obtaining section, described motion reference block there is movable information and this movable information with corresponding In the case that the movable information being judged as available motion reference block is inconsistent, it is judged to block can be utilized.