CN101573982B - Method and apparatus for encoding/decoding image using motion vector tracking - Google Patents
Method and apparatus for encoding/decoding image using motion vector tracking Download PDFInfo
- Publication number
- CN101573982B CN101573982B CN2007800491266A CN200780049126A CN101573982B CN 101573982 B CN101573982 B CN 101573982B CN 2007800491266 A CN2007800491266 A CN 2007800491266A CN 200780049126 A CN200780049126 A CN 200780049126A CN 101573982 B CN101573982 B CN 101573982B
- Authority
- CN
- China
- Prior art keywords
- corresponding region
- reference picture
- current block
- motion vector
- block
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Expired - Fee Related
Links
Images
Abstract
The present invention discloses a method and apparatus for encoding/decoding an image using motion vector tracking. The image encoding method includes determining corresponding areas of a plurality ofreference pictures that are to be used to predict a current block by tracking a motion vector route of a corresponding area of a reference picture referred to by the current block; generating a predi ction block of the current block by calculating a weighted sum of the corresponding areas of the plurality of reference pictures; and encoding a difference between the current block and the predictionblock.
Description
Technical field
The method and apparatus consistent with the present invention relates to the predictive coding/decoding of image, more specifically, relate to like this image is carried out coding/decoding, the route of the motion vector of the current picture of its Continuous Tracking, determine a plurality of reference pictures, and use this this current picture of reference picture prediction coding.
Background technology
When encoded video, the room and time redundancy of image sequence is eliminated to compress this image sequence.In order to eliminate time redundancy, be positioned at reference picture before or after the present encoding picture and be used to search zone to the regional similar reference picture of the picture of present encoding, detect the motion between the respective regions of the picture of present encoding and reference picture, and coding is by the residual error between the picture of the predictive picture carrying out motion compensation based on the motion of this detection and obtain and present encoding.
Video pictures is encoded among one or more bursts (slice).A burst comprises at least one macro block.Video pictures can be encoded in burst.According to standard H.264, video pictures is encoded into intra (I) burst (encoding) in a picture, prediction (P) burst (using a reference picture to encode) and bi-directional predicted (B) burst (using two reference picture prediction image samplings to encode).
In motion picture expert group 2 (MPEG-2) standard, use current picture picture and current picture picture afterwards before to carry out bi-directional predicted as the reference picture.According to advanced video coding (AVC) H.264/, bi-directional predictedly can use any two pictures, and be not restricted to before the current picture and picture afterwards as the reference picture.Be defined as bi-directional predicted picture (hereinafter being called " B picture ") by the picture that uses two predictive pictures.
Fig. 1 is that explanation is the figure of piece forecasting process of the current picture of B picture according to standard code H.264/AVC.H.264/AVC standard is by using two reference picture A and B (as macro block MB 1) at equidirectional, use two reference picture B and C (as macro block MB2) at different directions, use in two zones (as macro block MB3) of two zoness of different samplings of same reference picture A or use an optional reference picture B or D (as macro block MB4 or MB5) predicting the piece of B picture.
Generally, the view data that is encoded to the B picture has than the higher code efficiency of view data that is encoded to I or P picture.Use the B picture of two reference pictures can produce with respect to the P picture that uses a reference picture or the I picture that uses the picture intra-prediction more similar in appearance to the prediction data of current image date.In addition,,, can not cause big distortion, as having carried out certain low frequency filtering even between two reference pictures, error takes place because the B picture uses the mean value of two reference pictures as prediction data yet.
Because the B picture uses two reference pictures to obtain the code efficiency higher than P picture, if use more multiple reference pictures in prediction, then code efficiency increases., if carry out motion prediction and compensation in each reference picture, then operand increases, thereby the Standard of image compression of correlation technique is set to maximum two reference pictures.
Summary of the invention
The present invention proposes a kind of method and apparatus that is used for encoding/decoding image, and the motion vector route of the reference picture of its tracking current block is to use more multiple reference pictures prediction current block, so that improve coding/decoding efficient.
According to example embodiment of the present invention, the reference picture of greater number is used to the predictive coding current block, thereby improves the efficient of prediction and coding.
Description of drawings
Fig. 1 is that explanation is the figure of piece forecasting process of the current picture of bi-directional predicted (B) picture according to standard code H.264/AVC;
To be explanation determine to be used for to predict the figure of process of a plurality of reference pictures of current picture according to the method for encoding images of example embodiment of the present invention to Fig. 2;
To be explanation determine to be used for to predict the figure of process of a plurality of reference pictures of current picture according to the method for encoding images of another example embodiment of the present invention to Fig. 3;
To be explanation determine to be used for to predict the figure of process of a plurality of reference pictures of current picture according to the method for encoding images of another example embodiment of the present invention to Fig. 4;
Fig. 5 is the block diagram according to the picture coding device of example embodiment of the present invention;
Fig. 6 is the block diagram in the motion compensation units of Fig. 5 explanation according to example embodiment of the present invention;
Fig. 7 is the figure of piece of the different sizes of the explanation example embodiment according to the present invention motion that is used to predict variable-block in standard H.264/MPEG-4AVC;
Fig. 8 is the image according to the motion generation of passing through the prediction variable-block of example embodiment of the present invention;
Fig. 9 is used to explain according to the method for encoding images of example embodiment of the present invention determine by the figure along the process of the corresponding region of other reference pictures of the sub-corresponding region reference of the reference picture of moving mass boundary demarcation;
Figure 10 is used to explain according to the method for encoding images of another example embodiment of the present invention determine by the figure along the process of the corresponding region of other reference pictures of the sub-corresponding region reference of the reference picture of moving mass boundary demarcation;
To be explanation come dispensed to give the figure of process of weight of the corresponding region of reference picture according to the method for encoding images of example embodiment of the present invention to Figure 11;
Figure 12 is the flow chart of explanation according to the method for encoding images of example embodiment of the present invention;
Figure 13 is the block diagram according to the picture decoding apparatus of example embodiment of the present invention;
Figure 14 is the flow chart of explanation according to the picture decoding method of example embodiment of the present invention.
Embodiment
According to an aspect of the present invention, provide a kind of method for encoding images, comprising: by following the tracks of the corresponding region of a plurality of reference pictures of determining to be used to predict the current block of current picture by the motion vector route of the corresponding region of the reference picture of current block reference; The weighted sum of the corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; And the coding current block with predict poor between the piece.
According to a further aspect in the invention, a kind of picture coding device is provided, comprise: the reference picture determining unit, it is by following the tracks of the corresponding region of being determined to be used to predict a plurality of reference pictures of current block by the motion vector route of the corresponding region of the reference picture of current block reference; The weight estimation unit, the weighted sum of its corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; And coding unit, its coding current block with predict poor between the piece.
According to a further aspect in the invention, provide a kind of picture decoding method, comprising: discern the predictive mode of current block by reading the prediction mode information that is included in incoming bit stream; If current block is confirmed as using the corresponding region of a plurality of reference pictures to predict, then by following the tracks of the corresponding region of determining to be used to predict a plurality of reference pictures of current block by the motion vector route of the corresponding region of the corresponding region of the reference picture of the motion vector route reference that is included in the current block in the bit stream and reference picture; The weighted sum of the corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; And by being included in current block and the difference of prediction between the piece and the prediction piece Calais's decoding current block mutually in the bit stream
According to a further aspect in the invention, provide a kind of picture decoding apparatus, comprising: the predictive mode recognition unit, it discerns the predictive mode of current block by reading the prediction mode information that is included in the incoming bit stream; The reference picture determining unit, if current block is confirmed as using the corresponding region of a plurality of reference pictures to predict, then it is by following the tracks of the corresponding region of being determined to be used to predict a plurality of reference pictures of current block by the motion vector route of the corresponding region of the corresponding region of the reference picture of the motion vector route reference that is included in the current block in the bit stream and reference picture; The Weight prediction unit, the weighted sum of its corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; And decoding unit, it will be by being included in current block and the difference of prediction between the piece and the prediction piece Calais's decoding current block mutually in the bit stream.
Describe example embodiment of the present invention in detail referring now to accompanying drawing.
Method for encoding images according to example embodiment of the present invention uses the motion vector of the reference picture of being indicated by the motion vector of current picture to come the corresponding region of other reference pictures of Continuous Tracking, definite thus a plurality of reference pictures that will be used for the prediction of current picture, calculate the weighted sum of a plurality of reference pictures, and the predicted value that produces current picture.
The process of the definite a plurality of reference pictures that used by method for encoding images and device according to example embodiment of the present invention is described to Fig. 4 referring now to Fig. 2.
To be explanation determine to be used for to predict the figure of process of a plurality of reference pictures of current picture according to the method for encoding images of example embodiment of the present invention to Fig. 2.
With reference to figure 2, suppose the piece 21 (hereinafter referred to as " current block ") that will be encoded in current picture A is carried out the general motion prediction, determine motion vector MV1 thus, the corresponding region 22 of the reference picture 1 that its indication and current block 21 are the most similar.Suppose that also current picture A is prediction (P) picture, current block 21 is only with reference to the moving mass of a reference picture.Yet the present invention can be applicable to that tracking has each motion vector of the moving mass of two motion vectors in bi-directional predicted (B) picture, just as following the tracks of at the moving mass with a motion vector shown in the P of Fig. 2 picture.
With reference to figure 2, the motion vector MV1 indication that is produced by the motion prediction of current block 21 relatively has minimum error with current block 21 in reference picture 1 zone.In the prior art, the value of the corresponding region 22 of reference picture 1 is defined as the predicted value of current block 21, and coded residual, i.e. poor between the original pixel value of predicted value and current block 21.
The method for encoding images of example embodiment of the present invention is by using the corresponding region (as prior art) of first reference picture of being indicated by the motion vector of current block, and also be used for utilizing the movable information of the corresponding region of reference picture to predict the corresponding region of second reference picture of the corresponding region of first reference picture by use, predict current block.For example, be used to determine with the motion vector MV2 of the corresponding region 22 of current block 21 corresponding reference pictures 1 and be used for the corresponding region 23 of reference picture 2 of corresponding region 22 of prediction reference pictures 1.The motion vector MV3 of the corresponding region 23 of reference picture 2 is used to determine and is used for the corresponding region 24 of reference picture 3 of corresponding region 23 of prediction reference pictures 2.The motion vector MVn of the corresponding region 25 of reference picture n-1 is used to determine and is used for the corresponding region 26 of reference picture A of corresponding region 25 of prediction reference pictures n-1.As will be described later, tracking is by the corresponding region of first reference picture of the motion vector of current block indication or follow the tracks of by the process of the corresponding region of second reference picture of the motion vector indication of the corresponding region of first reference picture and carried out continuously, only comprises that up to reference picture intra-frame prediction block that intra-frame prediction block or reference picture comprise has the corresponding region greater than threshold value.
In this example embodiment, by pursuit movement vector route, predetermined weight be multiply by in each corresponding region of a plurality of reference pictures, and results added is produced the prediction piece of current block 21, described motion vector route is such as being motion vector route by the corresponding region 22 of the reference picture 1 of the motion vector MV1 of current block 21 indication, by the motion vector route of the corresponding region 23 of the reference picture 2 of the motion vector MV2 indication of the corresponding region 22 of reference picture 1, and by the motion vector route of the corresponding region 24 of the reference picture 3 of the motion vector MV3 indication of the corresponding region 23 of reference picture 2.
To be explanation determine to be used for to predict the figure of process of a plurality of reference pictures of current picture according to the method for encoding images of another example embodiment of the present invention to Fig. 3.With reference to figure 3, suppose I
0Be intra (I) picture, P
1And P
5Be the P picture, B
2And B
3It is the B picture.Now explanation is identified for predicting B
2The process of the corresponding region of a plurality of reference pictures of the current block 31 of picture.
Suppose B
2The current block 31 of picture has two motion vector MV1 and MV2, and this is the general motion prediction result.If resemble B
2The current block 31 of picture is the same, and the current block that will encode has two motion vectors, and then each motion vector route of the corresponding region of track reference picture is to determine the corresponding region of reference picture.Reference picture P by the first motion vector MV1 of current block 31 indication
1The movable information of corresponding region 33 be used to determine and be used for prediction reference pictures P
1The reference picture I of corresponding region 33
0Corresponding region 34.Because reference picture I
0Be the I picture that comprises intra-frame prediction block, reference picture I
0Corresponding region 34 do not have movable information, therefore stop to follow the tracks of.
Similarly, by the reference picture B of the second motion vector MV2 of current block 31 indication
3Corresponding region 32 have two motion vectors because reference picture B
3It is the B picture.At reference picture B
3Two motion vectors of corresponding region 32 in, the motion vector on the left side is tracked to determine to be used for prediction reference pictures B
3The reference picture P of corresponding region 32
1Corresponding region 41 and be used for prediction reference pictures P
1The reference picture I of corresponding region 41
0Corresponding region 42.The motion vector on the right is tracked to determine to be used for prediction reference pictures B
3The reference picture P of corresponding region 32
5Corresponding region 38.As mentioned above, reference picture B
3The tracing process of the right motion vector of corresponding region 32 will carry out continuously, till the intra-frame prediction block that does not have movable information or the reference picture that comprises intra-frame prediction block have corresponding region greater than threshold value.
In this example embodiment, two motion vector MV1 and MV2 by following the tracks of current block 31 multiply by predetermined weight with the corresponding region 32,33,34 of a plurality of reference pictures and each of 38, and with results added, produce the prediction piece of current block 31.
To be explanation determine to be used for to predict the figure of process of a plurality of reference pictures of current picture according to the method for encoding images of another example embodiment of the present invention to Fig. 4.Except only use the picture of encoding before current picture, the motion vector tracking process of this example embodiment is similar to the process of the last example embodiment of describing with reference to figure 3.According to standard H.264/AVC, reference picture is not restricted to before the current picture and two pictures afterwards, and can be two pictures on optional direction.Therefore, with reference to figure 4, can only use the picture before current picture to carry out predictive coding.
Follow the tracks of two motion vector MV1 of current block 43 and MV2 and be identified for predicting the corresponding region 44 to 52 of a plurality of reference pictures of current block 43.
As mentioned above, the method for encoding images of example embodiment of the present invention and device use the corresponding region by the reference picture of the motion vector indication of current block, with be used for the movable information of the corresponding region by using this reference picture and predict the corresponding region of other reference pictures of the corresponding region of this reference picture, to predict this current block.If the corresponding region of current block or reference picture has two motion vectors, then follow the tracks of each motion vector to determine the corresponding region of other reference pictures.
Fig. 5 is the block diagram according to the picture coding device 500 of example embodiment of the present invention.For convenience of description, suppose that this picture coding device follows H.264/AVC standard.Yet the picture coding device of example embodiment of the present invention can be applicable to the different method for encoding images that uses motion prediction and compensation.
With reference to figure 5, picture coding device 500 comprises motion estimation unit 502, motion compensation units 504, intraprediction unit 506, converter unit 508, quantifying unit 510, rearrangement unit 512, entropy coding unit 514, inverse quantization unit 516, inverse transformation block 518, filter unit 520, frame memory 522 and control unit 525.
The predicted value of 506 pairs of current blocks of intraprediction unit is carried out infra-frame prediction.
In case the Forecasting Methodology of the corresponding region of a plurality of reference pictures of use by inter prediction, infra-frame prediction or example embodiment produces the prediction piece of this current block, then corresponding with current block and the error amount of prediction between piece residual error is produced, and be converted into frequency domain by converter unit 508, quantize by quantifying unit 510 then.The residual error of entropy coding unit 514 codings through quantizing, output bit flow thus.
Picture through quantizing is rebuild so that obtain reference picture by inverse quantization unit 516 and inverse transformation block 518.The current picture of rebuilding is stored in the frame memory 522, for use in the prediction of next picture then through carrying out the filter unit 520 of eliminating blocking effect.
Fig. 6 is the block diagram in the motion compensation units 504 of Fig. 5 explanation according to example embodiment of the present invention.With reference to figure 6, the motion compensation units 600 of example embodiment of the present invention comprises reference picture determining unit 610 and weight estimation unit 620.
Reference picture determining unit 610 uses the motion vector of the current block that is produced by motion estimation unit 502 to determine the corresponding region of reference picture, and the route of the motion vector of the corresponding region of track reference picture, the corresponding region of determining to be used to predict a plurality of reference pictures of current block thus.
The weighted sum of the corresponding region of a plurality of reference pictures of weight estimation unit 620 calculating is to produce the prediction piece of current block.Weight estimation unit 620 comprise the corresponding region of determining a plurality of reference pictures weight weight calculation unit 621 and the corresponding region of a plurality of reference pictures be multiply by weight and with the prediction piece generation unit 622 of results added with the prediction piece that produces current block.
To describe the operation of reference picture determining unit 610 of the corresponding region of a plurality of reference pictures of determining to be used to predict current block now in detail.
Fig. 7 is the figure of piece of the different sizes of the explanation motion that is used to predict variable-block in standard H.264/MPEG-4AVC.Fig. 8 is the image by the motion generation of prediction variable-block.
With reference to figure 7, can use four kinds of methods to divide macro blocks: macro block can be divided into that a 16x16 macroblock partition, two 16x8 are cut apart, two 8x16 are cut apart or four 8x8 are cut apart, with the motion of predicted macroblock.In the 8x8 pattern, can use each of four 8x8 macro blocks of four kinds of methods divisions: each 8x8 macro block can be divided into a 8x8 macroblock partition, two 8x4 macroblock partition, two 4x8 macroblock partition or four sub-macroblock partition of 4x4.These various combinations of cutting apart with sub-macro block can realize in each macro block.Macro block is called tree-shaped motion compensation to this type of division of the sub-piece of different sizes.
With reference to figure 8, the motion that has low-energy in image predicts that with big cutting apart the motion that has high-octane in image is predicted with little cutting apart.Border between the moving mass that is used to use tree-shaped motion compensation to divide current picture is defined as the moving mass border.
As mentioned above, according to the motion vector of the corresponding region of the method for encoding images track reference picture of example embodiment of the present invention, so that determine to be used to predict the corresponding region of a plurality of reference pictures of current block., as shown in Figure 8 because reference picture is divided into the moving mass with different sizes, with the corresponding region of current block corresponding reference picture not exact match but be included in a plurality of moving mass in a certain moving mass.In this case, the corresponding region of reference picture comprises a plurality of motion vectors.The process of following the tracks of a plurality of motion vectors in the corresponding region that is included in reference picture will be described now.
Fig. 9 is used to explain according to the method for encoding images of example embodiment of the present invention determine by the figure along the process of the corresponding region of other reference pictures of the sub-corresponding region reference of the reference picture of moving mass boundary demarcation.With reference to figure 9, be included in a plurality of moving mass by the corresponding region 91 of the reference picture 1 of the motion vector MV1 of current block 90 indication.More specifically, with the corresponding region 91 of current block 90 corresponding reference pictures 1 not with one of them couplings of a plurality of moving mass, but be included among moving mass A, B, C, the D.In this case, reference picture determining unit 610 is along the corresponding region 91 of the moving mass boundary demarcation reference picture 1 of reference picture 1, and determines by the reference picture 2 of the motion vector indication of each moving mass of the reference picture 1 that comprises sub-corresponding region a, b, c and d and 3 corresponding region.More specifically, reference picture determining unit 610 is determined the corresponding region a ' 93 of reference picture 2 by the motion vector MVa that uses the moving mass A under the sub-corresponding region a, determine the corresponding region b ' 94 of reference picture 2 by the motion vector MVb that uses the moving mass B under the sub-corresponding region b, determine the corresponding region c ' 96 of reference picture 3 by the motion vector MVc that uses the moving mass C under the sub-corresponding region c, and the corresponding region d ' 95 that determines reference picture 3 by the motion vector MVd that uses the moving mass D under the sub-corresponding region d.
In this example embodiment, part comprise with moving mass A, the B of the corresponding region 91 of current block 90 corresponding reference pictures 1, C, D with reference to reference picture 2 and 3.Yet, even if moving mass A, B, C, D change their reference picture, the sports ground information of moving mass A, B, C, D (being motion vector and the reference picture information of moving mass A, B, C, D) also can be used to determine the corresponding region of other reference pictures corresponding with the sub-corresponding region of reference picture 1.
Figure 10 is used to explain according to the method for encoding images of another example embodiment of the present invention determine by the figure along the process of the corresponding region of other reference pictures of the inferior corresponding region reference of the reference picture of moving mass boundary demarcation.With reference to Figure 10, when the corresponding region 100 about the reference picture of current block partly is included among moving mass A, B1, B2, C and the D, as mentioned above, the corresponding region 100 of reference picture is by the moving mass boundary demarcation along reference block, and the sports ground information of each moving mass under sub-corresponding region a, b1, b2, c and the d is used to determine the corresponding region of other reference pictures.More specifically, reference picture determining unit 610 is determined the corresponding region of other reference pictures corresponding with sub-corresponding region a by the sports ground information of using the affiliated moving mass A of sub-corresponding region a, determine the corresponding region of other reference pictures corresponding with sub-corresponding region b1 by the sports ground information of using the moving mass B1 under the sub-corresponding region b1, determine the corresponding region of other reference pictures corresponding with sub-corresponding region b2 by the sports ground information of using the moving mass B2 under the sub-corresponding region b2, determine the corresponding region of other reference pictures corresponding by the sports ground information of using the moving mass C under the sub-corresponding region c, by the corresponding region of definite other reference pictures corresponding of the sports ground information of using the moving mass D under the sub-corresponding region d with sub-corresponding region d with sub-corresponding region c.
Determine that the process of reference picture is used to determine according to the corresponding region of first reference picture of being indicated by the motion vector of current block the corresponding region of second reference picture, also is used for determining the 3rd reference picture according to the corresponding region of second reference picture.Only when the corresponding region is included in the moving mass with motion vector information, can carry out continuously based on the tracking of motion vector., when the corresponding region is included in the intra-frame prediction block or be included in corresponding region in the intra-frame prediction block greater than threshold value, carry out this tracking with reference to the corresponding reference picture.For example, again with reference to figure 9, if with piece A, B, C, D under the corresponding region 91 of current block 90 corresponding reference pictures 1 be intra-frame prediction block entirely, then no longer carry out and follow the tracks of and the corresponding region 91 of reference picture 1 only is used to predict current block 90.Equally, if block A, B, C are the moving mass with motion vector, and piece D is an intra-frame prediction block, as the sub-corresponding region d that belongs to piece D during greater than threshold value, are used to predict current block 90 by the corresponding region 91 of reference picture 1 being multiply by the value that weight obtains.Determine whether to carry out continuously the corresponding region that the process of following the tracks of correspondingly is applied to other reference pictures of determining according to reference picture.
Be included in corresponding region in the intra-frame prediction block if one of corresponding region is included in the intra-frame prediction block less than threshold value, then carry out the tracing process of the corresponding region of determining other reference pictures continuously.To this, use the motion vector of the adjacent moving mass of intra-frame prediction block to distribute a virtual motion vector, and determine corresponding region by other reference pictures of this virtual motion vector indication to this intra-frame prediction block.In above-mentioned example, suppose that piece A, B, C comprise motion vector MVa, MVb and MVc respectively, piece D is an intra-frame prediction block, and the sub-corresponding region d of piece D is less than threshold value, this tracing process is carried out continuously.In this case, as mentioned above, for sub-corresponding region a, the b and the c that belong to piece A, B, C, with motion vector MVa, MVb and the median of MVc or the virtual motion vector that mean value is distributed to piece D of piece A, B, C, and determine corresponding region by other reference pictures of this virtual motion vector indication.
With reference to figure 6, if reference picture determining unit 610 is followed the tracks of the corresponding region that the route of the motion vector of current block is also determined a plurality of reference pictures, weight calculation unit 621 dispensed are given the weight of whole corresponding regions again.
To be explanation come dispensed to give the figure of process of weight of the corresponding region of reference picture according to the method for encoding images of example embodiment of the present invention to Figure 11.With reference to Figure 11, suppose D
tThe expression current block, D
T-1Expression and current block D
tThe corresponding region of corresponding reference picture t-1, D
T-2, a, D
T-2, b, D
T-2, c and D
T-2, d (comprehensively is called ' D
T-2') expression respectively with corresponding region D
T-1Each corresponding region of sub-zoning a, b, c and d corresponding reference picture t-2, and P
tExpression current block D
tThe prediction piece.
[mathematical expression 1]
P
t＝α*D
t-1+β*D
t-2 ...(1)
Can use various algorithms to determine to distribute to the weight and the β of the corresponding region of reference picture.Example embodiment of the present invention is used at prediction piece P
tWith current block D
tBetween produce the weight of minimal error.Calculate prediction piece P according to equation 2
tWith current block D
tBetween variance and (SSE).
[mathematical expression 2]
SSE＝∑(D
t-P
t)
2＝∑[D
t-(α*D
t-1+β*D
t-2)]
2 ...(2)
Can 0 determine weight and β by calculating according to the partial differential equation (PDE) of equation 3 and obtaining the result.
[mathematical expression 3]
Use current block adjacent block pixel and come the PDE of calculation equation 3 with the neighbor of the corresponding region of the pixel corresponding reference picture of the adjacent block of current block.This is because can use the information about the pixel of the adjacent block of current block of early decoding to determine weight and need not to transmit the weight that is used for predicting current block.Therefore, example embodiment of the present invention use current block adjacent block pixel and with the neighbor of the corresponding region of the pixel corresponding reference picture of the adjacent block of current block, determine weight so that use by the pretreated data of encoder, avoided transmitting the needs of the weight of the corresponding region of distributing to reference picture.
Be similar to the corresponding region D that uses reference picture t-1
T-1Corresponding region D with reference picture t-2
T-2About current block D
tPrediction piece P
tCalculating, consider current block D
tThe locus, can be by using the corresponding region D of reference picture t-1
T-1Neighbor N
T-1, a, N
T-1, b and N
T-1, the corresponding region D of c and reference picture t-2
T-2Neighbor N
T-2, a, N
T-2, b and N
T-2, c calculates the pixel N of the adjacent block of current block
tIn this case, calculate by using the corresponding region D of reference picture t-1 according to equation 4
T-1Neighbor N
T-1Corresponding region D with reference picture t-2
T-2Neighbor N
T-2The current block D that obtains
tThe predicted value N of pixel of adjacent block
tThe pixel N of the adjacent block of ' current block
tBetween SSE.
[mathematical expression 4]
SSE=∑ (the N of neighbor
t-N
t')
2=∑ [N
t-(α * N
T-1+ β * N
T-2)]
2... (4)
In equation 1, if the value of weight and β by normalization, alpha+beta=1 so, β=1-α.β=1-α is substituted in the equation 1 to provide following equation 5 and 6.
[mathematical expression 5]
P
t＝α*D
t-1+(1-α)*D
t-2 ...(5)
[mathematical expression 6]
SSE＝∑(D
t-P
t)
2＝∑[D
t-(α*D
t-1+(1-α)*D
t-2)]
2?...(6)
According to equation 7, the PDE according to the SSE of equation 6 obtains to satisfy by calculating
[mathematical expression 7]
Weight.
[mathematical expression 8]
As mentioned above, use the pixel N of the adjacent block of first pre-treatment respectively
t, neighbor N
T-2, and neighbor N
T-1To replace current block D
t, corresponding region D
T-2, corresponding region D
T-1, need not transmit each weight of distributing to the corresponding region so that determine weight.
By assigning weight for each reference picture when the corresponding region of the reference picture that uses larger amt and determining that the weight that produces minimal error between current block and prediction piece determines weight.
More specifically, if D1, D2, D3 ... Dn represents to be used to predict current block D
tThe corresponding region of the individual reference picture of n (n is an integer), and W1, W2, W3 ... Wn represents to distribute to the weight of each corresponding region, then uses P
t=W1*D1+W2*D2+W3*D3+...+Wn*Dn calculates current block D
tPrediction piece P
t(promptly predict piece P by calculating SSE
tWith current block D
tBetween error amount square) PDE, use weight as parameter and to obtain the result be 0, determine weights W 1, W2, W3 ... Wn.As mentioned above, the neighbor of the correspondence of the corresponding region of the pixel of the adjacent block of current block and reference picture is used to calculate PDE.
With reference to figure 6, prediction piece generation unit 622 multiply by weight with the corresponding region of a plurality of reference pictures again, with results added, and the prediction piece of generation current block.
According to the motion compensation units 504 conversion residual errors of example embodiment of the present invention (its be to use between prediction piece that the corresponding region of a plurality of reference pictures obtains and the current block poor), quantize this residual error, and this residual error of entropy coding, output bit flow thus.
The sign of one bit can be inserted will be according to the header according to the method for encoding images bitstream encoded of example embodiment of the present invention, and whether this sign indicates each piece to use the corresponding region of a plurality of reference pictures to carry out motion prediction.For example, indicate according to the conventional art bitstream encoded " 0 ", and " 1 " indication is according to example embodiment bitstream encoded of the present invention.
Figure 12 is the flow chart of explanation according to the method for encoding images of example embodiment of the present invention.With reference to Figure 12, follow the tracks of by the motion vector route of the corresponding region of the reference picture of current block reference corresponding region (operation 1210) with a plurality of reference pictures of being identified for predicting this current block.As mentioned above, when pressing the corresponding region of this reference picture of moving mass boundary demarcation, the motion vector of the moving mass under each corresponding region is used to determine the corresponding region of a plurality of reference pictures.
The definite weight (operation 1220) that will distribute to the corresponding region of a plurality of reference pictures.As mentioned above, the neighbor of the corresponding region of neighbor by using current block and a plurality of reference pictures corresponding with the neighbor of current block is defined as such value with the weight of corresponding region: it minimizes the difference between the neighbor of original neighbor and the current block predicted from the neighbor of corresponding region.
Produce the prediction piece (operation 1230) of current block by each value phase Calais of the corresponding region of a plurality of reference pictures be multiply by the weight acquisition.
Residual error as the difference between prediction piece and the current block is transformed, quantification and entropy coding, output bit flow (operation 1240) thus.
Figure 13 is the block diagram according to the picture decoding apparatus 1300 of example embodiment of the present invention.With reference to Figure 13, the picture decoding apparatus 1300 of example embodiment of the present invention comprises entropy decoding unit 1310, rearrangement unit 1320, inverse quantization unit 1330, inverse transformation block 1340, motion compensation units 1350, intraprediction unit 1360 and filter unit 1370.
Entropy decoding unit 1310 and rearrangement unit 1320 receive the bit stream of compression and the bit stream that receives are carried out the entropy decoding, produce the coefficient that quantizes thus.Inverse quantization unit 1330 and inverse transformation block 1340 are carried out the re-quantization and the inverse transformation of quantization parameter, extract the transition coding coefficient thus, motion vector information and prediction mode information.Prediction mode information can comprise sign, and whether its indication current block to be decoded is to encode by the corresponding region interpolation weight of using a plurality of reference pictures according to the method for encoding images of example embodiment of the present invention.As mentioned above, can use the motion vector information of the current block of the decoding corresponding region of a plurality of reference pictures of current block of determining to be used for to decode, so there is no need to transmit information about the corresponding region of a plurality of reference pictures of the current block that will be used for decoding in the mode the same with method for encoding images.
Intraprediction unit 1360 uses the adjacent block of current block to produce the prediction piece of current block, and this adjacent block is decoded before the intraframe predictive coding current block.
Motion compensation units 1350 is operated in the mode identical with the motion compensation units 504 of Fig. 5 explanation.In other words, when the weighted sum of the corresponding region by calculating a plurality of reference pictures and during predictive coding current block to be decoded, the corresponding region that motion compensation units 1350 uses the motion vector that is included in the current block in the bit stream to follow the tracks of the reference picture of previous coding, determine the corresponding region of a plurality of reference pictures thus, the definite weight that will distribute to the corresponding region of each reference picture, weight be multiply by in the corresponding region of a plurality of reference pictures, with results added, and the predicted value of generation current block.As mentioned above, use the neighbor of the corresponding region of the neighbor of current block of early decoding and a plurality of reference pictures corresponding to determine weight to the corresponding region of a plurality of reference pictures with the neighbor of current block.
From bit stream, extract the error amount D ' n between current block and the prediction piece, with itself and the prediction piece addition that produces by motion compensation units 1350 and intraprediction unit 1360, produce the video data uF ' n that rebuilds thus then.UF ' n finishes the decoding of current block thus through filter unit 1370.
Figure 14 is the flow chart of explanation according to the picture decoding method of example embodiment of the present invention.With reference to Figure 14, the prediction mode information that is included in the incoming bit stream is read so that discern the predictive mode (operation 1410) of current block.
If current block to be decoded is confirmed as being to use the corresponding region of a plurality of reference pictures to predict, then follow the tracks of by the corresponding region of the reference picture of the motion vector route reference of the corresponding region of motion vector route that is included in the current block in the bit stream and reference picture to determine to be used to predict the corresponding region (operation 1420) of a plurality of reference pictures of current block.
The neighbor of the corresponding region of the neighbor of the current block of early decoding and a plurality of reference pictures corresponding with the neighbor of current block is used to determine the weight of the corresponding region of distributing to a plurality of reference pictures, and the weighted sum of the corresponding region by calculating a plurality of reference pictures produces the prediction piece (operation 1430) of current block.
With the predicted value of current block be included in current block in the bit stream and the poor addition between the predicted value, the current block of decoding thus (operation 1440).
Example embodiment of the present invention also can be embodied as the computer-readable code on the computer-readable recording medium.Computer readable recording medium storing program for performing is any data storage device that can store subsequently by the data of computer system reads.The example of computer readable recording medium storing program for performing comprises read-only memory (ROM), random access storage device (RAM), CD-ROM, tape, floppy disk and light data storage device.Thereby computer readable recording medium storing program for performing also can be distributed in the computer system of network coupled with distribution mode storage and computer readable code executed.
Although partly illustrate and describe the present invention with reference to example embodiment, but those skilled in the art are to be understood that, do not breaking away under the situation of the spirit and scope of the present invention that limit by following claims, can carry out the various modifications on form and the details therein.
Claims (22)
1. method for encoding images comprises:
By following the tracks of the corresponding region of a plurality of reference pictures of determining to be used to predict the current block of current picture by the motion vector route of the corresponding region of the reference picture of current block reference;
The weighted sum of the corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; And
The coding current block with predict poor between the piece,
Wherein the step of the motion vector route of the corresponding region of track reference picture comprises: determine the corresponding region by second reference picture of the motion vector indication of current block, corresponding region up to the n reference picture of determining to indicate by the motion vector of n-1 reference picture
Wherein n is more than or equal to 3, and
Wherein the n reference picture is the reference picture of n-1 reference picture.
2. the method for claim 1, determine that wherein the step of the corresponding region of a plurality of reference pictures comprises:
Determine the corresponding region of first reference picture corresponding with current block by the motion of prediction current block;
Corresponding region along the moving mass border of first reference picture with first reference picture is divided into sub-corresponding region; And
Determine corresponding region by second reference picture of the motion vector indication of the moving mass of first reference picture of the sub-corresponding region that comprises first reference picture.
3. method as claimed in claim 2, determine that wherein the corresponding region of second reference picture comprises:
If one of sub-corresponding region of first reference picture is included in the intra-frame prediction block, then use the motion vector of the adjacent moving mass of intra-frame prediction block to determine the virtual motion vector of this intra-frame prediction block; And
Determine corresponding region by second reference picture of this virtual motion vector indication.
4. method as claimed in claim 3, wherein the median of the motion vector of the adjacent moving mass of intra-frame prediction block or mean value are used as the virtual motion vector of intra-frame prediction block.
5. the method for claim 1, wherein the corresponding region of n reference picture only is included in one or more intra-frame prediction block, and
Wherein, if only some is included in the intra-frame prediction block in the corresponding region of n reference picture, then the zone of this part is greater than threshold value.
6. the method for claim 1, the step that wherein produces the prediction piece of current block comprises:
Determine the weight of the corresponding region of a plurality of reference pictures; And
Produce the prediction piece of current block by each results added that the corresponding region of a plurality of reference pictures be multiply by weight respectively and will multiply each other.
7. method as claimed in claim 6, wherein use current block adjacent block first pre-treatment pixel and with the neighbor of the corresponding region of the pixel corresponding reference picture of the first pre-treatment of the adjacent block of current block weight is defined as such value, the difference between the value of the predicted value of the neighbor of the current block that the weighted sum of the neighbor of the corresponding region that it will be by calculating reference picture obtains and the neighbor of current block minimizes.
8. the method for claim 1 comprises that also the sign that uses a plurality of reference pictures to carry out the piece predictive coding indication is inserted in the presumptive area of the bit stream that produces by coded image.
9. the method for claim 1, the step of wherein determining the corresponding region of a plurality of reference pictures comprises, if the part of corresponding region that is included in first reference picture in the intra-frame prediction block is greater than threshold value, then only the corresponding region of first reference picture is defined as and is used to predict the corresponding region of the reference picture of current block
The step that wherein produces the prediction piece of current block comprises the prediction piece that will determine by the value that the corresponding region of first reference picture be multiply by predetermined weight acquisition as current block.
10. picture coding device comprises:
The reference picture determining unit, it is by following the tracks of the corresponding region of being determined to be used to predict a plurality of reference pictures of current block by the motion vector route of the corresponding region of the reference picture of current block reference;
The weight estimation unit, the weighted sum of its corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; With
Coding unit, its coding current block with predict poor between the piece,
Wherein the reference picture determining unit is determined the corresponding region by second reference picture of the motion vector indication of current block, up to the corresponding region of determining by the n reference picture of the motion vector indication of n-1 reference picture,
Wherein n is more than or equal to 3, and
Wherein the n reference picture is the reference picture of n-1 reference picture.
11. device as claimed in claim 10, wherein the reference picture determining unit will be divided into sub-corresponding region by the corresponding region of first reference picture of the motion vector indication of current block along the moving mass border of first reference picture; And determine corresponding region by second reference picture of the motion vector indication of the moving mass of first reference picture of the sub-corresponding region that comprises first reference picture.
12. device as claimed in claim 11, if wherein one of sub-corresponding region of first reference picture is included in the intra-frame prediction block, then the reference picture determining unit uses the motion vector of the adjacent moving mass of intra-frame prediction block to determine the virtual motion vector of this intra-frame prediction block, and determines the corresponding region by second reference picture of this virtual motion vector indication.
13. device as claimed in claim 12, wherein the median of the motion vector of the adjacent moving mass of intra-frame prediction block or mean value are used as the virtual motion vector of intra-frame prediction block.
14. device as claimed in claim 10, wherein the corresponding region of n reference picture only is included in one or more intra-frame prediction block, and
Wherein, if only some is included in the intra-frame prediction block in the corresponding region of n reference picture, then the zone of this part is greater than threshold value.
15. device as claimed in claim 10, wherein the weight estimation unit comprises:
Weight calculation unit is determined the weight of the corresponding region of a plurality of reference pictures; With
Prediction piece generation unit, it produces the prediction piece of current block by each results added that the corresponding region of a plurality of reference pictures be multiply by weight respectively and will multiply each other.
16. device as claimed in claim 15, wherein weight calculation unit use current block adjacent block first pre-treatment pixel and with the neighbor of the corresponding region of the pixel corresponding reference picture of the first pre-treatment of the adjacent block of current block weight is defined as such value, the difference between the value of the predicted value of the neighbor of the current block that the weighted sum of the neighbor of the corresponding region that it will be by calculating reference picture obtains and the neighbor of current block minimizes.
17. device as claimed in claim 10, wherein coding unit will indicate the sign that uses a plurality of reference pictures to carry out the piece predictive coding to be inserted in the presumptive area of the bit stream that produces by coded image.
18. device as claimed in claim 10, if the part of corresponding region that wherein is included in first reference picture in the intra-frame prediction block is greater than threshold value, then only the corresponding region of this first reference picture is defined as and is used to predict the corresponding region of the reference picture of current block
Wherein the weight estimation unit will be by multiply by the corresponding region of first reference picture definite prediction piece as current block of value that predetermined weight obtains.
19. a picture decoding method comprises:
Discern the predictive mode of current block by reading the prediction mode information that is included in the incoming bit stream;
If current block is confirmed as using the corresponding region of a plurality of reference pictures to predict, then by following the tracks of the corresponding region of determining to be used to predict a plurality of reference pictures of current block by the motion vector route of the corresponding region of the corresponding region of the reference picture of the motion vector route reference that is included in the current block in the bit stream and reference picture;
The weighted sum of the corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; And
By will being included in current block and the difference of prediction between the piece and the prediction piece Calais's decoding current block mutually in the bit stream,
Wherein the motion vector route of the corresponding region of track reference picture comprises: determine corresponding region by second reference picture of the motion vector indication of current block, and up to the corresponding region of determining by the n reference picture of the motion vector indication of n-1 reference picture,
Wherein n is more than or equal to 3, and
Wherein the n reference picture is the reference picture of n-1 reference picture.
20. method as claimed in claim 19 determines that wherein the step of the corresponding region of a plurality of reference pictures comprises:
By along the moving mass boundary demarcation of first reference picture corresponding region by first reference picture of the motion vector indication of current block; And
Determine corresponding region by second reference picture of the motion vector indication of the moving mass of first reference picture of the sub-corresponding region that comprises first reference picture.
21. method as claimed in claim 19, the step that wherein produces the prediction piece of current block comprises:
Use current block adjacent block first pre-treatment pixel and determine value with the neighbor of the corresponding region of the pixel corresponding reference picture of the first pre-treatment of the adjacent block of current block as the weight of the corresponding region of a plurality of reference pictures, the difference between the value of the predicted value of the neighbor of the current block that the weighted sum of the neighbor of the corresponding region that this weighted value will be by the calculating reference picture obtains and the neighbor of current block minimizes; And
Produce the prediction piece of current block by each results added that the corresponding region of a plurality of reference pictures be multiply by weight respectively and will multiply each other.
22. a picture decoding apparatus comprises:
The predictive mode recognition unit, it discerns the predictive mode of current block by reading the prediction mode information that is included in incoming bit stream;
The reference picture determining unit, if current block is confirmed as using the corresponding region of a plurality of reference pictures to predict, then it is by following the tracks of the corresponding region of being determined to be used to predict a plurality of reference pictures of current block by the motion vector route of the corresponding region of the corresponding region of the reference picture of the motion vector route reference of the current block that is included in bit stream and reference picture;
The Weight prediction unit, the weighted sum of its corresponding region by calculating a plurality of reference pictures produces the prediction piece of current block; With
Decoding unit, it be by will being included in current block and the difference of prediction between the piece and the prediction piece Calais's decoding current block mutually in the bit stream,
Wherein the reference picture determining unit is determined the corresponding region by second reference picture of the motion vector indication of current block, up to the corresponding region of determining by the n reference picture of the motion vector indication of n-1 reference picture,
Wherein n is more than or equal to 3, and
Wherein the n reference picture is the reference picture of n-1 reference picture.
Applications Claiming Priority (6)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US85629006P | 2006-11-03 | 2006-11-03 | |
| US60/856,290 | 2006-11-03 | ||
| KR1020070000706A KR101356734B1 (en) | 2007-01-03 | 2007-01-03 | Method and apparatus for video encoding, and method and apparatus for video decoding using motion vector tracking |
| KR1020070000706 | 2007-01-03 | ||
| KR10-2007-0000706 | 2007-01-03 | ||
| PCT/KR2007/005531 WO2008054179A1 (en) | 2006-11-03 | 2007-11-02 | Method and apparatus for encoding/decoding image using motion vector tracking |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN101573982A CN101573982A (en) | 2009-11-04 |
| CN101573982B true CN101573982B (en) | 2011-08-03 |
Family
ID=42193881
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN2007800491266A Expired - Fee Related CN101573982B (en) | 2006-11-03 | 2007-11-02 | Method and apparatus for encoding/decoding image using motion vector tracking |
Country Status (2)
| Country | Link |
|---|---|
| JP (1) | JP5271271B2 (en) |
| CN (1) | CN101573982B (en) |
Families Citing this family (14)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR20090095012A (en) * | 2008-03-04 | 2009-09-09 | 삼성전자주식회사 | Method and apparatus for encoding and decoding image using consecutive motion estimation |
| KR101691199B1 (en) * | 2008-04-11 | 2016-12-30 | 톰슨 라이센싱 | Method and apparatus for template matching prediction(tmp) in video encoding and decoding |
| KR101522850B1 (en) * | 2010-01-14 | 2015-05-26 | 삼성전자주식회사 | Method and apparatus for encoding/decoding motion vector |
| JP5979405B2 (en) * | 2011-03-11 | 2016-08-24 | ソニー株式会社 | Image processing apparatus and method |
| CN110830797B (en) * | 2012-01-18 | 2023-09-15 | 韩国电子通信研究院 | Video decoding device, video encoding device and method for transmitting bit stream |
| CN117041547A (en) | 2016-04-29 | 2023-11-10 | 世宗大学校产学协力团 | Method and apparatus for encoding/decoding image signal |
| CN109479142B (en) * | 2016-04-29 | 2023-10-13 | 世宗大学校产学协力团 | Method and apparatus for encoding/decoding image signal |
| ES2737845B2 (en) * | 2016-07-05 | 2021-05-19 | Kt Corp | METHOD AND APPARATUS TO PROCESS VIDEO SIGNAL |
| CN117041561A (en) * | 2016-12-07 | 2023-11-10 | 株式会社Kt | Method for decoding or encoding video and apparatus for storing video data |
| WO2018132150A1 (en) * | 2017-01-13 | 2018-07-19 | Google Llc | Compound prediction for video coding |
| CN111050177B (en) * | 2018-10-15 | 2021-07-20 | 腾讯科技（深圳）有限公司 | Video encoding method, video decoding method, video encoding apparatus, video decoding apparatus, computer device, and storage medium |
| CN114885158B (en) * | 2018-12-21 | 2023-10-20 | 华为技术有限公司 | Method and apparatus for mode dependent and size dependent block level restriction of position dependent prediction combinations |
| CN112632426B (en) * | 2020-12-22 | 2022-08-30 | 新华三大数据技术有限公司 | Webpage processing method and device |
| CN117041597B (en) * | 2023-10-09 | 2024-01-19 | 中信建投证券股份有限公司 | Video encoding and decoding methods and devices, electronic equipment and storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5602593A (en) * | 1994-02-22 | 1997-02-11 | Nec Corporation | Overlapped motion compensation using a window function which varies in response to an input picture |
| CN1592131A (en) * | 1999-04-28 | 2005-03-09 | D.S.P.C.科技有限公司 | System and method for joint time tracking of multiple paths |
| EP1519589A2 (en) * | 1998-09-10 | 2005-03-30 | Microsoft Corporation | Object tracking in vector images |
| US6901110B1 (en) * | 2000-03-10 | 2005-05-31 | Obvious Technology | Systems and methods for tracking objects in video sequences |
Family Cites Families (11)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| JPH07264596A (en) * | 1994-03-15 | 1995-10-13 | Daewoo Electron Co Ltd | Image encoding device |
| JPH09163376A (en) * | 1995-12-05 | 1997-06-20 | Nippon Telegr & Teleph Corp <Ntt> | Inter-frame coding, decoding method and device for moving image |
| JPH09161071A (en) * | 1995-12-12 | 1997-06-20 | Sony Corp | Device and method for making areas correspond to each other |
| JP2002223440A (en) * | 2001-01-25 | 2002-08-09 | Nippon Hoso Kyokai <Nhk> | Shape coder |
| JP2003284075A (en) * | 2002-01-18 | 2003-10-03 | Toshiba Corp | Method and apparatus for coding moving image, and method and apparatus for decoding |
| JP4130783B2 (en) * | 2002-04-23 | 2008-08-06 | 松下電器産業株式会社 | Motion vector encoding method and motion vector decoding method |
| KR100533028B1 (en) * | 2003-06-03 | 2005-12-02 | 엘지전자 주식회사 | Method for deciding types of motion vector and macroblock |
| JP2004227519A (en) * | 2003-01-27 | 2004-08-12 | Matsushita Electric Ind Co Ltd | Image processing method |
| KR101044934B1 (en) * | 2003-12-18 | 2011-06-28 | 삼성전자주식회사 | Motion vector estimation method and encoding mode determining method |
| JP2006345445A (en) * | 2005-06-10 | 2006-12-21 | Sony Corp | Moving picture conversion apparatus and moving picture conversion method, and computer program |
| JP2007043651A (en) * | 2005-07-05 | 2007-02-15 | Ntt Docomo Inc | Dynamic image encoding device, dynamic image encoding method, dynamic image encoding program, dynamic image decoding device, dynamic image decoding method, and dynamic image decoding program |
-
2007
- 2007-11-02 CN CN2007800491266A patent/CN101573982B/en not_active Expired - Fee Related
- 2007-11-02 JP JP2009535217A patent/JP5271271B2/en not_active Expired - Fee Related
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US5602593A (en) * | 1994-02-22 | 1997-02-11 | Nec Corporation | Overlapped motion compensation using a window function which varies in response to an input picture |
| EP1519589A2 (en) * | 1998-09-10 | 2005-03-30 | Microsoft Corporation | Object tracking in vector images |
| CN1592131A (en) * | 1999-04-28 | 2005-03-09 | D.S.P.C.科技有限公司 | System and method for joint time tracking of multiple paths |
| US6901110B1 (en) * | 2000-03-10 | 2005-05-31 | Obvious Technology | Systems and methods for tracking objects in video sequences |
Also Published As
| Publication number | Publication date |
|---|---|
| JP2010509800A (en) | 2010-03-25 |
| JP5271271B2 (en) | 2013-08-21 |
| CN101573982A (en) | 2009-11-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN101573982B (en) | Method and apparatus for encoding/decoding image using motion vector tracking | |
| KR101356734B1 (en) | Method and apparatus for video encoding, and method and apparatus for video decoding using motion vector tracking | |
| KR101356735B1 (en) | Mothod of estimating motion vector using global motion vector, apparatus, encoder, decoder and decoding method | |
| CN103096080B (en) | Apparatus for estimating motion vector of current block | |
| KR101218434B1 (en) | Method and systems to estimate motion based on reconstructed reference frames at a video decoder | |
| CN100566426C (en) | The method and apparatus of encoding and decoding of video | |
| CN102362498B (en) | Selectivity is carried out the apparatus and method of syntax element encodes/decoding and is used it to carry out the apparatus and method of coding/decoding to image | |
| CN101513071A (en) | Method and apparatus for determining expected distortion in decoded video blocks | |
| CN101595735A (en) | Video is carried out the method and apparatus of Code And Decode | |
| CN101009833A (en) | Method of and apparatus for deciding encoding mode for variable block size motion estimation | |
| EP2555523A1 (en) | Image encoding device, image decoding device, image encoding method, and image decoding method | |
| US20110150074A1 (en) | Two-pass encoder | |
| KR20080088042A (en) | Method and apparatus for video encoding and decoding | |
| CN107846593B (en) | Rate distortion optimization method and device | |
| CN102484711A (en) | Prediction of a movement vector of a current image partition pointing to a reference zone that covers multiple reference image partitions and encoding and decoding using one such prediction | |
| CN104954792A (en) | Method and device for carrying out subjective video quality optimization coding on P-frame sequence | |
| CN102265615A (en) | Image prediction using the repartitioning of a reference causal area, and encoding and decoding using such a prediction | |
| CN100521789C (en) | Method of estimating backward motion vectors within a video sequence | |
| CN103314584A (en) | Method of coding a sequence of images and corresponding reconstruction method | |
| Paul et al. | Adaptive search range selection for scalable video coding extension of H. 264/AVC | |
| KR20100098225A (en) | Method and apparatus for computational complexity control of video encoding |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C14 | Grant of patent or utility model | ||
| GR01 | Patent grant | ||
| CF01 | Termination of patent right due to non-payment of annual fee |
Granted publication date: 20110803 Termination date: 20161102 |
|
| CF01 | Termination of patent right due to non-payment of annual fee |