CN101669366A - Scalable video coding encoder with adaptive reference fgs and fgs motion refinement mechanism and method thereof - Google Patents
Scalable video coding encoder with adaptive reference fgs and fgs motion refinement mechanism and method thereof Download PDFInfo
- Publication number
- CN101669366A CN101669366A CN200780045636A CN200780045636A CN101669366A CN 101669366 A CN101669366 A CN 101669366A CN 200780045636 A CN200780045636 A CN 200780045636A CN 200780045636 A CN200780045636 A CN 200780045636A CN 101669366 A CN101669366 A CN 101669366A
- Authority
- CN
- China
- Prior art keywords
- fgs
- layer
- current
- contraction
- expansion factor
- 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.)
- Pending
Links
Images
Landscapes
- Compression, Expansion, Code Conversion, And Decoders (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
Abstract
Provided are alternatives for improving coding efficiency when an AR-FGS technique and an FGS motion refinement technique are applied to scalable video coding. When prediction of a residual signal anFGS layer is not performed, a ' prediction signal of a block related to the FGS layer is predicted in the same manner as the manner of predicting a prediction signal of a base quality layer. A scalingfactor is allowed to have a non-zero value if required, and the residual signal of the FGS layer is used to determine a scaling factor of a higher FGS layer. The AR-FGS and FGS motion refinement techniques are restricted from being simultaneously used for key pictures.
Description
Technical field
The present invention relates to adopt the scalable video (SVC) of fine granulation SNR scalability (FGS) motion refinement (motionrefinement) technology and adaptive reference (AR, adaptive reference) FGS technology.
Background technology
In scalable video (SVC), FGS is the key character that is used for controlling subtly the video quality of SNR dimension.When having removed the FGS layer,, the inter prediction structure of SVC vision signal can be transmitted to subsequently image owing to causing image (picture) quality deterioration.
Can control deterioration in image quality by adaptive reference (AR) the FGS technology that is used to improve code efficiency propagates.In addition, the FGS motion refinement technique that is used for being provided with the motion vector of each FGS layer can be used to improve the code efficiency of FGS layer.Yet, when using AR FGS and FGS motion refinement technique together, because according to the FGS motion refinement technique, not from predict the residual signals of FGS layer piece corresponding to basic layer (that is, base-quality layer or lower FGS layer) piece of FGS layer piece, so AR FGS technology is not suitably worked.
IT R﹠amp by MIC (information and Department of Communication Force)/IITA (information technology advances research institute); This work is supported in D planning [2005-S-103-02, " Development of Ubiquitous Content Access Technology forConvergence of Broadcasting and Communications (being used to converge the development of broadcasting and the ubiquitous access to content technology of communicating by letter) "].
Description of drawings
Fig. 1 illustrates the AR-FGS structure of traditional SVC;
Fig. 2 illustrates the AR-FGS structure among the SVC according to the preferred embodiment of the invention;
Fig. 3 A to Fig. 3 E illustrates at according to the decoding processing in the standardisation documents of first alternative of describing with reference to figure 2 of the embodiment of the invention;
Fig. 4 A and Fig. 4 B illustrate the decoding processing that is used for according to the standardisation documents of second alternative of describing with reference to figure 2 of the embodiment of the invention;
Fig. 5 illustrates and is used for the grammer of the 3rd alternative according to the preferred embodiment of the invention;
Fig. 6 illustrates the decoding processing that is used for according to the standardisation documents of the 3rd alternative of the embodiment of the invention;
Fig. 7 illustrates the grammer that is used for according to the 4th alternative of the embodiment of the invention;
Fig. 8 illustrates the grammer that is used for according to the 5th alternative of the embodiment of the invention;
Fig. 9 illustrates the grammer that is used for according to the 6th alternative of the embodiment of the invention;
Figure 10 is the block diagram according to the SVC encoder of AR-FGS after the employing improvement of the embodiment of the invention and FGS motion refinement technique;
Figure 11 is the flow chart that illustrates according to the operation of the illustrated SVC encoder of Figure 10 of the embodiment of the invention;
Figure 12 is the block diagram according to the SVC encoder of AR-FGS after the employing of the embodiment of the invention improvement corresponding with the 3rd alternative and FGS motion refinement technique; And
Figure 13 is the block diagram that adopts the SVC encoder of AR-FGS after the improvement corresponding with the 4th, the 5th and the 6th alternative and FGS motion refinement technique according to another embodiment of the present invention.
Summary of the invention
Technical problem
FGS motion refinement technique among the SVC can be used to improve the code efficiency of FGS layer.The FGS motion refinement technique allow the FGS layer have movable information with base-quality layer in different block modes.
In the case, can not predict the residual signals of this FGS layer piece from the coordination piece (co-located block) the basic layer of FGS layer piece, and the residual signals of base-quality layer is not suitable for controlling the adaptivity of AR-FGS.In addition, current AR-FGS has only considered the attribute of the residual signals of base-quality layer, and thereby may have problems when using AR-FGS technology and FGS motion refinement technique simultaneously.
Correspondingly, the invention provides the alternative that is used to solve contingent problem when using AR-FGS and FSG motion refinement technique simultaneously, thereby improved the adaptivity of AR-FGS.
Technical scheme
According to an aspect of the present invention, provide when AR-FGS and FGS motion refinement technique are applied to SVC simultaneously, can improve the alternative of code efficiency.
The residual signals of the piece in not predicting the FGS layer is predicted the prediction signal of the piece of FGS layer according to the mode identical with the prediction signal that is used for predicting base-quality layer.
If desired, then contraction-expansion factor can have nonzero value, and is used to determine the contraction-expansion factor of higher FGS layer for its residual signals of not carrying out the FGS piece of residual signals prediction.
When residual signals when prediction between active coating always, determine self-adaptive processing based on the residual signals of base-quality layer.
FGS and FGS motion refinement technique are not used for key images simultaneously.
According to an aspect of the present invention, a kind of AR-FGS after the improvement and SVC encoder of FGS motion refinement technique of using is provided, comprise: the prediction signal determining unit, be used for when between base-quality layer or low FGS layer and current FGS layer, not during inter-layer prediction, determining the prediction signal of current FGS layer piece according to the contraction-expansion factor of current FGS layer; And the contraction-expansion factor determining unit, be used for residual signals based on current FGS layer piece, be identified for predicting contraction-expansion factor corresponding to the higher FGS layer piece of current FGS layer piece.
According to a further aspect in the invention, a kind of AR-FGS after the improvement and SVC encoder of FGS motion refinement technique of using is provided, comprise: inter-layer prediction is provided with the unit, is used to be arranged on inevitable inter-layer prediction between basic layer (that is, base-quality layer or low FGS layer) and each FGS layer; And the contraction-expansion factor determining unit, be used for determining the contraction-expansion factor of higher FGS layer based on the residual signals of basic layer.
According to a further aspect in the invention, a kind of AR-FGS after the improvement and SVC encoder of FGS motion refinement technique of using is provided, comprise: the FGS-MR deactivated cell, when being used for image as the GOP of incoming bit stream, prevent that the FGS motion refinement technique is applied to this key images corresponding to key images.
According to a further aspect in the invention, a kind of AR-FGS after the improvement and SVC encoder of FGS motion refinement technique of using is provided, comprise: selectivity FGS-MR deactivated cell, only be used under the situation of image of the GOP of incoming bit stream therein, when the AR-FGS technology is applied to this key images, prevent that just the FGS motion refinement technique is applied to this key images corresponding to key images.
According to a further aspect in the invention, provide a kind of AR-FGS after the improvement and SVC encoder of FGS motion refinement technique of using, comprising: the FGS-MR deactivated cell is used to prevent that the FGS motion refinement technique is applied to key images; And the AR-FGS deactivated cell, when being used for image as the GOP of incoming bit stream, stop the AR-FGS technology to be applied to this key images corresponding to key images.
According to a further aspect of the invention, a kind of AR-FGS after the improvement and SVC decoder of FGS motion refinement technique of using is provided, wherein, in the operation that is used for each FGS layer is decoded, when the FGS motion refinement technique is applied to current FGS layer and between current FGS layer and base-quality layer or low FGS layer not during inter-layer prediction, contraction-expansion factor according to current FGS layer comes the prediction signal of current FGS layer piece is decoded, and the SVC decoder is determined this contraction-expansion factor based on the residual signals of current FGS layer piece.
According to a further aspect in the invention, a kind of AR-FGS after the improvement and SVC decoder of FGS motion refinement technique of using is provided, wherein, so that between basic layer and each FGS layer during inevitable inter-layer prediction, the inevitable residual signals based on basic layer of this SVC decoder be determined the contraction-expansion factor of higher FGS layer when the bit stream that received of configuration.
According to a further aspect in the invention, a kind of AR-FGS after the improvement and SVC decoder of FGS motion refinement technique of using is provided, wherein, this SVC decoder is not checked the sign of having represented the FGS motion refinement technique to be applied to the key images of the GOP in the bit stream that is received.
According to a further aspect in the invention, a kind of SVC decoder that uses the AR-FGS after the improvement is provided, wherein, when reception comprises the bit stream of inter-layer prediction signalization, this SVC decoder is identified for predicting that based on the residual signals of current FGS layer piece contraction-expansion factor corresponding to the higher FGS layer piece of current FGS layer piece, this inter-layer prediction signalization have been represented inter-layer prediction is set will carrying out between basic layer and each FGS layer.
According to a further aspect in the invention, provide a kind of AR-FGS after the improvement and SVC coding/decoding method of FGS motion refinement technique of using, this SVC coding/decoding method comprises: determine that whether present frame is corresponding to key images; And during corresponding to key images, determine whether to use the AR-FGS technology, and when present frame does not correspond to this key images, determine whether to use the FGS motion refinement technique at present frame.
Beneficial effect
When AR-FGS and FGS motion refinement were applied to SVC simultaneously, the present invention had improved code efficiency.In addition, the present invention can solve the problem that produces that causes because the self adaptation of current AR-FGS has only considered the attribute of residual signals of basic layer when using AR-FGS and FGS motion refinement simultaneously.
Embodiment
Referring now to the accompanying drawing that wherein shows example embodiment of the present invention the present invention is described more fully.Yet the present invention may be embodied as many different forms, and should not be understood that to be limited to the embodiment that proposes here; On the contrary, propose these embodiment and make that the disclosure will be thorough and complete, and the disclosure will be passed on design of the present invention all sidedly to those skilled in the art.Run through in the accompanying drawing, same Reference numeral refers to same element.
SVC is the important technology that is used for the video communication of isomerous environment.The SVC technology allows block (truncation) of original video bitstream under the constraint of terminal or network, the output bit flow that presents with the difference that provides corresponding to original contents.In following three dimensions, support the scalability of SVC video, that is, and space, time and SNR.
In SVC, FGS is the control of video quality subtly.For each spatial resolution, at first come base-quality layer is encoded by the method that is similar to H.264/AVC.Then, can add nearly three FGS layers to base-quality layer, so that strengthen the SNR quality of corresponding base-quality layer.Can extract these FGS layers from the arbitrfary point, so that satisfy bit rate conditions.
Because the influence of FGS layer that is removed and inter prediction structure, so video quality (SNR) deterioration can be transmitted to subsequent images.This propagation is called as the drift error among the SVC.For fear of drift error, can only use the information of the base-quality layer of previous frame to obtain the inter prediction of key images.Yet owing to do not use best inter prediction, this solution causes low code efficiency.
For the flexible balance between code efficiency and the error robustness is provided, AR-FGS has been proposed.The AR-FGS technology is controlled the part of the FGS information that is used to form inter prediction adaptively based on the characteristic of base-quality layer.
In addition, the FGS motion refinement technique also increases the code efficiency of FGS layer.The FGS motion refinement technique allows each FGS layer to have motion vector to make block mode be different from base-quality layer.
Yet, because can not predict the attribute of the residual signals of this FGS layer piece from the coordination piece the basic layer of FGS layer piece, so may there be the problem of the attribute of the residual signals of only having considered base-quality layer for the adaptivity of control AR-FGS.
Be described with reference to Figure 1 the problem that when in traditional SVC, using FGS motion refinement technique and AR-FGS technology simultaneously, produces.
Fig. 1 illustrates the AR-FGS structure of traditional SVC.With reference to figure 1, spatial resolution is made up of a base-quality layer 100 and first to the 3rd additional FGS layer 110,120 and 130.The processing of the one FGS layer 110 is described as example.
The reconstruction signal of piece 101 comprises: prediction signal 102 and residual signals 103.Prediction signal 102 corresponding to from the reconstruction signal 104 of the previous image block of base-quality layer and the prediction signal of motion compensation, and from the difference between the reconstruction signal 104 of the reconstruction signal 105 of a FGS layer 110 and previous image and the prediction signal sum of motion compensation.
In self-adapting stretching unit 106, the first contraction-expansion factor S1 multiply by from the difference between the reconstruction signal 104 of the reconstruction signal 105 of a FGS layer 110 and previous image and the prediction signal of motion compensation.When the first contraction-expansion factor S1 is 0, only obtain prediction signal 102 from base-quality layer, and even when from the piece 105 extraction FGS information of a FGS layer 110 the video quality deterioration do not occur yet.Yet when the first contraction-expansion factor S1 is not zero, if not from the piece 105 extraction FGS information of a FGS layer 110, prediction signal 102 will have better video quality.The following describes two kinds of situations wherein controlling the first contraction-expansion factor S1.
◆ when Closing Switch K11
At first, the inter-layer prediction of consideration from base-quality layer (or low FGS layer) to higher FGS layer.When K switch 11 (111) be connected to a FGS layer 110 and thereby prediction from basic layer 100 when higher FGS layer takes place, determine the first contraction-expansion factor S1 based on the coefficient of the residual signals 107 of base-quality layer 100.When the coefficient of residual signals 107 is not 0 (when K switch 21 (121) is switched to node 1), be set to 0 coefficient of correspondence that obtains prediction signal 102 by the first contraction-expansion factor S1.When the coefficient of residual signals 107 is 0, be set to the coefficient of correspondence that nonzero value is determined prediction signal 102 by the first contraction-expansion factor S1.The nonzero value of the first contraction-expansion factor S1 depends on content and application program.
When all coefficients of residual signals 107 all are 0, in spatial domain, take place flexible.When arbitrary coefficient of residual signals 107 is non-zero, in transform domain, carry out flexible.That is, change differential signal to transform domain, and then described differential signal is stretched from spatial domain.
◆ when cut-off switch K11
Consider not exist second situation from base-quality layer (or low FGS layer) to the inter-layer prediction of higher FGS layer.For example, when K switch 11 (111) be not connected to a FGS layer 110 and thereby when not having to carry out prediction from base-quality layer 100 to higher FGS layer, utilize except K switch 21 (121) be switched to node 2 and based on the coefficient of the residual signals 103 of a FGS layer 110 be provided with the first contraction-expansion factor S1, with first situation in identical method determine the first contraction-expansion factor S1.Correspondingly, produced the problem that the residual signals 103 that uses current FGS layer is determined contraction-expansion factor S1.
Now, that propose, be used to solve the problem that produces when using AR-FGS technology and FGS motion refinement technique together in the present invention alternative will be described with reference to Figure 2.
Fig. 2 illustrates the AR-FGS structure among the SVC according to the preferred embodiment of the invention.With reference to figure 2, come control connection K switch 11, K12 and K13 according to the FGS motion refinement technique.With reference to as described in the figure 1, when producing inter-layer prediction between residual block 103 and 107, using in the time of AR-FGS technology and FGS motion refinement technique does not become problem as top.Yet, when not having inter-layer prediction, when definite contraction-expansion factor Si, have problems.The alternative that is used to address this problem is described below with reference to Fig. 2.
[first alternative]
When disconnect to connect K switch 1i and thereby between execution level during the residual signals prediction, the contraction-expansion factor Si of i FGS layer is set to 0, makes that the prediction signal of the related blocks in the i FGS layer is corresponding with the prediction signal of base-quality layer 200.For example, the prediction signal of the related blocks 202 in the FGS layer 210 becomes with the prediction signal of base-quality layer identical.
As cut-off switch K1i and Closing Switch K1
I+1The time, determine the contraction-expansion factor S (i+1) of (i+1) FGS layer based on the residual signals of i FGS layer.In addition, as cut-off switch K1i, Closing Switch K1
I+1And Closing Switch K1
I+2The time, determine the contraction-expansion factor S (i+1) of (i+1) FGS layer and the contraction-expansion factor S (i+2) of (i+2) FGS layer based on the i residual signals.For example, when i is 1, determine contraction-expansion factor S (i+1) and S (i+2) based on residual signals 203.
◆ when cut-off switch K11
1) when stopping using the prediction of (inactivate) residual signals between the base-quality layer 200 and first fsg layer 210, contraction-expansion factor S1 is set to 0.Then, obtain the residual signals 203 of a FGS layer 210 according to the mode identical with the residual error layer 207 that obtains base-quality layer 200.
For example, the prediction signal 202 of a FGS layer 210 is identical with the prediction interval of base-quality layer 200, and irrespectively the residual signals 203 of a FGS layer 210 is encoded with the residual signals 207 of base-quality layer 200.In the case, not by using prediction but by using the quantization parameter different to come residual signals 203 is encoded with the quantification that is used for residual signals 207 is encoded.Residual signals 203 can be used for determining the residual signals and the contraction-expansion factor S2 of the 2nd FGS layer 220.
2) when cut-off switch K11 (211) and Closing Switch K12 (212), K switch 22 (222) is switched to node 2, and the residual signals 203 of a FGS layer 210 is used to determine the contraction-expansion factor S2 of the 2nd FGS layer 220.
3) when Closing Switch K12 (212) and K13 (213), K switch 23 (233) is switched to node 1, and determines contraction-expansion factor S2 and S3 based on residual signals 203.
Fig. 3 A to 3E illustrates the standardisation documents at first alternative that proposes in the present invention.In Fig. 3 A to 3E, with the each several part of shadow representation according to the standardization document of first alternative modification that proposes in the present invention.In Fig. 3 A to 3E, extention be when the FGS layer is decoded in determining the decoding processing of contraction-expansion factor, when using FGS motion refinement (motion_refinement_flag (motion refinement sign)=1) and not using residual signals prediction (residual_prediction_flag (residual prediction flag)=0), be used for contraction-expansion factor sF and be set to 0 grammer.
[second alternative]
In first alternative, the contraction-expansion factor Si of i FGS layer is set to 0, make when disconnect when connecting K switch 1i and not carrying out inter-layer residue prediction, that the prediction signal of the related blocks of i FGS layer becomes is identical with basic layer prediction signal.Whether contraction-expansion factor Si is set to aspect 0, and second alternative is different from first alternative.
Promptly, in second alternative, even cause not existing inter-layer residue prediction because disconnecting connection K switch 1i, if desired, then also contraction-expansion factor Si is set to nonzero value, and for it there is not the residual signals of the FGS piece of inter-layer prediction to be used to determine the contraction-expansion factor of higher FGS layer.
For example, when cut-off switch K11 (211) and Closing Switch K12 (212), K switch 22 (222) is switched to node 2, and the residual signals 203 of a FGS layer 210 is used to determine the contraction-expansion factor S2 of the 2nd FGS layer 220.
Can also use second alternative solve from the contraction-expansion factor S1 of a FGS layer 110 prediction piece 102, above with reference to figure 1 described problem.
Fig. 4 A and Fig. 4 B illustrate the standardisation documents according to the decoding processing of second alternative of describing with reference to figure 2 of the embodiment of the invention.In Fig. 4 A and Fig. 4 B, with the each several part of shadow representation according to the standardization document of second alternative modification that proposes in the present invention.
In Fig. 4 A and Fig. 4 B, additional decoding processing allows to determine the contraction-expansion factor Si of i FGS layer when cut-off switch K1i, by the residual signals 107 of (i-1) FGS layer, this is different from current standardisation documents, in current standardisation documents, when cut-off switch K1i, by the residual signals of i FGS layer, determine the contraction-expansion factor Si's (for example, when the cut-off switch K11, determine contraction-expansion factor S1) of the i FGS factor by residual signals 103.
Variable sigBCoeff has represented the value corresponding to residual signals, and is used for determining contraction-expansion factor.In current standardisation documents, when motion_refinement_flag is 1 and residual_prediction_flag when being 0, the sigBcoeff of i FGS layer determines the contraction-expansion factor of i FGS layer.That is, residual signals 103 is determined contraction-expansion factor S1.Yet, in the present invention, generate variable sigBCoeffTem and revise standardisation documents, make sigBCoeff have the residual signals value of (i-1) FGS layer, so that solve the problem of current standardisation documents.
[the 3rd alternative]
In the time between basic layer and each FGS layer, must carrying out the FGS motion refinement, the 3rd alternative is proposed for AR-FGS.In the 3rd alternative, always K switch 11 (211), K12 (212) and K13 (213) are set to closure.That is, in the 3rd alternative, always activate inter-layer prediction, make that residual signals is predicted between execution level.Correspondingly, all K switch 21 (221), K22 (222) and K23 (223) are switched to node 1, and thereby the residual signals 207 of base-quality layer is used for determine contraction-expansion factor Si always.
Fig. 5 illustrates the grammer of the 3rd alternative according to the preferred embodiment of the invention.Deletion among Fig. 5 is to have represented the grammer of whether carrying out residual signals prediction (inter-layer prediction), make and when residual_prediction_flag is 1, carry out the residual signals prediction, and when residual_prediction_flag is 0, do not carry out the residual signals prediction.Yet, as mentioned above,, need not to transmit this sign, thereby delete this grammer to decoder if residual prediction flag is set to carry out the residual signals prediction.
Fig. 6 illustrates the standardisation documents that is used for according to the decoding processing of the 3rd alternative of the embodiment of the invention.In Fig. 6, because residual_prediction_flag always 1 in the FGS layer, so deletion and the relevant description of decoding processing when residual_prediction_flag is 0, and delete and be used to check that whether residual_prediction_flag is 1 processing.
In Fig. 5 and 6, with the part of standards document of shadow representation according to the modification of the 3rd alternative.
[the 4th alternative]
Only the key images in the image sets in SVC (GOP) is used the AR-FGS technology.Thereby, do not use the FGS motion refinement, to solve the problem that when using AR-FGS and FGS motion refinement simultaneously, produces to key images.Correspondingly, need not to revise existing AR-FGS technology so that receive the FGS motion refinement technique.
Fig. 7 illustrates the grammer that is used for according to the 4th alternative of the embodiment of the invention.In Fig. 7, whether use_base_prediction_flag (use fundamental forecasting sign) has represented present image corresponding to the sign of key images.
Although in the traditional standardized document, whether use the motion_refinement_flag of motion refinement technique, only check motion_refinement_flag in the present invention for the image that is not key images for all images inspection representative.
[the 5th alternative]
In the 5th alternative, when using the AR-FGS technology, do not use the FGS motion refinement for key images; And when not using the AR-FGS technology, use the FGS motion refinement technique for key images.The difference of the 5th alternative and the 4th alternative is not use the FGS motion refinement technique to all key images.
Fig. 8 illustrates the grammer that is used for according to the 5th alternative of the embodiment of the invention.With reference to figure 8, when not using AR-FGS, (, then use AR-FGS) if adaptive_ref_fga_flag (adaptive reference FGA sign) is 1, use motion_refinement_flag to indicate whether to use motion refinement technique.
[the 6th alternative]
In the 6th alternative, do not use AR-FGS technology and FGS motion refinement technique for the key images among the SVC.In the case, although the code efficiency of encoded video signal is not high, has also reduced the bit stream complexity and reduced the propagation of video quality deterioration.
Fig. 9 illustrates the grammer that is used for according to the 6th alternative of the embodiment of the invention.With reference to figure 9, do not use for key images and to have represented the adaptive_ref_fga_flag that whether uses the AR-FGS technology and to have represented the motion_refinement_flag that whether uses motion refinement technique.
[the AR-FGS application process after the improvement]
Except first to the 6th alternative, the invention allows for the AR-FGS application process behind a kind of improve, described method more is similar to the fact of the residual signals of i FGS layer than the residual signals of base-quality layer based on the residual signals of (i-1) FGS layer, when interlayer in AR-FGS prediction is used to residual signals (when not using the FGS motion refinement technique, but although perhaps when using FGS motion refinement technique inter-layer prediction also to be used to residual signals), by using the residual signals of (i-1) FGS layer, determine the contraction-expansion factor Si of i FGS layer.
Especially, when Closing Switch K12 (212) in Fig. 2, residual signals 203 can be used to determine contraction-expansion factor S2.In the case, K switch 22 (222) is switched to node 2.
AR-FGS application process after the improvement can make up with the 3rd, the 4th and the 5th alternative.For example, when Closing Switch K22 (211) and K12 (212) both the time, determine contraction-expansion factor S1 by residual signals 207, and determine contraction-expansion factor S2 by residual signals 203.
Figure 10 is the block diagram according to the SVC encoder of AR-FGS after the employing improvement of the embodiment of the invention and FGS motion refinement technique.With reference to Figure 10, the SVC encoder comprises: prediction signal determining unit 1010 and contraction-expansion factor determining unit 1020.
When between base-quality layer or low FGS layer and current FGS layer not during inter-layer prediction, prediction signal determining unit 1010 is determined the prediction signal of current FGS layer piece according to the contraction-expansion factor of current FGS layer.When the contraction-expansion factor of current FGS layer is 0, determines the prediction signal of current FGS layer piece according to above-mentioned first alternative, and when the contraction-expansion factor of current FGS layer is not 0, determine the prediction signal of current FGS layer piece according to above-mentioned second alternative.
Contraction-expansion factor determining unit 1020 is identified for predicting the contraction-expansion factor corresponding to the higher FGS layer piece of current FGS layer piece based on the residual signals of current FGS layer piece.In the case, inter-layer prediction being set will carry out between current FGS layer and higher FGS layer.The detail operations of contraction-expansion factor determining unit 1020 is relevant with first and second alternatives.
Figure 11 is the flow chart that illustrates the operation of the illustrated SVC encoder of Figure 10.With reference to Figure 11, in operation S1010, determine whether inter-layer prediction between base-quality layer or low FGS layer and current FGS layer.When having determined between base-quality layer or low FGS layer and current FGS layer not inter-layer prediction,, then in operation S1030, determine prediction signal according to first alternative if the contraction-expansion factor of current FGS layer is 0; And if the contraction-expansion factor of current FGS layer is not 0, then in operation S1040, determine prediction signal according to second alternative.Then, in operation S1050, determine the contraction-expansion factor of higher FGS layer based on the residual signals of current FGS layer piece.
Figure 12 is employing according to the embodiment of the invention AR-FGS after corresponding to the improvement of the 3rd alternative and the block diagram of the SVC encoder of FGS motion refinement technique.With reference to Figure 12, the SVC encoder comprises: inter-layer prediction is provided with unit 1210 and contraction-expansion factor determining unit 1220.
Inter-layer prediction is provided with unit 1210 and is arranged on inevitable inter-layer prediction between basic layer and each FGS layer.Contraction-expansion factor determining unit 1220 is always determined the contraction-expansion factor of higher FGS layer based on the residual signals of basic layer.The operation of the illustrated SVC encoder of Figure 12 has been described in the 3rd alternative in more detail.
Figure 13 is the block diagram that adopts according to another embodiment of the present invention corresponding to the SVC encoder of the AR-FGS of the improvement of the 4th, the 5th and the 6th alternative and FGS motion refinement technique.With reference to Figure 13, when the image among the GOP of incoming bit stream during corresponding to key images, FGS-MR deactivated cell 1310 prevents that the FGS motion refinement technique is applied to this key images.AR-FGS deactivated cell 1320 stops the AR-FGS technology to be applied to this key images.
The 4th alternative is corresponding to the SVC encoder that only comprises FGS-MR deactivated cell 1310, and the 6th alternative is corresponding to comprising both SVC encoders of FGS-MR deactivated cell 1310 and AR-FGS deactivated cell 1320.The 5th alternative is corresponding to the SVC encoder that only just optionally uses FGS-MR deactivated cell 1310 when the AR-FGS technology is applied to key images.
As mentioned above, Figure 11,12,13 and 14 illustrated SVC encoders optionally can be made up with one of first to the 6th alternative.
The present invention can also be embodied as the computer-readable code on the computer readable recording medium storing program for performing.Computer readable recording medium storing program for performing is that can store thereafter can be by any data storage device of the data of computer system reads.The example of computer readable recording medium storing program for performing comprises: read-only memory (ROM), random-access memory (ram), CD-ROM, tape, floppy disk, optical data storage device and carrier wave (such as, by the data transmission of internet).Computer readable recording medium storing program for performing also can be distributed on the network-coupled computer system, makes with distributed way storage and computer readable code executed.
Although specifically illustrate and described the present invention with reference to example embodiment of the present invention, but those of ordinary skill in the art will understand, can carry out the various changes on form and the details therein, and not break away from the spirit and scope of the present invention that limit by following claim.
Claims (34)
1. one kind is used the AR-FGS after the improvement and the SVC encoder of FGS motion refinement technique, comprising:
The prediction signal determining unit is used for when not during inter-layer prediction, determine the prediction signal of current FGS layer piece according to the contraction-expansion factor of current FGS layer between base-quality layer or low FGS layer and current FGS layer; And
The contraction-expansion factor determining unit is used for the residual signals based on current FGS layer piece, is identified for predicting the contraction-expansion factor corresponding to the higher FGS layer piece of current FGS layer piece.
2. according to the SVC encoder of claim 1, wherein, this contraction-expansion factor is the utilization rate that is used to predict the previous piece of current FGS layer piece.
3. according to the SVC encoder of claim 1, wherein, the contraction-expansion factor of this current FGS layer is set to 0, and obtains the prediction signal of current FGS layer piece according to the same way as with the mode of the prediction signal that is used to obtain the base-quality layer piece.
4. according to the SVC encoder of claim 1, wherein, the contraction-expansion factor of this current FGS layer is not 0.
5. according to the SVC encoder of claim 1, wherein, inter-layer prediction between current FGS layer piece and higher FGS layer piece.
6. according to the SVC encoder of claim 1, wherein, the reconstruction signal of current FGS layer piece comprises prediction signal and residual signals.
7. one kind is used the AR-FGS after the improvement and the SVC encoder of FGS motion refinement technique, comprising:
Inter-layer prediction is provided with the unit, is used to be arranged on inevitable inter-layer prediction between basic layer and each FGS layer; And
The contraction-expansion factor determining unit is used for determining based on the residual signals of basic layer the contraction-expansion factor of higher FGS layer.
8. according to the SVC encoder of claim 7, wherein, this contraction-expansion factor is the utilization rate that is used to predict the previous piece of current FGS layer piece.
9. one kind is used the AR-FGS after the improvement and the SVC encoder of FGS motion refinement technique, comprise: the FGS-MR deactivated cell, when being used for image as the GOP of incoming bit stream, prevent from the FGS motion refinement technique is applied to described key images corresponding to key images.
10. according to the SVC encoder of claim 9, wherein, only when the AR-FGS technology not being applied to key images, the FGS-MR deactivated cell just allows the FGS motion refinement technique is applied to key images.
11., also comprise being used to prevent that the AR-FGS technology is applied to the AR-FGS deactivated cell of key images according to the SVC encoder of claim 9.
12. one kind is used the AR-FGS after the improvement and the SVC encoder of FGS motion refinement technique, comprising:
Inter-layer prediction is provided with the unit, is used to be arranged on inter-layer prediction between basic layer and each FGS layer; And
The contraction-expansion factor determining unit is used for the residual signals based on current FGS layer piece, is identified for predicting the contraction-expansion factor of the higher FGS layer piece corresponding with current FGS layer piece.
13. the SVC encoder according to claim 12 also comprises: when FGS-MR deactivated cell, this FGS-MR deactivated cell are used for image as the GOP of incoming bit stream corresponding to key images, prevent from the FGS motion refinement technique is applied to key images.
14. according to the SVC encoder of claim 13, wherein, only when the AR-FGS technology was not applied to key images, the FGS-MR deactivated cell just allowed the FGS motion refinement technique is applied to key images.
15. the coding method in the SVC encoder that uses AR-FGS after the improvement and FGS motion refinement technique, this coding method comprises:
When between base-quality layer or low FGS layer and current FGS layer, not during inter-layer prediction, determining the prediction signal of current FGS layer piece according to the contraction-expansion factor of current FGS layer; And
Based on the residual signals of current FGS layer piece, be identified for predicting the contraction-expansion factor of the higher FGS layer piece corresponding with current FGS layer piece.
16. according to the coding method of claim 15, wherein, this contraction-expansion factor is the utilization rate that is used to predict the previous piece of current FGS layer piece.
17. according to the coding method of claim 15, wherein, the contraction-expansion factor of this current FGS layer is set to 0, and obtains the prediction signal of current FGS layer piece according to the same way as with the mode of the prediction signal that is used to obtain base layer block.
18. according to the coding method of claim 15, wherein, the contraction-expansion factor of current FGS layer is not 0.
19. according to the coding method of claim 15, wherein, inter-layer prediction between current FGS layer piece and higher FGS layer piece.
20. according to the coding method of claim 15, wherein, the reconstruction signal of current FGS layer piece comprises prediction signal and residual signals.
21. the coding method in the SVC encoder that uses AR-FGS after the improvement and FGS motion refinement technique, this coding method comprises:
Be arranged on inevitable inter-layer prediction between basic layer and each FGS layer; And
Determine the contraction-expansion factor of higher FGS layer based on the residual signals of basic layer.
22. according to the coding method of claim 21, wherein, this contraction-expansion factor is the utilization rate that is used to predict the previous piece of current FGS layer piece.
23. the coding method in the SVC encoder that uses AR-FGS after the improvement and FGS motion refinement technique comprises:, prevent that the FGS motion refinement technique is applied to described key images when the image among the GOP of incoming bit stream during corresponding to key images.
24., wherein, only when the AR-FGS technology not being applied to key images, just allow the FGS motion refinement technique is applied to key images according to the coding method of claim 23.
25. the coding method according to claim 23 also comprises: prevent from the AR-FGS technology is applied to key images.
26. the coding method in the SVC encoder that uses AR-FGS after the improvement and FGS motion refinement technique, this coding method comprises:
Be arranged on inter-layer prediction between basic layer and each FGS layer; And
Based on the residual signals of current FGS layer piece, be identified for predicting the contraction-expansion factor of the higher FGS layer piece corresponding with current FGS layer piece.
27. the coding method according to claim 26 also comprises:, prevent from the FGS motion refinement technique is applied to key images when the image among the GOP of incoming bit stream during corresponding to key images.
28., wherein, only when the AR-FGS technology is not applied to key images, just allow the FGS motion refinement technique is applied to key images according to the coding method of claim 27.
29. one kind is used the AR-FGS after the improvement and the SVC decoder of FGS motion refinement technique, wherein, in the operation that is used for each FGS layer is decoded, when the FGS motion refinement technique is applied to current FGS layer and between current FGS layer and base-quality layer or low FGS layer not during inter-layer prediction, contraction-expansion factor according to current FGS layer comes the prediction signal of current FGS layer piece is decoded, and this contraction-expansion factor is to be determined based on the residual signals of current FGS layer piece by the SVC encoder.
30. according to the SVC decoder of claim 29, wherein, the contraction-expansion factor of this current FGS layer is set to 0.
31. one kind is used the AR-FGS after the improvement and the SVC decoder of FGS motion refinement technique, when the bit stream that received of configuration so that must inter-layer prediction between basic layer and each FGS layer the time, this SVC decoder must be determined the contraction-expansion factor of higher FGS layer based on the residual signals of basic layer, so that the bit stream that is received is decoded.
32. SVC decoder that uses the AR-FGS after the improvement, when reception comprises the bit stream of inter-layer prediction signalization, this SVC decoder is identified for predicting that based on the residual signals of current FGS layer piece the contraction-expansion factor of the higher FGS layer piece corresponding with current FGS layer piece, this inter-layer prediction signalization have been represented inter-layer prediction is set will carrying out between basic layer and each FGS layer.
33. one kind is used the AR-FGS after the improvement and the SVC coding/decoding method of FGS motion refinement technique, this SVC coding/decoding method comprises:
Determine that whether present frame is corresponding to key images; And
, determine whether to use the AR-FGS technology, and when present frame does not correspond to this key images, determine whether to use the FGS motion refinement technique during at present frame corresponding to key images.
34. the SVC coding/decoding method according to claim 33 also comprises: corresponding to key images and when not using the AR-FGS technology, determine whether to use the FGS motion refinement technique at present frame.
Applications Claiming Priority (4)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| US85204606P | 2006-10-16 | 2006-10-16 | |
| US60/852,046 | 2006-10-16 | ||
| KR1020070002653 | 2007-01-09 | ||
| KR1020070104240 | 2007-10-16 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN101669366A true CN101669366A (en) | 2010-03-10 |
Family
ID=41804864
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN200780045636A Pending CN101669366A (en) | 2006-10-16 | 2007-10-16 | Scalable video coding encoder with adaptive reference fgs and fgs motion refinement mechanism and method thereof |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN101669366A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103918261A (en) * | 2011-07-21 | 2014-07-09 | 卢卡·罗萨托 | Signal processing and inheritance in a tiered signal quality hierarchy |
| CN104904214A (en) * | 2013-01-07 | 2015-09-09 | Vid拓展公司 | Motion information signaling for scalable video coding |
| CN109417617A (en) * | 2016-06-22 | 2019-03-01 | 韩国电子通信研究院 | Intra-frame prediction method and device |
-
2007
- 2007-10-16 CN CN200780045636A patent/CN101669366A/en active Pending
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103918261A (en) * | 2011-07-21 | 2014-07-09 | 卢卡·罗萨托 | Signal processing and inheritance in a tiered signal quality hierarchy |
| CN103918261B (en) * | 2011-07-21 | 2018-01-19 | 卢卡·罗萨托 | Signal transacting and succession in hierarchical signal quality level |
| CN104904214A (en) * | 2013-01-07 | 2015-09-09 | Vid拓展公司 | Motion information signaling for scalable video coding |
| CN109417617A (en) * | 2016-06-22 | 2019-03-01 | 韩国电子通信研究院 | Intra-frame prediction method and device |
| CN109417617B (en) * | 2016-06-22 | 2023-10-20 | Lx 半导体科技有限公司 | Intra-frame prediction method and device |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| RU2668723C2 (en) | Method and equipment for coding and decoding video signals | |
| JP6874032B2 (en) | Picture encoding / decoding method and equipment using this | |
| AU2012275789B2 (en) | Motion prediction in scalable video coding | |
| RU2613738C2 (en) | Signaling of state information for decoded picture buffer and reference picture lists | |
| CN101010964B (en) | Method and apparatus for using frame rate up conversion techniques in scalable video coding | |
| EP2106666B1 (en) | Improved inter-layer prediction for extended spatial scalability in video coding | |
| RU2387094C1 (en) | Method and device for coding/decoding video signal | |
| CN112243128B (en) | Inter-layer prediction method and method of transmitting bit stream | |
| US20090041130A1 (en) | Method of transmitting picture information when encoding video signal and method of using the same when decoding video signal | |
| KR20080094041A (en) | Video coding with fine granularity spatial scalability | |
| MX2008000569A (en) | Coding dependency indication in scalable video coding. | |
| KR20100015611A (en) | Inter-view prediction with downsampled reference pictures | |
| KR20060122664A (en) | Method for decoding a video signal encoded in inter-layer prediction manner | |
| JP2009522971A (en) | Switchable filter upsampling mechanism for scalable video coding | |
| WO2007109993A1 (en) | Error control system, method, encoder and decoder for video coding | |
| CN101945285A (en) | Embedded interactive application service transcoding and encoding system | |
| KR102407912B1 (en) | Bidirectional intra prediction signaling | |
| JP4284265B2 (en) | Moving picture coding apparatus, moving picture coding method, moving picture decoding apparatus, and moving picture decoding method | |
| CN101669366A (en) | Scalable video coding encoder with adaptive reference fgs and fgs motion refinement mechanism and method thereof | |
| KR20080004392A (en) | Scalable video encoding/decoding method and apparatus thereof | |
| WO2006132509A1 (en) | Multilayer-based video encoding method, decoding method, video encoder, and video decoder using smoothing prediction | |
| US20100046620A1 (en) | Scalable video coding encoder with adaptive reference fgs and fgs motion refinement mechanism and method thereof | |
| KR102602633B1 (en) | Method and apparatus for image encoding/decoding | |
| KR100759722B1 (en) | H.264-based scalable encoding method for performing motion-compensation using different prediction structure according to frame size based layers | |
| KR100488421B1 (en) | Lossy coding method of binary image |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| C02 | Deemed withdrawal of patent application after publication (patent law 2001) | ||
| WD01 | Invention patent application deemed withdrawn after publication |
Application publication date: 20100310 |