专利汇可以提供PREDICTION OF IMAGES BY REPARTITIONING OF A PORTION OF REFERENCE CAUSAL ZONE, CODING AND DECODING USING SUCH A PREDICTION专利检索,专利查询,专利分析的服务。并且A method is provided for decoding a data stream representative of an image or of a sequence of images, the stream comprising data representative of at least one portion of one of the images, the method implementing a step of predicting a motion vector of a partition of a current image portion, with respect to a reference causal zone which has been partitioned according to a mode of coding determined for the zone. The prediction step comprises, in relation to said current partition: partitioning anew at least one portion of the reference causal zone into a plurality of reference partitions, determining motion vectors respectively associated with reference partitions obtained subsequent to the new partitioning, determining a predictor motion vector on the basis of said determined motion vectors, and predicting the motion vector of said current partition on the basis of said determined predictor motion vector.,下面是PREDICTION OF IMAGES BY REPARTITIONING OF A PORTION OF REFERENCE CAUSAL ZONE, CODING AND DECODING USING SUCH A PREDICTION专利的具体信息内容。
The present invention pertains generally to the field of image processing, and more precisely to the coding and to the decoding by competition of digital images and of sequences of digital images.
Several coding and decoding methods exist for the transmission of images. Mention may be made in particular of major types of coding such as so-called “intra” coding where an image is coded in an autonomous manner, that is to say without reference to other images, or else so-called “inter” coding which consists in coding a current image with respect to past images so as to express and transmit only the difference between these images.
The coding methods of the aforementioned type generally comprise a predictive coding step according to which image portions, called blocks or macroblocks, of a current image are predicted with respect to other reference blocks or macroblocks, that is to say previously coded and then decoded.
In the case for example of the H264/MPEG-4 AVC standard (AVC standing for “Advanced Video Coding”), the predictive coding of a macroblock consists in chopping the macroblocks according to a plurality of partitions generally having the form of blocks of smaller size.
More precisely, during the predictive coding of a macroblock in accordance with the aforementioned standard, with the exception of a macroblock of 16×16 type, the macroblock is generally chopped according to a plurality of partitions generally having the form of blocks of smaller size. The spatial prediction of such a macroblock in an image consists in predicting each block of smaller size forming this macroblock with respect to one or more blocks of another macroblock of this same image having already been coded and then decoded, such a macroblock being called a reference macroblock. This prediction is possible only if the reference macroblock neighbors the macroblock to be predicted and is situated in certain predetermined directions with respect to it, that is to say generally above and to the left, in a so-called “causal” neighborhood.
The drawback of such a type of prediction resides in the fact that for a given transmission rate from the coder to the decoder, it achieves worse video quality than inter coding since it does not exploit the temporal correlations between the images of the sequence.
In the case of inter coding, the current macroblock to be coded may be partitioned according to the 16×16, 8×16, 16×8 and. 8×8 modes. If the 8×8 mode is selected, each 3×8 block is again partitioned according to the 8×8, 4×8, 8×4 and 4×4 modes. Each current block is compared with one or more blocks of a reference causal zone, of spatial or temporal type, which comprises macroblocks already decoded at the level of the coder during the coding by the latter of a current macroblock. A temporal predictor is then defined by a vector which describes the motion between the current block and the reference block.
A drawback of the prediction represented in
It follows from this that such a prediction lacks precision.
A drawback of the prediction represented in
One of the aims of the invention is to remedy drawbacks of the aforementioned prior art.
For this purpose, according to a first aspect, the present invention relates to a method for decoding a data stream representative of an image or of a sequence of images, said stream comprising data representative of at least one portion of one of the images, the method implementing a step of predicting a motion vector of a partition of a current image portion, with respect to a reference causal zone which has been partitioned according to a mode of coding determined for the zone.
According to the invention, the prediction step of such a decoding method comprises, in relation to said current partition, the steps consisting in:
Such a repartitioning advantageously makes it possible to predict the current motion vector of a current partition on the basis of motion vectors belonging to more suitable reference partitions, in particular from a shape, size and arrangement point of view, than those arising from the initial partitioning of at least one reference macroblock which was performed during the coding of this macroblock.
The invention thus makes it possible to significantly improve the precision of the prediction of the current vectors under inter coding.
It should moreover be noted that the reference motion vectors which are determined subsequent to the new partitioning are not used for the coding of the current partition.
According to a second aspect, the present invention relates to a method for coding an image or a sequence of images generating a data stream comprising data representative of at least one portion of one of the images, such a method implementing a step of predicting a motion vector of a partition of a current image portion, with respect to a reference causal zone which has been partitioned according to a mode of coding determined for this portion.
According to the invention, the prediction step of such a coding method comprises, in relation to said current partition, the steps consisting in:
According to a third aspect, the present invention relates to a method for predicting a motion vector of a partition of a current image portion, with respect to at least one reference causal zone which has been partitioned according to a mode of coding determined for the zone.
According to the invention, such a prediction method comprises, in relation to the current partition, the steps consisting in:
In one embodiment, the reference partitions obtained subsequent to the new partitioning are in the closest neighborhood of the current partition.
Such an arrangement thus makes it possible to further refine the precision of the prediction.
In another embodiment, a reference partition has the same position as the current partition associated with it, in an image belonging to the reference causal zone.
In yet another embodiment, the reference partition is of identical shape and identical size to those of the current partition associated with it.
Such an arrangement substantially increases the precision of the temporal prediction of the current vectors. Indeed, the reference motion vectors obtained thus describe a motion closer to the real motion between the current image portion and the reference image portion than the motion described by the reference motion vectors of the prior art.
According to yet other embodiments, the reference partitions obtained subsequent to the new partitioning overlap or are disjoint.
Correlatively, according to a fourth aspect, the present invention relates to a device for decoding a data stream representative of an image or of a sequence of images, the stream comprising data representative of at least one portion of one of the images, such a decoding device comprising a device for predicting a motion vector of a partition of a current image portion, with respect to a reference causal zone which has been partitioned according to a mode of coding-determined for the zone.
According to the invention, the prediction device of such a decoding device comprises:
Correlatively, according to a fifth aspect, the present invention relates to a device for coding an image or a sequence of images generating a data stream comprising data representative of at least one portion of one of the images, such a device comprising a device for predicting a motion vector of a partition of a current image portion, with respect to a reference causal zone which has been partitioned according to a mode of coding determined for said zone.
According to the invention, the prediction device of such a coding device comprises:
Correlatively, according to a sixth aspect, the present invention relates to a device for predicting a motion vector of a partition of a current image portion, with respect to a reference causal zone which has been partitioned according to a mode of coding determined for the zone.
According to the invention, such a prediction device comprises:
The invention further relates to a computer program comprising instructions for implementing one of the methods according to the invention, when it is executed on a computer.
The coding method, the prediction method, the decoding device, the coding device and the prediction device exhibit at least the same advantages as those afforded by the decoding method according to the present invention.
Other characteristics and advantages will become apparent on reading preferred embodiments described with reference to the figures in which:
An embodiment of the invention will now be described, in which the coding method according to the invention is used to code a sequence of images according to a binary stream close to that obtained by a coding according to the H.264/MPEG-4 AVC standard. In this embodiment, the coding method according to the invention is for example implemented in a software or hardware manner by modifications of a coder complying initially with the H.264/MPEG-4 AVC standard. The coding method according to the invention is represented in the form of an algorithm comprising steps C0 to C7, represented in
It should be noted that the decoding method according to the invention is also implemented in a software or hardware manner by modifications of a decoder complying initially with the H.264/MPEG-4 AVC standard.
The coding method according to the invention is implemented in a coding device CO represented in
With reference to
This partitions selection module SP uses for example a procedure for choosing by exhaustive competition or else a procedure for choosing with the aid of an algorithm with a-priori. Such procedures are well known to the person skilled in the art (cf: G. J. Sullivan and T. Wiegand, “Rate-distortion optimization for video compression” IEEE Signal Proc. Mag., pp. 74-90, 1998) They will not therefore be described hereinbelow.
Said partitions are grouped together in a database BD of the coder CO. Such partitions may be of rectangular or square shape or else of other geometric shapes, such as for example substantially linear shapes,
In the example represented, the selection module SP selects a square-shaped partition of size 4×4.
The following step C1 represented in
Subsequent to the partitioning step C1, in the course of a step C2 represented in
Such a prediction module PREDCO is intended to predict the motion vector of each partition P1, . . . P16 of the current macroblock MBpart, on the basis of a predictor motion vector calculated on the basis of a portion of an already coded and then decoded reference causal zone. In the example represented in
With reference to
Still with reference to
In the course of a step C3 represented in
In the course of a step C4 represented in
In the course of a step C5, the calculation module CAL2 of
MVp1=Med(MVr1, MVr2 and MVr3)
In the course of a step C6, the prediction module PREDCO of
The prediction calculation module PREDCO of
The spatial prediction which has just been described hereinabove is particularly efficacious because of the fact that the motion vector of the current partition P1 is predicted on the basis of motion vectors which belong to reference partitions whose shape and size are much more suitable than in the prior art.
Steps C3 to C6 are thereafter repeated so as to predict the current motion vectors (not represented in
Once various possible predictions have been calculated by the prediction calculation module PREDCO, in the course of a step C7 represented in
Each predicted macroblock is coded, in the course of a step C8, as in the H.264/MPEG-4 AVC standard.
Once this structural coding has been performed by the decision module DCNCO, the coefficients of residuals if they exist, corresponding to the blocks of the image IE, are dispatched to the transform and quantization module TQCO, to undergo discrete cosine transforms followed by a quantization. The slices of macroblocks with these quantized coefficients are thereafter transmitted to an entropy coding module CE so as to produce, with the other images of the video sequence that are already coded in the same manner as the image IE, a binary video stream F coded according to the invention.
The binary stream F is firstly dispatched to an entropy decoding module DE, decoding being inverse to that performed by the entropy coding module CE represented in
An image reconstruction module RI then receives decoded data corresponding to the data produced by the module DCNCO (
The first step D0 is the decoding of data structures coded in a slice of a current macroblock of the image IE to be decoded. In a manner known per se, the reconstruction module RI determines on the basis of the data of said macroblock slice:
The following step D1 represented in
In the course of a step D2 represented in
In the course of steps D3 to D6, the prediction module PREDDO performs the same algorithm as that performed by the prediction module PREDDO of the aforementioned coder CO, so as to obtain predicted motion vectors in accordance with the procedure described hereinabove.
In the course of a step D7, a decision module DCNDO chooses the optimal prediction according to a rate distortion criterion well known to the person skilled in the art.
Each predicted macroblock is thereafter decoded, in the course of a step D8, as in the H.264/MPEG-AVC standard.
Once all the macroblocks of the image IE have been decoded, the image reconstruction module RI provides as output from, the decoder DO, an image ID corresponding to the decoding of the image IE.
Having regard to the fact that the prediction algorithm performed at the decoder DO is in every respect the same as that performed at the coder CO, the information cost induced by the predictors used is thereby greatly reduced.
The second embodiment described hereinbelow is distinguished from the previous embodiment mainly by the fact that the prediction module PREDCO represented in
In accordance with the second embodiment, the aforementioned step of selecting partitions C0 is not performed.
In the course of step C1 of
In the example represented in
In the course of step C2 of
In the second embodiment described here, the prediction module PREDCO is intended to predict the three motion vectors associated respectively with the three partitions P′1, P′2 and P′3 of the current macroblock MBpart, on the basis respectively of three reference motion vectors of one and the same reference macroblock, denoted MBrIE-1, having the same position as the current macroblock MBpart in the previous image IE-1. In the example represented in
As mentioned hereinabove with reference to the first embodiment, the reference macroblock MBrIE-1 has previously been coded and then decoded.
In the course of step C3 of
In the course of step C4 represented in
In the course of step C5 of
In the course of step C6, the prediction module PREDCO of
The prediction calculation module PREDCO of
The following steps of choosing optimal prediction C7 and of coding C8 are respectively similar to the aforementioned steps C7 and C8 of the first embodiment and for this reason, will not be described again.
The reconstruction module RI of the decoder DO thereafter implements steps D0 to D8 similar to those of the decoding method in accordance with the first embodiment.
The first step D0 is the decoding of data structures coded in a slice of a current macroblock of the image IE to be decoded. In a manner known per se, the reconstruction module RI determines on the basis of the data of said macroblock slice:
In the course of the following step D1, the macroblocks partitioning module PMBDO chops the macroblock according to the aforementioned three partitions P′1, P′2 and P′3.
Subsequent to the partitioning step D1, in the course of a step D2 similar to that of the aforementioned first embodiment, the partitioning module PMBDO transmits the current macroblock to be decoded and which has just been partitioned into three partitions, to the aforementioned prediction module PREDDO.
In the course of steps D3 to D6, the prediction module PREDDO performs the same algorithm as that performed by the module PREDCO, in the course of the aforementioned steps C3 to C6.
In the course of step D7, the decision module DCNDO chooses the optimal prediction according to a rate distortion criterion well known to the person skilled in the art.
Each predicted macroblock is thereafter decoded, in the course of step D8, as in the H.264/MPEG-AVC standard.
Once all the macroblocks of the image IE have been decoded, the image reconstruction module RI provides as output from the decoder DO, an image ID corresponding to the decoding of the image IE.
It goes without saying that the embodiments which have been described hereinabove have been given purely by way of wholly non-limiting indication, and that numerous modifications may easily be made by the person skilled in the art without, however, departing from the scope of the invention.
Thus for example, the reference macroblock MBrIE-1 may be chopped differently from that of the current macroblock to be predicted MBpart.
Moreover, the reference partitions Pr′1, Pr′2 and Pr′3 overlap.
标题 | 发布/更新时间 | 阅读量 |
---|---|---|
大规模协同环境下的协同交互优化方法 | 2020-06-10 | 864 |
一种海洋生态环境损害因果关系判定方法 | 2020-05-17 | 211 |
基于ARINC653操作系统生成代码的方法与装置 | 2020-05-21 | 553 |
顾及城镇格局特征的人口分布时空演变与认知 | 2020-05-11 | 630 |
一种保证事务一致性和线性一致性的分布式系统和方法 | 2020-05-13 | 779 |
Determining event causality including employment of partitioned event space | 2020-06-22 | 161 |
Ensuring globally consistent transactions | 2020-05-23 | 620 |
ESTIMATION DE MOUVEMENT AU NIVEAU DU DÉCODEUR PAR MISE EN CORRESPONDANCE MODÈLES | 2020-05-31 | 843 |
ESTIMATION DE MOUVEMENT AU NIVEAU DU DÉCODEUR PAR MISE EN CORRESPONDANCE MODÈLES | 2020-06-08 | 41 |
PROCÉDÉ DE CODAGE ET DE DÉCODAGE D'IMAGES, DISPOSITIF DE CODAGE ET DE DÉCODAGE D'IMAGES ET PROGRAMMES D'ORDINATEUR CORRESPONDANTS | 2020-06-09 | 279 |
高效检索全球专利专利汇是专利免费检索,专利查询,专利分析-国家发明专利查询检索分析平台,是提供专利分析,专利查询,专利检索等数据服务功能的知识产权数据服务商。
我们的产品包含105个国家的1.26亿组数据,免费查、免费专利分析。
专利汇分析报告产品可以对行业情报数据进行梳理分析,涉及维度包括行业专利基本状况分析、地域分析、技术分析、发明人分析、申请人分析、专利权人分析、失效分析、核心专利分析、法律分析、研发重点分析、企业专利处境分析、技术处境分析、专利寿命分析、企业定位分析、引证分析等超过60个分析角度,系统通过AI智能系统对图表进行解读,只需1分钟,一键生成行业专利分析报告。