EP3991433A1

EP3991433A1 - Method for encoding and decoding an image of a video sequence, and associated device

Info

Publication number: EP3991433A1
Application number: EP20734037.3A
Authority: EP
Inventors: Félix Henry; Gordon Clare
Original assignee: Fondation B Com
Current assignee: Fondation B Com
Priority date: 2019-06-25
Filing date: 2020-06-24
Publication date: 2022-05-04
Also published as: JP2022539552A; FR3098073B1; KR20220024714A; FR3098073A1; WO2020260412A1; BR112021025469A2; US12113961B2; CN114051729A; US20220368889A1

Abstract

Proposed is a method for encoding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, wherein the method involves encoding at least one block of the chrominance component using a BDPCM mode.

Description

Method of encoding and decoding an image of a video sequence and associated device

Technical field of the invention

The present invention relates generally to the technical field of video compression.

It relates in particular to a method for encoding and decoding an image of a video sequence and associated device.

State of the art

Video compression standards such as HEVC and post HEVC ("High Efficiency Vido Coding" standards, described for example in "High Efficiency Video Coding, Coding Tools and Specification", Matthias Wien, Signais and Communication Technology, 2015) describe a step of prediction of part of the image to be coded from a reference image. This can be located in the same image as the zone considered (INTRA prediction mode), or in another image of the video sequence (INTER prediction mode).

A mode of prediction of the elements of a block, called BDPCM (for "Block Differential Puise Code Modulation", in English), comprises a prediction of the value of a pixel of an image from that (s) of one or more of its neighbors. Different variants of this mode exist. A first variant is based on the prediction of a quantified residue as described in “CE8-related: Quantized residual BDPCM, Marta Karczewicz, Video Experts Team (JVET) of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29 / WG 1 1, 14th Meeting: Geneva, CH, 19-27 March 2019 ”. A second variant is based on the values of reconstructed pixels, located in the vicinity of the considered area, as described in “CE8: BDPCM with horizontal / vertical predictor and independently decodable areas (test 8.3. 1b), Mohsen Abdoli, Gordon Clare, Félix Henry, Pierrick Philippe, Joint Video Experts Team (JVET) of ITU-T SG 16 WP 3 and ISO / IEC JTC 1 / SC 29 / WG 1 1, 13th Meeting: Marrakech, MA, 9-18 Jan. 2019 ”.

Presentation of the invention

In this context, the invention proposes a method for coding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the method comprising coding of at least one block of the chrominance component, using a so-called BDPCM mode.

Using the BDPCM mode for encoding a block of the chrominance component can improve the compression gain of the image.

Other non-limiting and advantageous characteristics of the process according to the invention, taken individually or in any technically possible combination, are as follows.

According to one embodiment, at least one block of the chrominance component can be encoded using the BDPCM mode when a predetermined criterion is satisfied.

Thus, it is possible to choose the chrominance components to which the BDPCM mode applies, for example when these satisfy a predetermined criterion.

For example, the predetermined criterion is satisfied when said two components are of identical size.

The criterion can also be an encoding criterion indicative of a choice made by an encoder able to implement the encoding method. This choice made by the encoder results, for example, from an analysis of the video sequence. The choice of the encoder can thus be a function of a histogram, in other words a statistical graph, representing the distribution of the intensities of the pixels of one or more images of the video sequence, that is to say the number of pixels for each light intensity.

According to one embodiment, the encoding may further comprise the encoding of information indicative of the BDPCM mode used to encode the block of the chrominance component.

According to one embodiment, when the predetermined criterion is satisfied, an indicator, indicating the use of the BDPCM mode, is coded for the block of the chrominance component.

According to another aspect, the invention also proposes a method for coding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the method comprising for at least one block of the chrominance component, an evaluation of a so-called BDPCM mode and another mode for encoding said block, and depending on the result of the evaluation, a selection of one of the modes to encode said block, and an encoding of information representative of the mode selected to encode said block.

By evaluating the BDPCM mode for at least one block of the chrominance component, the BDPCM mode can be applied if the result of the evaluation is satisfactory. The gain in compression of the image is thus improved. The chrominance components for which it is not relevant to apply the BDPCM mode can thus be discarded, if the evaluation does not give a satisfactory result.

According to one embodiment, the BDPCM mode and the other mode can be evaluated for at least one block of the chrominance component when a predetermined criterion is satisfied.

According to one embodiment, the predetermined criterion can be satisfied when said two components are of identical size.

According to one embodiment, the encoding may further comprise the encoding of information indicating the sampling format of the image, said information being indicative of the sizes of said components.

According to one mode of implementation, the evaluation can include coding of the block of the chrominance component considered according to the BDPCM mode on the one hand and according to the other mode on the other hand, and according to another criterion. predetermined (for example a rate-distortion cost), a selection of the block encoded according to the BDPCM mode or of the block encoded according to the other mode.

According to one mode of implementation, the method can further comprise, for at least one block of the luminance component, an evaluation of the so-called BDPCM mode and of the other mode for encoding said block of the luminance component, and as a function of from the result of the evaluation, a coding of information representative of one of the modes for coding said block of the luminance component.

According to one embodiment, the components preferably have the same partitioning into blocks, the coding of the image being carried out block by block.

As a variant, the coding of the image can be carried out component by component. According to one mode of implementation, the coding can also comprise, for the image considered,

the coding of a first information relating to at least said image, indicating that the BDPCM mode can be used to encode at least one block of at least one component of the image, and if the first information indicates that the BDPCM mode can be used to encode at least one block of at least one component of the image, encoding of a second piece of information relating to at least said image, indicating that the BDPCM mode can be used to encode at least one block of the component chroma of the image.

According to one embodiment, the method can comprise, for at least one block of the chrominance component, if the second information relating to said image indicates that the BDPCM mode can be used to code at least one block of the chrominance component. of the image, the encoding of an indicator relating to at least one block of the chrominance component, indicating that the BDPCM mode is used to encode said block of the chrominance component.

According to one embodiment, the method can comprise for at least one block of the chrominance component, if said criterion is satisfied, the coding of an indicator relating to said block, indicating that the BDPCM mode is used to code said block. of the chrominance component.

According to one embodiment, the image further comprises a second chrominance component having the same partitioning as the other chrominance component, and if the BDPCM mode is selected for the block of the other chrominance component, said information relating to the block of the other chrominance component may include an indicator indicating that the BDPCM mode is selected for the block of the other chrominance component and the corresponding block of the second chrominance component, and a direction indicator, allowing determining a block predictor for the block of the other chrominance component and the corresponding block of the second chrominance component.

According to one embodiment, a block is formed of elements, the coding of a block using the BDPCM mode can comprise, for at least one element of the block, the coding of a quantized value of a residue , representative of a difference between a first and a second value, the first value corresponding to a value quantified of a difference between the value of the element considered and a reference element whose position is indicated by the given direction, and the second value corresponding to the quantized value of a difference between said other element and another element of reference signaled by the given direction.

According to another aspect, the invention also proposes a method of coding an image of a video sequence comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, the components of luminance and of chrominance having the same partitioning into blocks, the method can comprise for at least one block of the luminance component, an evaluation of a mode called BDPCM and of another mode for encoding said block, and

as a function of the result of the evaluation, a selection of one of the modes to encode said block of the luminance component and the corresponding block of the chrominance component, and a coding of information representative of the mode selected to encode the block of the luminance component and corresponding block of the chrominance component.

Preferably, if the BDPCM mode is selected, said information relates to the block of the luminance component, and comprises an indicator indicating that the mode selected for the block of the luminance component and for the corresponding block of the chrominance component is the BDPCM mode and a direction indicator for determining a block predictor for the block of the luminance component and for the corresponding block of the chrominance component.

According to another aspect, the invention also proposes a method for decoding an encoded image, from a stream of encoded video data, the image comprising at least two components, including a component representative of the luminance and a component. representative of the chrominance, each component being partitioned into blocks, the method comprising a decoding of information indicating for at least one block of the chrominance component, whether or not it is encoded using a mode called BDPCM .

When the image further comprises a second chrominance component having the same partitioning as said chrominance component (or "first" chrominance component), the decoded value of said information may also be used to indicate whether the corresponding block of the second chrominance component is encoded using a so-called BDPCM mode.

Said decoded value of said information may also indicate a direction, making it possible, on the basis of respective block predictors, a function of said direction, to decode the block of said (first) chrominance component and the corresponding block of the second chrominance component.

According to another aspect, the invention also proposes a method for decoding an encoded image, from a stream of encoded video data, the image comprising at least two components, including a component representative of the luminance and a component. representative of the chrominance, each component being partitioned into blocks, the method comprising a decoding of a first item of information representative of an encoding criterion, and if the decoded value of the information indicates that the encoding criterion is satisfied, decoding of a second item of information indicating, for at least one block of the chrominance component, whether or not it is encoded using a so-called BDPCM mode.

Preferably, for the last two aspects of the aforementioned invention, the image can further comprise a second chrominance component having the same partitioning as the other chrominance component, and the decoded value of the second information is also used for indicate whether the corresponding block of the second chrominance component is encoded using a so-called BDPCM mode, said decoded value of the second information being able also to indicate a direction, allowing from respective block predictors, a function of said direction, to decode the block of the other chrominance component and the corresponding block of the second chrominance component.

Preferably, the method may further comprise decoding other information indicating for at least one block of the chrominance component, whether it is encoded using a so-called BDPCM mode.

As already indicated, the encoding criterion may for example be indicative of a choice made by an encoder having encoded said video data stream. This choice made by the encoder may result from an analysis of the video data stream.

According to another aspect, the invention also proposes a method of decoding an encoded image, from a stream of encoded video data, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the method comprising a decoding of a first item of information relating to at least said image, indicating that the BDPCM mode can be used to decode at least one block of at least one component of the image, and if the first item of information indicates that the BDPCM mode can be used to decode at least one block of at least one component of the image, decoding of a second item of information relating to at least said image, indicating that the BDPCM mode can be used to decode at least one block of the chrominance component of the picture.

The method can also comprise, if the second information relating to said image indicates that the BDPCM mode can be used to decode at least one block of the chrominance component of the image, the decoding an indicator relating to said at least one block of the chrominance component, indicating that the BDPCM mode can be used to decode said block of the chrominance component.

According to another aspect, the invention also proposes a method for decoding an encoded image, from a stream of encoded video data, the image comprising at least two components, including a component representative of the luminance and a component. representative of the chrominance having the same partitioning into blocks, the method comprising a decoding of a first item of information representative of an encoding criterion, and if the decoded value of the information indicates that the encoding criterion is satisfied, a decoding of a second item of information relating to at least one block of the luminance component indicating whether the block of the luminance component and the corresponding block of the chrominance component are or are not encoded using a so-called BDPCM mode .

According to another aspect, the invention also proposes a device for encoding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the device comprising at least one microprocessor configured to implement an encoding of at least one block of the chrominance component, using a mode called BDPCM.

According to another aspect, the invention also proposes a device for decoding an encoded image from a stream of encoded video data, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the device comprising at least one microprocessor configured to implement a decoding of information indicating for at least one block of the chrominance component, whether it is encoded or not using a mode called BDPCM.

Detailed description of the invention

In addition, various other characteristics of the invention emerge from the appended description given with reference to the drawings which illustrate non-limiting embodiments of the invention and in which:

- Figure 1 illustrates different color formats;

FIG. 2 illustrates different color components according to a given color format;

- Figure 3 shows a first mode of implementation of a coding method according to the invention;

- Figure 4 shows a second embodiment of a coding method according to the invention;

- Figure 5 shows a third embodiment of a coding method according to the invention;

- Figure 6 shows a fourth embodiment of a coding method according to the invention;

- Figure 7 shows a fifth embodiment of an encoding method according to the invention;

- Figure 8 shows a first mode of implementation of a decoding method according to the invention;

- Figure 9 shows a second embodiment of a decoding method according to the invention;

FIG. 10 illustrates an exemplary implementation of a step of an embodiment of an encoding / decoding method according to the invention; and

- Figure 11 schematically illustrates an embodiment of a device according to the invention.

FIG. 1 illustrates for a given color standard (for example for the “YUV” or “YCbCr” standard), different sampling patterns which can be applied to components representative of the chrominance of an image. Conventionally, an image comprises a component representative of the luminance referenced Y here (hereinafter called “luminance component” or “luma” for the purposes of simplification), and two chrominance components (hereinafter called “chrominance component” or “chroma” for simplicity). The latter are respectively referenced Cb and Cr in FIG. 1.

Figure 1 illustrates three formats, each associated with a sampling pattern, for an image area consisting of 8 pixels Pxi, forming two blocks of 2 pixels by 2 pixels. Since human vision is less sensitive to color than to brightness, less chrominance information is generally retained than luminance information without degrading the perceived quality of the image.

In the first format or structure called 4: 4: 4, none of the components are downsampled. Each pixel Pxi comprises three color components.

In the second format called 4: 2: 2, the chrominance components are downsampled for every other pixel.

In the third so-called 4: 2: 0 format, only the chrominance components of the first pixel of a block formed of four pixels (2 pixels ^* 2 pixels) are retained.

Other formats exist, the examples given above are not exhaustive.

Thus, as illustrated in FIG. 2 for the 4: 2: 0 format, when a chrominance component Cb, Cr is downsampled, its size is smaller than that of the luminance component Y. For example for the format 4 : 2: 0, the size of the chrominance components Cb and Cr is divided by 4 with respect to the size of the luminance.

Each component is partitioned into BLC blocks. The partitioning into blocks may be identical for each of the components as illustrated in FIG. 2 or particular for one or a set of components.

In a first embodiment illustrated in FIG. 3, it is proposed to code, during a step E1, at least one block of the chrominance component according to a mode called BDPCM (for “Block Differential Poise Code Modulation”) , in particular when a predetermined criterion CT is satisfied. As a variant, the BDPCM mode can for example be applied by default.

Applying BDPCM mode to chroma improves the data compression rate. It is possible to apply this mode when a certain criterion is satisfied, for example so as not to increase the complexity of the encoding process. Different possible criteria are described below.

These so-called encoding criteria are not linked to the quality of the compression itself, but to a decision by the encoder to use or not a tool depending on considerations such as: encoding time, relevance of the tool in question for the type of content to be encoded, the energy consumption, or the targeted bit rate (this list is not exhaustive).

In other words, an encoding criterion is a general choice of the encoder, in relation to an encoding context that can evolve from one content to another or in relation to a targeted performance to another. It is thus distinguished from a criterion which is based solely on local information specific to a given block or which results from an implementation choice made by a user.

As explained above, several variants of the BDPCM mode exist.

In a first variant (called R-BDPCM) described in the aforementioned article “CE8-related: Quantized residual BDPCM”, when the luminance component of a block is coded according to the BDPCM mode, a direction information item (for example a indicator) is coded for the luminance block considered. The indicator is representative of a vertical or horizontal direction.

In this first variant, an INTRA type prediction is performed on the elements of the block considered initially. This prediction is identical to that described in the HEVC standard. Thus, if the direction indicator for the block corresponds to a horizontal direction, it is the horizontal INTRA prediction which is applied, i.e. the value of the element of the considered block is predicted ( or one of its components), from the value of the neighboring element, to the left of this element. Otherwise, the vertical INTRA prediction is applied, i.e. the value of the element of the block considered (or of one of its components) is predicted from the value of a neighboring element above this element. The value of the element of the luminance component of the predicted considered block, located in row i and column j of the block, is referenced by P (i, j), the elements of a block being organized in rows and columns (i and j being integers). Secondly, the prediction residue RES1 of each element P (i, j) is calculated as below:

RES1 (i, j) = P (i, j) -PRED1 (i, j)

where PRED1 (i, j) is a predictor equal to the element to the left of P (i, j), REFLEFT (i, j), if the prediction direction defined is horizontal or to the element above P (i, j), REFTOP (ij), if the prediction direction defined is vertical.

Then the calculated prediction residue RES1 (i J) is quantified into a quantized residue called QRES1 (i, j) for the element P (i, j).

If the prior INTRA prediction was carried out in the horizontal direction, so-called residual values R (i, j) are obtained as follows:

R (i, 0) = QRES1 (i, 0)

R (i, j) = QRES1 (i, j) - QRES1 (i, j-1) for ¹0

If the previous INTRA prediction was performed in the vertical direction, the values R (i, j) are produced as follows:

R (0, j) = QRES1 (0, j)

R (i, j) = QRES1 (i, j) - QRES1 (i-1, j) for i ¹0

Then, all the values R (i, j) are encoded by known means, such as the encoding of the residual coefficients of the FIEVC standard for example.

During the decoding of the encoded data stream, if the direction of the prediction, transmitted in the stream, is horizontal, the quantized residue QRES1 (i, j) is reconstructed as follows:

On decoding, if the direction of the prediction is vertical, the quantized residue QRES1 (i, j) is reconstructed as follows:

QRES1 (i, j) = å ' _k = oR (k, j).

We can then reconstruct the reconstructed version REC1 (i, j) of pixel P (i, j), from the direction of prediction and the quantized residue QRES1 (i, j), by adding to the predictor PRED1 (i, j) , the dequantized value of QRES1 (i, j).

In a second variant (called S-BDPCM) described in the article “CE8: BDPCM with horizontal / vertical predictor and independently decodable areas (test 8.3. 1b)” above, a direction information item (for example an indicator) is also coded for the luminance component in the considered block. Likewise, the indicator is representative of a vertical or horizontal direction. In this variant, the elements are processed one after the other, in a predefined order. If the direction indicator matches the horizontal prediction direction, then all items in the first column of the block (left) are processed first, then all samples in the second column, and so on until the last column of the block. Within a column, the elements can be processed in parallel or in any order. A reference element called the predictor element, PRED2 (i, j) makes it possible to predict the element of the considered block located at the i ^th row of the j ^th column. He comes:

PRED2 (i, 0) = REFLEFT (i)

PRED2 (i, j) = REC2 (i, j-1) for j ¹0, where

REFLEFT (i) is the reference element (i.e. belonging to the decoded neighboring block) to the left of the block and to the i ^th line, and

REC2 (i, j-1) is the reconstructed element of the considered block located at the i ^th row and at the j-1 ^th column (that is to say immediately to the left of the considered element).

If the direction indicator matches the vertical prediction direction, all items in the first row of the block (top) are processed, then all samples in the second row, and so on up to the last row. Within the same row, the elements can be processed in parallel or in any order. Likewise, a reference element called the predictor element PRED2 (i, j) to predict the element located at the i ^th row of the j ^th column of the block considered is obtained as follows:

PRED2 (0, j) = REFTOP (j)

PRED2 (i, j) = REC2 (i-1, j) for i ¹0, where

REFTOP (i) is the reference element (that is to say belonging to the decoded neighboring block) above the considered block and at the j ^th column;

REC2 (i-1, j) is the decoded (or reconstructed) element of the considered block located at the i-1 ^st row and at the j ^th column (therefore, immediately above the current element).

A corresponding residue RES2 (i, j) is then calculated by the difference between the original value of the element located at the i ^th row of the j ^th column of the block considered and PRED (ij).

The residue RES2 (i, j) is quantified so as to obtain a quantified residue: QRES2 (i, j). During decoding, the quantized residue obtained from an encoded video stream, transmitted to a decoder, is then dequantized to obtain a dequantized residue: DQRES2 (i, j).

The reconstructed value REC2 (i, j) of the considered element is then obtained according to the following formula:

REC2 (i, j) = PRED2 (i, j) + DQRES2 (i, j), PRED2 (i, j). Note that REC2 (i, j) is used to predict the next element in the order of element processing.

It should be noted that in the case where the quantization used in the aforementioned variants is a uniform scalar quantization, the first and the second variant are equivalent. Indeed, if we assume that the prediction is horizontal (the reasoning is the same if it is vertical), and that the uniform scalar quantization uses a quantization step d. In this case, according to the first variant, a reconstructed element

REC1 (i, j) = REFL

According to the second variant, the reconstructed element REC2 (i, j) is written:

REC2 (i, j) = PRED2 (i, j) + DQRES2 (i, j)

= REC2 (i, j-1) + QRES2 (i, j) ^* d

= REC2 (i, j-2) + (QRES2 (i, j-1) + QRES2 (i, j)) ^* d

= REFLEFT (i) + åi _{k = 0} QRES2 (i, j) * d

So in both cases, the reconstructed element is equal to the value of the reference element REFLEFT (i), to which is added the sum of the quantized values of the residues of the previously reconstructed elements, multiplied by the quantization step. In the remainder of the text, the word BDPCM will be used to denote either of the variants of the mode.

The predetermined criterion CT to be satisfied may relate to the color format of the image. For example, when the image or several images or all the images of the video sequence is (are) not downsampled, the encoding of at least one block of the chrominance component (s) of one or more images is achieved using BDPCM mode. Another predetermined criterion CT may be a function of the overall content of one or more images or of the content of one of its components. The overall content can be analyzed by the encoder, using a histogram for example. Depending on the result of the analysis, the encoder chooses whether or not to implement a coding mode. For example, if an image or component of the image comprises images of a certain format such as “Screen Content” (literally “screen content” in French), the coding of at least one block of the chrominance component (s) of the image is performed using BDPCM mode. Another predetermined criterion may be the gain produced if the BDPCM mode is used to encode the block of the chrominance component, namely whether it is greater than a predetermined threshold or not. The list of criteria set out above is not exhaustive. Those skilled in the art will be able to use another criterion not mentioned here, which, when it is satisfied, involves the coding of at least one block of a chrominance component of an image.

FIG. 4 illustrates another embodiment of the invention, comprising, in addition to step E1, a step E2 of encoding information indicative of the BDPCM mode used to encode the block of the chrominance component.

This information can be a block-level binary indicator (that is to say relating to the block considered), which when activated for a block of the chrominance component, indicates that the BDPCM mode is used to encode the block in question.

Moreover, the information relating to a block of one chrominance component, can also be valid for the corresponding block of the other chrominance component, if the two chrominance components have the same partitioning into blocks. A single indicator is reported at the block level of the first chrominance component. In other words, a single indicator indicates whether BDPCM mode is used to encode on the one hand a block of a (first) chrominance component and on the other hand the corresponding block in the other (or second) chrominance component.

Preferably, said information may further be indicative of a prediction direction, vertical or horizontal, as explained above. As with the previous indicator, the direction information may be a binary indicator reported only for the block of the first chrominance component and also valid for the corresponding block of the second chrominance component.

FIG. 5 illustrates another embodiment of the invention. This mode of implementation comprises a step E10 comprising, in particular if the criterion CT is satisfied, an evaluation of the BDPCM mode and of another mode for coding a block of the chrominance component. As a variant, step E10 may be independent of criterion CT. In another variant, if the partitioning of the luminance and chrominance components is identical, the evaluation can be made on the corresponding luminance block and the selected mode can then be directly used to code the chrominance component.

The other mode can be, for example, a coding mode with only an INTRA prediction or an INTER prediction mode as described for the HEVC standard. The evaluation here includes coding of the chrominance component of the block considered according to the BDPCM mode on the one hand and according to the other mode on the other hand. Depending on another predetermined criterion, the block of the chrominance component encoded according to the BDPCM mode or the other mode is selected. This other predetermined criterion may for example be the bit rate-distortion cost. If this cost is greater than a given threshold, in this case the BDPCM mode is not used for encoding the chrominance component of the block considered.

Step E11 then comprises an encoding of information representative of the other encoding mode for the block of the chrominance component. Conversely, step E12 comprises encoding information representative of the BDPCM mode to encode the block of the chrominance component.

Steps E1 1 and E12 may further include coding of information representative of the CT criterion, if this is taken into account in the previous step. For example, if the criterion to be satisfied is to have an identical size for the luminance and chrominance components, the steps E1 1 and E12 can also comprise the coding of information indicating the sampling format of the image. . If this information indicates that the color format is 4: 4: 4, then the CT criterion is satisfied, because the color components have the same size in this format. For example, if the images in a video sequence can potentially be in 4: 4: 4 or 4: 2: 0 format, then the information can be a binary indicator which when activated, signals that the image (or sequence if used at sequence level) is in 4: 4: 4 aspect ratio.

FIG. 6 illustrates a mode of implementation of a coding of several components of a block of the image. The coding is carried out here component by component. All the blocks of the same component are coded before coding the blocks of another component. This mode of implementation is particularly suitable if the components do not have the same partitioning.

In a first step E400, a first component is considered, for example the luminance component. Then for this selected luminance component, we consider the first block, E401. This is coded according to a method chosen in step E402. For example, the method may include evaluating the BDPCM mode and another mode, then selecting the code block that results in the lowest bit rate-distortion cost. During a T2 test, it is verified that the coded block is the last of the component. If this is not the case, we go to the next block in step E403 until all the blocks of the luminance component are encoded.

If the component processed is not the last, test T3, another component is obtained in step E404, the first chrominance component for example. Likewise, the first block of the new component is selected in step E401. Then this block is coded using one of the methods described above. The coding here can be a function of a predetermined criterion CT. For example, the encoding of this chroma component may include an evaluation of the BDPCM mode and another mode, if the color aspect ratio of the whole picture or video clip is 4: 4: 4.

Alternatively, if the chrominance component and the luminance component have the same partitioning, the chrominance component block may be encoded using the selected mode to encode the corresponding luminance component block. For example, if the mode selected for the luminance component block is BDPCM mode, that is used by default to encode the corresponding block of the chrominance component.

As before, if the block of the chrominance component processed is not the last, test T2, we go to the next block E403. The process is repeated until all the blocks of the last component are processed, test T3.

The process can be repeated for all the images making up the video sequence.

FIG. 7 illustrates another embodiment for the coding of several components of a block of the image. The coding is carried out here block by block. All the components of the same block are coded before coding the next block. For this mode of implementation, the components preferably have the same partitioning into blocks.

During a step E500, a first block is obtained (for example, located at the top left of the image). The order of travel of the blocks is assumed to be known here. The luminance component of this block is encoded in step E501. Then the first chrominance component of this block is in turn encoded in step E502 according to one of the methods described above, possibly according to a CT criterion.

The second chrominance component is encoded in step E503 according to one of the methods described above, possibly according to a CT criterion. As a variant, the encoding mode used for the first chrominance component of the block is used by default for the encoding of the second chrominance component of the block.

If the processed block is not the last, test T5, we consider the next block. The process is repeated until the last block.

Alternatively, since the chrominance components and the luminance component have the same partitioning, the chrominances of the block can be encoded using the mode selected to encode the luminance of the block. For example, if the mode selected to encode the luminance component of the block is BDPCM mode, this is automatically used to encode the chrominance components of the block.

As a variant, at least two components of the block considered are coded in parallel.

The process can be repeated for all the images making up the video sequence.

The coding according to the embodiments of FIGS. 6 and 7 in particular, can comprise, for each image or for the whole of the video sequence, the coding of information indicating whether the criterion CT is satisfied or not. This information to be an indicator of the color format of the image or (all) images in the video sequence. For example, the indicator can signal that the input image is in 4: 2: 0 format (CT criterion not satisfied) or 4: 4: 4 format (CT criterion satisfied). As a variant, the embodiments of FIGS. 6 and 7 can comprise, for at least one image, the coding of a first item of information, indicating that the BDPCM mode can be used to encode at least one block of at least one component. of the image. And if the first information indicates that the BDPCM mode can be used to encode at least one block of at least one component of the image, an encoding of a second information relating to at least said image, indicating that the BDPCM mode can be used to encode at least one block of the chrominance component of the image.

The first and second information can relate to each frame of a video sequence for better dynamics, or to the entire video sequence.

If the second information is activated indicating that the BDPCM mode can be used on at least one block of the chrominance component, a flag can be encoded for the block in question. This indicator, when activated, indicates that BDPCM mode is used for the block in question.

FIG. 10 illustrates an example of implementation of this first and this second information at image level. Four indicators FL1, FL2, FL3 and FL4 are considered here. When the FL1 flag is on (= 1), in this case BDPCM mode can be used for at least one component of the image. When the FL2 flag is on (= 1), in this case the BDPCM mode can be used for at least one block of a chrominance component of the image. If the FL2 flag is on, for at least one block of the chrominance component, an FL4 flag signals (= 1) or not (= 0) that BDPCM mode is being used to encode the block in question. Another flag, FL3, can be used to signal the use of another encoding mode. Alternatively, the other encoding mode can be a default mode. In this case, no indicator is used.

For example, for a first image IM1, the FL1 and FL2 flags are both turned on. The indicators FL3 and FL4 for a block of the luminance component Y of this image IM1 can then take the values 0 and 1 respectively, indicating that the BDPCM mode is used to code the block concerned. The FL3 and FL4 indicators for a block of the chrominance component here Cb of this image IM1 can take the values “0” and “1” respectively, indicating that the BDPCM mode is used to encode the block concerned.

In another example, for an IM2 image, only the FL1 flag is enabled. The FL2 flag is off. The indicators FL3 and FL4 for a block of the luminance component Y of this image IM1 can then take the values 0 and 1 respectively, indicating that the BDPCM mode is used to code the block concerned. Only one FL3 flag is encoded for the block of the chrominance component here Cb, of the image IM2, because the FL2 flag does not allow the use of the BDPCM mode to encode a block of the chrominance component. The FL3 flag is activated indicating that INTRA mode is being used to encode this block. If it was disabled, another encoding mode could be used (eg INTER mode). Alternatively, no flag is encoded for this block of the chrominance component. The encoding mode is a default mode.

Using the FL1 and FL2 flags avoids encoding (and subsequently decoding) one flag for each block of luminance and / or chrominance, when BDPCM mode is not used in the picture .

Reference is again made to FIGS. 6 and 7. The coding, according to the embodiments of these figures, can comprise, for each image or for the whole of the video sequence, the coding of information indicating that the partitioning of the components is identical or not for luminance and chrominances. This flag may signal a first partitioning for the luminance component and a second partitioning for the two chrominance components.

Furthermore, the coding of the blocks according to the embodiments of Figures 6 and 7 can include if the BDPCM mode is selected for a component, a coding of a coding direction (vertical or horizontal). This direction can be coded for the block of luminance and a single chrominance component. The direction coded for the block of the chrominance component is by default valid for the corresponding block of the second chrominance component.

The coding according to the embodiments of FIGS. 6 and 7 can further comprise the coding of a first item of information relating to the image, capable of indicating whether the BDPCM mode can be used to encode at least one component of minus one block of the image, and if the first information indicates that the BDPCM can be used to encode at least one component of at least one block of the image, an encoding of a second item of information relating to said image, capable of indicating that the BDPCM mode can be used to encode the chrominance component d 'at least one block of the image.

Alternatively, this information may include a first indicator indicating the mode used for the luminance component block and a second indicator indicating the mode used for the chrominance component block.

As a variant, this information may comprise a single indicator indicating the mode used for the block of the luminance component and the mode used for the block of the chrominance component (s).

As a variant, this information may include a first indicator indicating the mode used for the block of the luminance component and a single second indicator indicating the mode used for the block of the two chrominance components.

FIGS. 8 and 9 illustrate two embodiments of a method for decoding a stream of encoded data representative of the image or images of a video sequence, according to the invention. These figures correspond respectively to the embodiments of Figures 6 and 7.

First, reference is made to FIG. 8, where the decoding is carried out component by component. The mode of implementation can comprise beforehand a step E599 of decoding a first coded information item, relating to the image (or to the video sequence) to which the component belongs, indicating whether the BDPCM mode is used or not for coding at least one component of at least one block of the image. This first information corresponds to the FL1 indicator in figure 10.

If the first information once decoded indicates that the BDPCM mode can be used to encode at least one component of at least one block of the image, a second information relating to said image is decoded. This second piece of information, corresponding to the FL2 flag in Figure 10, indicates that BDPCM mode was used to encode the chrominance component of at least one block of the image.

As a variant, the prior step E599 may comprise a decoding of information representative of a predetermined criterion, for example the color format of the considered image (or of all the images of the video sequence). Depending on the color format obtained, the BDPCM mode may or may not have been used for encoding a chrominance component of a block, and corresponding information may be decoded for the block of this chrominance component.

Following this preliminary step E599 (which may be optional), the decoding is carried out here component by component. All the blocks of the same component are decoded before decoding the blocks of another component.

In a first step E600, the encoded data representative of a first component, for example the luminance component, is considered. Then for this luminance component, we consider the encoded data of the first block in the order of decoding, in step E601. During a step E602, it is possible to decode an item of information (for example the indicators FL3, FL4 in FIG. 10) indicating for this block of the luminance component, if it is encoded using a so-called mode. BDPCM or other mode.

Depending on the value of the information, the block of the luminance component is decoded with one or the other mode. During a T6 test, it is verified that the decoded block is not the last of the processed component. If not, we go to the next block in step E603 until all the blocks of the luminance component are decoded.

If the component processed is not the last, test T7, the encoded data of another component is obtained in step E604, the first chrominance component for example. Likewise, the coded data relating to the first block of the new component are selected in step E601. Information representative of the coding mode used for the block of the considered chrominance component can be is decoded (for example an indicator FL3 or the indicators FL3, FL4 of FIG. 10). If authorized by the second information decoded during step E599, this information representative of the encoding mode can indicate the BDPCM mode. If not authorized by the second information decoded during step E599, this information representative of the encoding mode can signal another mode as explained with reference to FIG. 10.

As a variant, if not authorized by the second information item decoded during step E599, the encoding mode used can be a default mode, for example the INTRA mode. Alternatively, if the chrominance component and the luminance component have the same partitioning, the chrominance block can be decoded using the mode that was used to encode the corresponding luminance component block.

As before, if the chrominance component block processed is not the last one, test T6, we go to the next block E603. The process is repeated until all the blocks of the last component are processed, test T7.

The process is repeated here so that all of the encoded data of the video data stream is processed.

FIG. 9 illustrates another mode of implementation for the decoding of several components of a block of the image. The decoding is done here block by block. All the components of the same block are decoded before decoding the next block. For this mode of implementation, the components preferably have the same partitioning.

As for the mode of implementation shown in FIG. 8, the method may include a preliminary step E699, similar to step E599 described above.

Following this preliminary step E699 (which may be optional), the coded data corresponding to a first block are obtained E700. The order of travel of the blocks is assumed to be known here. Information indicative of the encoding mode which was used to encode the luminance component is decoded (eg FL3, FL4). The luminance component of this block is decoded in step E701 from this decoded encoding mode information.

Then information indicative of the encoding mode which was used to encode the chrominance component of the block concerned is decoded (for example FL3, FL4 or FL3 only).

Alternatively, the information decoded during the decoding of the luminance component can also indicate the encoding mode of the chrominance component of the block considered, if for example it is encoded on two bits or then by being a default mode for all the components. of the block concerned. The first chrominance component of this block is in turn decoded in step E702 using the mode indicated by the decoded information item, this mode possibly being the BDPCM mode.

The second chrominance component is decoded in step E703 according to one of the methods described above. As a variant, the encoding mode used for the first chrominance component of the block is automatically used for the decoding of the second chrominance component of the block.

As a variant, at least two components of the block considered are decoded in parallel. The process is repeated here so that all of the encoded data of the video data stream is processed.

According to a particular embodiment of the invention illustrated by FIG. 11, the steps of the coding method are implemented by computer program instructions. For this, the coding device 800 or the decoding device 900 has the conventional architecture of a computer and includes in particular a MEM memory 801, 901, a processing unit UT 802, 902, equipped for example with at least one microprocessor P 1, and controlled by the computer program Pg 803 stored in memory MEM. The computer program Pg comprises instructions for implementing the steps of the encoding or decoding method as described above, when the program is executed by the microprocessor.

On initialization, the code instructions of the computer program Pg are for example loaded into a RAM memory before being executed by the processor. The processor of the processing unit UT notably implements the steps of the encoding or decoding method described above in these various variant embodiments, according to the instructions of the computer program Pg.

The device 800, 900 can be arranged to cooperate at least with the following modules of an AND terminal:

a data transmission / reception I / R module, by means of which a binary stream or a compressed file FC is transmitted in a telecommunications network, for example a wired network or a wireless network; and a module M for storing the image or the sequence of images to be encoded, the encoded data stream / file obtained or the image or the sequence of images decoded.

According to one embodiment, the device 800 for coding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, comprises at least one microprocessor configured to implement a coding of at least one block of the chrominance component, using a so-called BDPCM mode.

According to one embodiment, the device 800 for coding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, comprises at least one microprocessor configured to implement for at least one block of the chrominance component, an evaluation of a so-called BDPCM mode and of another mode for encoding said block, and as a function of the result of the evaluation, selection of one of the modes for coding said block, and coding of information representative of the mode selected for coding said block.

According to one embodiment, the device 900 for decoding an image encoded from a stream of encoded video data, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance , each component being partitioned into blocks, the device comprising at least one microprocessor configured to implement a decoding of information indicating for at least one block of the chrominance component, whether it is encoded or not using d 'a mode called BDPCM.

Of course, various other modifications can be made to the invention within the scope of the appended claims.

Claims

1. Method for encoding an image of a video sequence, the image comprising at least two components, including a component representative of luminance and a component representative of chrominance, each component being partitioned into blocks, the method comprising coding ( E1) of at least one block of the chrominance component, using a so-called BDPCM mode.

The method of claim 1, wherein at least one block of the chrominance component is encoded using BDPCM mode when a predetermined criterion (CT) is satisfied.

3. Method according to claim 2, wherein the criterion is an encoding criterion indicative of a choice made by an encoder capable of implementing said encoding method.

4. The method of claim 3, wherein the choice made by the encoder results from an analysis of the video sequence.

5. The method of claim 2, wherein the predetermined criterion is satisfied when said two components are of identical size.

The method of claim 1, wherein the encoding further comprises encoding information indicative of the BDPCM mode used to encode the block of the chrominance component.

7. The method of claim 2, wherein when the predetermined criterion is satisfied, an indicator, indicating the use of BDPCM mode, is encoded for the block of the chrominance component.

8. Method for encoding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the method comprising for at least one block of the chrominance component, an evaluation (E10) of a mode called BDPCM and of another mode for encoding said block, and

as a function of the result of the evaluation, a selection of one of the modes for coding said block, and a coding (E1 1, E12) of information representative of the mode selected for coding said block.

9. The method of claim 8, wherein the BDPCM mode and the other mode are evaluated for at least one block of the chrominance component when a predetermined criterion (CT) is satisfied.

10. The method of claim 9, wherein the predetermined criterion (CT) is satisfied when said two components are of identical size.

1 1. The method of claim 10, wherein the encoding further comprises encoding information indicative of the sampling format of the image, said information being indicative of the sizes of said components.

12. The method of claim 8, wherein the evaluation comprises coding of the block of the chrominance component considered according to the BDPCM mode on the one hand and according to the other mode on the other hand, and according to another. predetermined criterion, a selection of the block coded according to the BDPCM mode or of the block coded according to the other mode.

13. The method of claim 8, further comprising for at least one block of the luminance component, an evaluation of the so-called BDPCM mode and of the other mode for encoding said block of the luminance component, and as a function of the result of the evaluation, coding of information representative of one of the modes to code said block of the luminance component.

14. The method of claim 8 or 13, wherein the components have the same block partitioning, the coding of the image being performed block by block (E500- E504).

15. The method of claim 8 or 13, wherein the encoding of the image is performed component by component (E400-E404).

16. The method of claim 8, wherein the encoding further comprises for the considered image,

17. The method of claim 16, comprising for at least one block of the chrominance component, if the second information relating to said image indicates that the BDPCM mode can be used to encode at least one block of the chrominance component of the. image, the encoding of an indicator relating to at least one block of the chrominance component, indicating that the BDPCM mode is used to encode said block of the chrominance component.

18. The method of claim 9, comprising for at least one block of the chrominance component, if said criterion is satisfied, the encoding of an indicator relating to said block, indicating that the BDPCM mode is used to encode said block of the component. chroma.

19. Coding method according to claim 8, wherein the image further comprises a second chrominance component having the same partitioning as the other chrominance component, and if the BDPCM mode is selected for the block of the other component. of chrominance, said information relating to the block of the other chrominance component comprises an indicator indicating that the BDPCM mode is selected for the block of the other chrominance component and the corresponding block of the second chrominance component, and an indicator of direction, to determine a block predictor for the block of the other chrominance component and the corresponding block of the second chrominance component.

20. The method of claim 1 or 8, wherein a block is formed of elements, the coding of a block using the BDPCM mode comprises for at least one element of the block, the coding of a quantized value d. 'a residue, representative of a difference between a first and a second value, the first value corresponding to a quantified value of a difference between the value of the element considered and a reference element whose position is indicated by the direction data, and the second value corresponding to the quantized value of a difference between said other element and another reference element indicated by the given direction.

21. Method for encoding an image of a video sequence, the image comprising at least two components, including a component representative of luminance and a component representative of chrominance, the luminance and chrominance components having the same partitioning into blocks , the method comprising for at minus one block of the luminance component, an evaluation of a mode called BDPCM and another mode for encoding said block, and

22. Coding method according to claim 21, wherein if the BDPCM mode is selected, said information relates to the block of the luminance component, and comprises an indicator indicating that the mode selected for the block of the luminance component and for the block. corresponding block of the chrominance component is the BDPCM mode and a direction indicator for determining a block predictor for the block of the luminance component and for the corresponding block of the chrominance component.

23. A method of decoding an encoded image, from a stream of encoded video data, the image comprising at least two components, including a component representative of luminance and a component representative of chrominance, each component being partitioned. in blocks, the method comprising a decoding (E601, E701) of an item of information indicating, for at least one block of the chrominance component, whether or not it is encoded using a mode called BDPCM.

24. The method of claim 23, wherein the image further comprises a second chrominance component having the same partitioning as said chrominance component, the decoded value of said information also being used to indicate whether the corresponding block of the second component chrominance is encoded using a so-called BDPCM mode.

25. The method of claim 24, wherein said decoded value of said information also indicates a direction, allowing from respective block predictors, a function of said direction, to decode the block of said chrominance component and the corresponding block of the. second chrominance component.

26. A method of decoding an encoded image, from a stream of encoded video data, the image comprising at least two components, including a component representative of luminance and a component representative of chrominance, each component being partitioned into blocks, the method comprising a decoding of a first piece of information representative of an encoding criterion, and if the decoded value of the information indicates that the encoding criterion is satisfied, a decoding of a second information indicating for at least one block of the chrominance component, whether or not it is encoded using a so-called BDPCM mode.

27. The method of claim 26, wherein the image further comprises a second chrominance component having the same partitioning as the other chrominance component, and wherein the decoded value of the second information is also used to indicate whether the corresponding block of the second chrominance component is encoded using a so-called BDPCM mode.

28. The method of claim 27, wherein said decoded value of the second information also indicates a direction, allowing from respective block predictors, a function of said direction, to decode the block of the other chrominance component and the block. corresponding to the second chrominance component.

29. The method of claim 26, further comprising decoding other information indicating for at least one block of the chrominance component, whether it is encoded using a so-called BDPCM mode.

30. The method of claim 26, wherein the encoding criterion is indicative of a choice made by an encoder having encoded said video data stream.

31. A method according to claim 30, wherein the choice made by the encoder results from an analysis of the video data stream.

32. A method of decoding an encoded image, from a stream of encoded video data, the image comprising at least two components, including a component representative of luminance and a component representative of chrominance, each component being partitioned in blocks, the method comprising a decoding of a first item of information (FL1) relating to at least said image, indicating that the BDPCM mode can be used to decode at least one block of at least one component of the image, and if the first information indicates that the BDPCM mode can be used to decode at least one block of at least one component of the image, a decoding of a second information (FL2) relating to at least said image, indicating that the BDPCM mode can be used to decode at least one block of the chrominance component of the image.

33. The method of claim 32, comprising, if the second information relating to said image indicates that the BDPCM mode can be used to decode at least one block of the chrominance component of the image, decoding a flag (FL4 ) relating to said at least one block of the chrominance component, indicating that the BDPCM mode is used to decode said block of the chrominance component.

34. A method of decoding a coded image, from a stream of coded video data, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance having the same partitioning in blocks, the method comprising a decoding of a first item of information representative of an encoding criterion, and if the decoded value of the item of information indicates that the encoding criterion is satisfied, a decoding of a second item of information relating to a minus one block of the luminance component indicating whether the block of the luminance component and the corresponding block of the chrominance component are encoded or not using a so-called BDPCM mode.

35. Device (800) for encoding an image of a video sequence, the image comprising at least two components, including a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the device comprising at least one microprocessor configured to implement coding of at least one block of the chrominance component, using a so-called BDPCM mode.

36. Device (900) for decoding an encoded image from an encoded video data stream, the image comprising at least two components, of which a component representative of the luminance and a component representative of the chrominance, each component being partitioned into blocks, the device comprising at least one microprocessor configured to implement a decoding of information indicating for at least one block of the chrominance component, whether or not it is encoded using a mode says BDPCM.