EP3991436A1 - Method for encoding and decoding video by signalling of a candidate sub-assembly - Google Patents

Abstract

The invention relates to a method for decoding a bitstream comprising a sequence of encoded images, the method comprising: - obtaining first information from first encoded data contained in the bitstream; - determining a list of coding parameters on the basis of the first information and at least one predefined list of coding parameters; - obtaining second information from second encoded data contained in the bitstream, the second information corresponding to a coding parameter among the coding parameters from the list of coding parameters; - determining an element of a restored image corresponding to an element of an image from the sequence of images by means of the coding parameter corresponding to the second information and encoded data that represents the element from the image in the sequence; wherein the size of the second encoded data depends on a number of coding parameters in the list of coding parameters.

Description

Description

Title: Video encoding and decoding method by signaling a candidate subset

Field and context of the invention

The present description relates to the field of video compression, and more specifically to video encoders and decoders.

[0002] The description can be applied in particular to the protocols for selecting prediction modes in MPEG video compression standards, such as for example AVC (h.264) for "Advanced Video Coding", HEVC (h.265) for “High Efficiency Video Coding”, and future MPEG compression standards, such as VVC (h.266) for “Versatile Video Coding”.

[0003] Conventional video compression schemes are based on the same principles since the first generation of the MPEG standard, MPEG2. The images in a video sequence or video stream are viewed sequentially and divided into parts, which in turn can be divided into parts. These elements can be, for example, macroblocks (MB), “coding tree units” (CTU) or else blocks, “coding units” (CU) or “prediction units” (PU).

[0004] The division of images into elements or / and pixel elements, during the encoding processes of a video sequence, allows the image data to be processed locally. Different types of processing associated with encoding parameters can be used to encode each element.

[0005] The different types of processing can generally be used to improve, for example, the video compression process during video encoding. One type of processing can be, for example, an intra prediction comprising intra prediction modes as coding parameters. According to another example, the type of processing may be the application of an adaptive sample shift filter comprising shift or offset values as encoding parameters. According to another example, the type of treatment may be a inter prediction comprising modes known as “merge modes” as coding parameters.

[0006] Depending on the compression standards and the types of encoding parameters, a number of encoding parameters are accessible for encoding and decoding. This number of coding parameters can, depending on the type and the compression standard, be relatively large and thus lead to a consequent signaling which can reduce the compression performance accordingly in the event of sub-optimal or incomplete use of the assembly. coding parameters Incomplete use can occur during precise implementation of an encoder with strong constraints in terms of speed of execution or complexity, leading the manufacturer not to use all the parameters. Suboptimal use of encoding parameters may occur depending on specific content and / or encoder point of use.

Generally, the evolution of a compression standard, such as the MPEG-type compression standard, for example, implies a compression gain, but also an increase in the number of coding parameters available for the same type of processing. For example, the intra prediction modes which are 35 for the HEVC standard, and 67 for the future VVC compression standard.

[0008] The encoding parameters used during video encoding are necessary during decoding of the corresponding binary stream. The encoder must thus signal them to the decoder in the binary stream, but in return for a certain signaling cost. This signaling can represent a certain number of bits which affects the performance of the compression, in particular if it is not compensated for a sufficient gain in compression quality. The signaling can consist, for example, of signaling a coding parameter. For example, when the encoding parameters are intra prediction modes in the HEVC standard, the signaling of the prediction mode selected by the encoder for an element represents up to 5 bits. That is to say that the encoder encodes the index value corresponding to this intra prediction mode among 35 modes of prediction. In the case of the future WC compression standard, this signaling is carried out on up to 6 bits to cover the 67 prediction modes.

[0009] Thus the cost of overall signaling can quickly become significant. This signaling is all the more important as the number of available encoding parameters is large and the number of elements to be encoded is large.

The present invention improves the situation.

To this end, a method is proposed for decoding a video binary stream corresponding to a sequence of images, the method comprising:

- Obtain first information from first encoded data contained in the binary stream;

- determining a list of encoding parameters based on the first information and at least one predefined list of encoding parameters;

obtaining second information from second encoded data contained in the binary stream, the second information corresponding to an encoding parameter from among the encoding parameters of the list of encoding parameters;

determining an element of a restored image corresponding to an element of an image of the sequence of images by means of said coding parameter corresponding to the second information and of encoded data representative of the element of the image of the sequence ;

wherein the size of the second encoded data depends on a number of encoding parameters in the encoding parameter list.

This makes it possible to decode a sequence of images encoded in a binary stream on the basis of a list of encoding parameters included in the predefined list of encoding parameters. Thus, the number of encoding parameters available for encoding and decoding can be reduced. It is therefore possible to reduce the size of the signaling assigned to the encoding parameters, that is to say the second encoded data. Advantageously, the list of parameters is strictly included in the predefined list, thus making it possible to reduce the size of the signaling whatever the element to be decoded. [0013] In addition, advantageously, this reduction in the number of upstream encoding parameters can make it possible to reduce the encoding or decoding calculation time and / or reduce the hardware resources required, in particular for example when the selection of an encoding parameter from a list of encoding parameters requires testing all of them in order to select the optimal encoding parameter.

By coding parameter is meant a parameter on the basis of which the encoded data representative of an element are obtained. In other words, by encoding parameter is meant a function or a parameter of a function making it possible to obtain the encoded data representative of an element. For example, an encoding parameter can be used to determine a predictive element of which the differential with the element to be encoded is the residue. Such an encoding parameter can thus be a prediction mode, i.e. the encoding parameter can indicate an element of an image of the image sequence (either of the same image or of an image different than that of the element which is to be encoded / decoded). The encoding parameter may also allow the calculation of a residual after the predictive element (ie indicated by a prediction mode) of the element to be encoded / decoded has been determined.

[0015] By an element of the image is meant a processing unit. These processing units divide or split the image into groups of pixels. This element can represent a set of pixels or correspond to a single pixel, an image being divided into elements which are then processed for encoding or decoding. For example, the elements can be macroblocks, CTUs (coding tree units), or CTBs (coding block units). Each element can be further divided into smaller fixed size 4x4, 8x8, 16x16, 32x32, 64x64, or 128x128 pixel elements or rectangular elements. These elements of smaller size can be for example blocks, CU (coding unit), CB (coding block), “prediction unit” (PU), “predict block” (PB), “transform unit” (TU), or even “transform blocks” (TB).

By obtaining first information from first data encoded in the binary stream is meant the decoding of the first encoded data comprising the first information. This first information can therefore be signaled in the binary stream by the encoder through the first encoded data. This first information can comprise, for example, information relating to a mode of management of coding parameters (indication of a mode of management of the coding parameters, parameterization of the mode of management used for the encoding, etc.). The term “coding parameter management mode” is understood to mean a mode defining the way in which the coding parameter lists are determined on the basis of the first information (for example coding characteristics) and of the predefined list of coding parameters.

[0017] Several management modes can be indicated in the first information in order to determine several lists of encoding parameters from the same predefined list of encoding parameters, thus making it possible to use the most suitable list for encoding or the decoding of each element of an image according to the different needs (for example characteristic of the image, characteristic of the element, etc.).

A predefined list of encoding parameters is understood to mean a list comprising a finite number of encoding parameters. This number may depend on the compression standard used for encoding the video sequence, and the type of encoding parameters. For example prediction modes or offset values of a so-called "offset" filter. The type of encoding parameters can be determined according to the first information. The predefined list of encoding parameters may include an indexing of the encoding parameters, for example from 0 to N-1 with N a positive integer.

[0019] By way of example, if the encoding parameters are intra prediction modes in the HEVC compression standard, the predefined list includes 35 encoding parameters or intra prediction modes. According to another example, in the case of the future VVC compression standard, the predefined list of coding parameters would then include 67 intra prediction modes for an element, for example CUs.

[0020] According to another example, if the encoding parameters are offset values used during the “sample adaptive offset” (SAO) filtering for pixels encoded on 10 bits in the HEVC standard or future VVC, the predefined list of encoding parameters can then include 32 offset values for an element, for example a CTU.

It can be understood by determining a list of encoding parameters, determining at least one list of encoding parameters strictly included in at least one predefined list of encoding parameters. The list of coding parameters can include an indexing of the coding parameters, for example from 0 to i, with i positive integer strictly less than N-1.

By obtaining second information from second data encoded in the binary stream is meant the decoding of second encoded data comprising second information. This second information can therefore be signaled in the binary stream by the encoder through the second encoded data. This second information makes it possible to indicate to the decoder the encoding parameter to be used from among the encoding parameters of the list determined for the determination of an element in the restored image.

Since the encoder and the decoder have the same list of determined encoding parameters, when the encoder signals the encoding parameter to be used for decoding from the list of encoding parameters, the decoder can find this encoding parameter from the same list of encoding parameters.

By determining an element of a restored image corresponding to an element of an image of the sequence of images (or simply determining an element of an image) is meant the fact of reconstituting the pixels contained in this element of the image by applying the encoding parameter indicated in the second item of information. The encoding parameter relates to a type of processing, for example a prediction, intra or inter, or the application of an ODS filter. Thus, the determination of the element by a specific type of processing consists in applying an encoding parameter of this type.

It can be understood by size of the second encoded data a size which can be defined by a number of bits. The second encoded data may correspond to an index of an encoding parameter signaled in the bit stream by the encoder. This index being that of one of the coding parameters included in the indexed list of coding parameters determined by the decoder. The number of bits used to encode the index is a function of the number of encoding parameters in the determined indexed list.

Thus during the decoding of an image of a sequence of images, that is to say of the decoding of the elements of this image, for each element can be signaled by the second information, the index of an encoding parameter to be used from among a reduced number of encoding parameters included in the list of encoding parameters determined by the decoder. The signaling cost in bits can thus be reduced.

Advantageously, the list of encoding parameters can be determined by selecting encoding parameters from the predefined list of encoding parameters or by selecting from among a plurality of predefined lists of encoding parameters.

Thus, either the list is determined by selection of the coding parameters within the predefined list or the selection is carried out upstream of obtaining the first information and therefore independently thereof, that is to say say that several sub-lists of coding parameters are predetermined, the determination of the list then consists in selecting one of these predetermined sub-lists. The predetermined coding parameter sublists are sublists of the predetermined list. This helps reduce the size of the first information needed to determine the list.

The selection of the encoding parameters within the predefined list makes it possible to adapt the number of encoding parameters which can be used according to the image and when several lists are determined by selection from the predefined list then it is possible to adapt the number of coding parameters according to the elements. Thus, two elements of the same image can be encoded / decoded with determined lists of different size. For example, the elements of an image encoded / decoded in intra of an intra-refresh zone (within the framework of an intra-refresh encoding) could be encoded / decoded by means of a list comprising more parameters of encoding that the elements of the image encoded / decoded intra outside the intra-refresh area. According to one or more embodiments, the first encoded data and / or the second encoded data may be included in a signaling part of the binary stream. The signaling part can make it possible to simplify the decoding of a video sequence by providing information for example on the structure of the images, the type of image, the encoding resolution, etc. The general signaling part of the HEVC standard consists of a "Video Parameter Set" (VPS) field (also called NAL), a "Sequence Parameter Set" (SPS) field, and a "Picture Parameter Set" field ”(PPS). The future WC standard would include at least the SPS and PPS fields.

The VPS field includes information relating to the video. The VPS field includes, for example, information relating to the “Profile”, the “Level” and the “Tier” which define decoding properties.

The SPS field includes information relating to the sequence of images considered. All the images in the sequence normally use the same SPS. The SPS field includes in particular information relating to the coding tools used, or else important parameters describing the properties of the encoded sequence (for example the size of the images).

The PPS field includes information relating to each image, even if several images can refer to the same PPS in order to reduce the size of the signage. The PPS notably includes additional information relating to the encoding tools used in the image (s) which refer to it (for example the "intra prediction mode"). It is for example, in this PPS field, that the first and second information can be signaled.

[0034] According to another example, the first and second information can be indicated in the header of the slices or "slices" of the image of the sequence of images.

In addition, the first encoded data and / or the second encoded data can be signaled according to a signaling mechanism specific to the type of coding parameter. Advantageously, the list of encoding parameters can be a function of encoding characteristics of elements of the image.

[0037] Thus, it is possible to use different determined lists to decode elements of the image according to the specificities of these elements. Thus, a first list can be used to encode / decode a first element and a second list to encode / decode a second element of the same image. For this it is possible to determine for a given image (or a group of given images), on the basis of the first information, a plurality of lists of encoding parameters from which a list is selected as a function of the encoding characteristics of the element to decode / encode from this given image. The determined lists are obtained before proceeding with the decoding / encoding of the elements of the image to be decoded, that is to say that the determined lists are obtained before the decoding of a set of elements of the image. When encoding / decoding an element, a list from among the plurality of lists determined for the image is selected as a function of the encoding characteristics of this element, these encoding characteristics are specific to the element to be encoded / decoded . The first information can include the encoding characteristics of the elements. In both cases, it is thus possible to obtain great freedom in determining the list or lists of encoding parameters to adapt to the specificities of each element to be encoded / decoded. Advantageously, at least two lists are determined and at least two elements of the image are respectively encoded / decoded with different determined lists.

[0038] The encoding characteristics of elements of the image are specific to the element to be encoded / decoded. They can thus depend only on this element, namely that if we change one of the elements of the image to be encoded / decoded we do not change the encoding characteristics of the element to be encoded / decoded.

Advantageously, the encoding characteristics of elements of the image comprise an element size and / or a quantization step (QP) of element and / or a chromatic type of element and / or a type. element brightness and / or a type of intra element encoding and / or type of inter element encoding and / or size of transform and / or type of transform and / or type of filter.

[0040] It can be understood by an element size a size defined by a number of pixels, for example 64x64 pixels.

By a chromatic type of element and a type of luminosity of element, one understands a component of luminosity called "luma", Y, and a color component called "chroma", U or V, of the pixels of the. element.

By type of intra element encoding is meant that the element of the image I is encoded intra that is to say that the element is not encoded from an element another image than image I.

By type of inter element encoding is meant that the element of the image I is encoded in Inter, that is to say that the element is encoded at least from an element d 'another image than image I.

By a first element encoded from a second element is meant that the prediction mode (Intra, Inter, etc.) used to encode the first element and which is signaled in the encoded data representative of the first element is aimed at or indicates to use the second element (that is, the values of the second element once decoded) to encode the first element.

It can be understood by a transform size, a discrete cosine transform (DCT) size of 4x4, 8x8, 16x16, 32x32 or 64x64 pixels.

It can be understood by a type of transform, a DCT-II type or a DST-VII type or a DCT-VIII type.

It can be understood by type of filter, a sample adaptive offset filter (SAO) of the type "contour" ("Edge") or "Band" ("Band"), a filter. Unblocking “Deblocking Filter” or an adaptive loop filter “Adaptive Loop Filtering, ALF”.

Advantageously, the determination of the list of coding parameters can be done by selecting from the predefined list of coding parameters coding parameters specified by the first information. Thus, the first information obtained by decoding the first encoded data comprises information explicitly indicating the encoding parameters to be selected. For example, the first information comprises a list of indices corresponding to the coding parameters of the predefined list or lists when the latter are indexed, for example from 0 to N-1, N being a positive integer. This is a so-called “explicit” management mode. Thus in the so-called “explicit” management mode, instructions can explicitly indicate the encoding parameters that the decoder can select from among one or more predefined lists of encoding parameters making it possible to constitute for example a first list of encoding parameters.

Advantageously, the instructions can indicate several sets of coding parameters for determining several lists of coding parameters.

Advantageously, the predefined list of encoding parameters can be indexed from 0 to N-1, with N a number of encoding parameters in the predefined list of encoding parameters and in which the determination of the list of encoding parameters can be done by selection on the basis of the first information:

- coding parameters of the predefined list of indices k + nm, with n strictly positive integer and k positive or zero integer and n + k less than or equal to N-1 and m between 0 and the default integer part of (Nk-1) / n; and,

- coding parameters from the predefined list of indices strictly less than k ’with k’ positive or zero integer less than or equal to k.

Thus, it is possible to determine one or more lists of coding parameters in a simple manner and with a small signaling size. Thus, it may be possible to transmit a minimum of instructions in the first data to limit the cost of the overall signaling in the bit stream. In addition, such a selection from the predefined list is regular, that is to say the selected coding parameters are distributed regularly in the predefined list. Thus, if the encoding parameter (pi) of the predefined list that the encoder / decoder would use to encode / decode the element has not been selected in the determined list used to encode / decode the element of the image then the index difference between the encoding parameter (pi) and the encoding parameter (p ₂ ) of the determined list actually used by the encoder / decoder to encode / decode the element remains less than n / 2 and in any case less than n. Now, due to the ordering of the coding parameters, the difference between residues obtained with two coding parameters of close indices is on average smaller than the difference between residues obtained with two coding parameters of more distant indices. Thus, with a list determined by such a selection, we obtain, on average, residues that are closer to those which would have been obtained using the predefined list. Thus, the amount of encoded data necessary to encode the residues of the elements of the image is reduced while reducing the signaling required to signal each encoding parameter. Moreover, this type of selection makes it possible to obtain determined lists whose structure / construction is similar to that of the predefined list, which makes it possible not to stray too far from the construction of the predefined list. Thus, the CABAC arithmetic encoder will be less impacted.

The encoding parameters of index k + n.m are selected from the predefined list, with m ranging from 0 to the default integer part of (N-k-1) / n. The term k can make it possible to define an offset in the selection of the coding parameters.

The management mode allowing such a selection is called management mode by "decimation". The positive integers k and n, as well as information indicating that the "decimation" management mode is activated, are sufficient for the decoder to be able to determine a list of coding parameters according to this management mode. Thus, the first information can include the integers k, n. The integer n represents a decimation factor for selecting an encoding parameter from a predefined list of encoding parameters for every n encoding parameters.

[0055] According to one or more embodiments, n and m can be determined based on a hierarchical level of the image relative to the images of the sequence of images.

By hierarchical level of the image with respect to the images of the sequence of images or more generally of a grouping of images or group of images (GOP), it is understood that in the case where the images of 'a set of images use previously encoded images, hierarchical level 0 is that whose images serve as references to images subsequently encoded and which itself only use intra predictions or images of hierarchical level 0; hierarchical level 1 can use images of hierarchical level 0 or 1, and so on. Bottom-level images are often not referenced by any other image.

Thus, the lists of encoding parameters used to encode / decode each image can therefore include a number of different encoding parameters depending on the hierarchical level of the images.

Advantageously, the determination of the list of coding parameters can be done by selection (on the basis of the first information) of the coding parameters from the predefined list according to the levels of use frequencies of the coding parameters of the predefined list.

[0059] The frequency of use of an encoding parameter is understood to mean the number of times this encoding parameter has been used for the determination of elements of previously rendered images. Specifically, the number of times this encoding parameter has been used for decoding elements of the image being decoded and / or previous images.

Such a selection from the predefined list makes it possible to select the coding parameters most likely to be used with regard to the preceding images. This is relevant as well when one encodes / decodes elements in intra as in inter. This is because spatial predictions with high usage frequency levels in a previous frame are more likely to have high usage frequency levels in the image being encoded / decoded. Thus, the encoding parameters of the predefined list that the encoder / decoder would use to encode / decode the elements of the image being processed are more likely to be selected from the determined lists. Thus, with a list determined by such a selection, we obtain, on average, residues that are closer to those which would have been obtained using the predefined list. Thus, the amount of encoded data necessary to encode the residues of the elements of the image is reduced while reducing the signaling required to signal each encoding parameter. By way of example, the selection from the predefined list as a function of the levels of use frequencies can be done by selecting the coding parameters with a level of use frequencies greater than a threshold or by selecting a number. encoding parameter preset having the highest frequency levels of use among the encoding parameters of the preset list.

The management mode allowing such a selection is called "histogram" management mode. The threshold or the predefined number of encoding parameters having the highest usage frequency levels may be included in the first information. Thus, the information included in the first information is information indicating that the “histogram” management mode is activated as well as the associated threshold or number. Thus, this selection mode requires little signaling.

Advantageously, the predefined list of coding parameters can be indexed and the determination of the list of coding parameters can be done by selection on the basis of the first information:

- primary encoding parameters specified by the first information; and

- secondary encoding parameters such as an index of each secondary encoding parameter is included in \ J _k [i (p) _k - m _¾ ; i (p) _k + m _¾ ] where i (p) _k are the indices of the primary encoding parameters and p _k are strictly positive integers specified by the first information.

Thus, it is possible to determine one or more lists of coding parameters with little signaling. Indeed, it is possible to transmit a minimum of instructions in the first data, and limit the overall signaling cost in the bit stream.

Some coding parameters (the primary coding parameters) are explicitly indicated in the first information. The selection of the other encoding parameters (secondary encoding parameters) is done by taking the index parameters included in the neighborhood of the primary encoding parameters. For example the parameters of index [2 (r), _¾ - mk; 2 (p), _¾ + m *] included in the neighborhood of the k th primary encoding parameter. The management mode allowing such a selection is called "group" management mode. Thus, the information that can be included in the first information to implement such a selection is information indicating that the “group” management mode is activated as well as the indices of the primary coding parameters and the associated p _k values. Thus, this selection mode requires little signaling. To further reduce signaling, it is possible to give the same value to all p _k .

Advantageously, the predefined list of coding parameters or the plurality of predefined lists are predefined lists of prediction modes or predefined lists of offset values of a filter.

It can be understood by prediction modes, intra or inter modes making it possible to determine an element (in other words to determine the pixels of the element) of an image from other elements belonging to the same image for elements encoded in intra and from elements of other images for elements encoded in inter.

The term offset value is understood to mean a value that can be added to each pixel of an element resulting from an intra or inter prediction in order to correct the errors that have arisen during the encoding (in particular those relating to the quantization) .

[0070] A second aspect of the invention relates to a method for encoding a video binary stream corresponding to a sequence of images which may comprise:

- for an image of the sequence of images, determine an encoding setting;

- determine a list of encoding parameters based on the encoding setting and at least one predefined list of encoding parameters;

get video bitstream including:

- encoded data representative of first information, the first information indicating the encoding setting;

- encoded data representative of an element of the image obtained by means of an encoding parameter of the list of determined encoding parameters, the encoded data representative of the element comprises encoded data representative of second information, the size Datas encodings representative of the second information depends on a number of encoding parameters in the list of encoding parameters.

As previously indicated, the list of encoding parameters may include a small number of encoding parameters making it possible to reduce the size of the signaling of the encoding parameters and to reduce the amount of calculation required to encode the elements of the image.

[0072] By encoding setting, it is meant encoding input data. For example, the encoding settings can be:

- instructions specifying the management modes to be used,

- the encoding characteristics of elements, and / or

- the management modes to be used according to the encoding characteristics.

By video bit stream is meant the bit stream resulting from the encoding of the sequence of images.

By encoded data representative of one or more elements, it is meant the data included in a binary sequence and therefore in the video binary stream which are obtained by encoding of said element or elements. This is data relating to an element of the image resulting from the video processing applied to the sequence of images. This encoded data contains the information on the basis of which the decoder will decode the sequence of encoded images to render the element of the image.

These encoded data are obtained from the element they represent (for example, to determine a residue), but also from other elements either of the image (Intra prediction) or of other images of the video stream (Inter prediction). In other words, the encoded data representative of a first element are obtained from a second element when the prediction mode (or more generally the encoding parameter) used to encode the first element and which is signaled in the representative encoded data of the first element refers to or indicates to use the second element (that is, the values of the second element once decoded) to encode the first element.

The encoded data can include the second encoded information. The second data may correspond to the signaling of the encoding parameter to be used for determining the element of the picture to be rendered during decoding. For example, the second encoded data may comprise the index of the encoding parameter to be used from among the indices of the list of encoding parameters, which is the same whether it is for the encoder or the decoder.

By way of example, these representative encoded data may include the index of a prediction mode as well as a residue encoded after quantization and representing the difference between the values of a predictive element and the element of l. 'picture.

[0078] A third aspect of the invention relates to a computer program comprising instructions for implementing all or part of the methods described above, when this program is executed by a processor.

A fourth aspect of the invention relates to a device for decoding a video binary stream corresponding to a sequence of images, the device comprising:

- a processor; and

- non-transient computer support comprising instructions which, when executed by the processor, configures the device to:

- Obtain first information from first encoded data contained in the binary stream;

- determining a list of encoding parameters based on the first information and at least one predefined list of encoding parameters;

obtaining second information from second encoded data contained in the binary stream, the second information corresponding to an encoding parameter from among the encoding parameters of the list of encoding parameters;

determining an element of a restored image corresponding to an element of an image of the sequence of images by means of said coding parameter corresponding to the second information and of encoded data representative of the element of the image of the sequence , and wherein the size of the second encoded data depends on a number of encoding parameters in the encoding parameter list. A fifth aspect of the invention relates to a device for encoding a video binary stream corresponding to a sequence of images, the device comprising:

- a processor; and

- non-transient computer support comprising instructions which, when executed by the processor, configures the device to:

- for an image of the sequence of images, determine an encoding setting;

- determine a list of encoding parameters based on the encoding setting and at least one predefined list of encoding parameters;

- get a video binary stream including:

- encoded data representative of first information, the first information indicating the encoding setting;

- encoded data representative of an element of the image obtained by means of an encoding parameter of the list of determined encoding parameters, the encoded data representative of the element comprises encoded data representative of second information, the size encoded data representative of the second information depends on a number of encoding parameters in the list of encoding parameters.

Brief description of the drawings

[0081] Other features, details and advantages will become apparent on reading the detailed description below, and on analyzing the accompanying drawings, in which:

[0082] [Fig. 1 a] illustrates an example of an element used for encoding or decoding an image according to one embodiment.

[0083] [Fig. 1 b] illustrates an example of the number of intra angular prediction modes available for an encoder or decoder according to an MPEG type compression standard, for example the future WC standard.

[0084] [Fig. 2] illustrates the decoding method proposed according to one or more embodiments.

[0085] [Fig. 3a] illustrates an example of a so-called “GOP” image grouping structure. [0086] [Fig. 3b] illustrates another example of a so-called “GOP” image grouping structure.

[0087] [Fig. 4a] illustrates an example of implementation of the method proposed according to one or more embodiments.

[0088] [Fig. 4b] illustrates an example of implementation of the method proposed according to one or more embodiments.

[0089] [Fig. 5] illustrates a non-limiting example of the implementation of the proposed method when the management mode is said to be “explicit”.

[0090] [Fig. 6a] illustrates a non-limiting example of implementation of the proposed method when the management mode is said to be “decimation”.

[0091] [Fig. 6b] illustrates a non-limiting example of the implementation of the method proposed when the management mode is said to be “decimation”.

[0092] [Fig. 7] illustrates non-limiting examples of implementation of the method proposed when the management mode is called “histogram”.

[0093] [Fig. 8] illustrates a non-limiting example of the implementation of the proposed method when the management mode is said to be “group”.

[0094] [Fig. 9] illustrates the encoding method proposed according to one or more embodiments.

[0095] [Fig. 10] illustrates an example of an encoding device and an example of a device for decoding a sequence of images for implementing the proposed method.

Description of embodiments

[0096] Embodiments of a computer readable medium include, but are not limited to, computer storage media and communication media, including any medium facilitating the transfer of a computer program from a location to another. By “computer storage medium (s)” is meant any physical medium that can be accessed by a computer. Examples of computer storage media include, but are not limited to, flash memory disks or components or any other memory devices. flash (e.g. USB sticks, memory sticks, memory sticks, key disks), CD-ROMs or other optical data storage devices, DVDs, magnetic disk data storage devices or others magnetic data storage devices, data memory components, RAM, ROM, EEPROM memories, smart cards, SSD (Solid State Drive) type memories, and any other form of medium usable for transporting or storing or memorizing data or data structures which can be read by a computer processor.

In addition, various forms of computer readable medium can transmit or carry instructions to a computer, such as a router, a gateway, a server, or any data transmission equipment, whether it is wired transmission (by coaxial cable, optical fiber, telephone wires, DSL cable, or Ethernet cable), wireless (by infrared, radio, cellular, microwave), or virtualized transmission equipment (virtual router, virtual gateway, end of virtual tunnel, virtual firewall). The instructions may, depending on the embodiments, include code of any computer programming language or computer program element, such as, without limitation, assembly languages, C, C ++, Visual Basic, HyperText Markup Language (HTML), Extensible Markup Language (XML), HyperText Transfer Protocol (HTTP), Hypertext Preprocessor (PHP), SQL, MySQL, Java, JavaScript, JavaScript Object Notation (JSON), Python, and bash scripting.

In addition, the terms "in particular", "for example", "example", "typically" are used in the present description to denote examples or illustrations of non-limiting embodiments, which do not necessarily correspond to preferred or advantageous embodiments over other possible aspects or embodiments.

Video data are generally the subject of source coding aimed at compressing them in order to limit the resources required for their transmission and / or their storage. There are many encoding standards, such as H.264 / AVC, H.265 / HEVC, MPEG-2, or the future WC compression standard (h.266), which can be used for this purpose. [0100] In conventional video compression schemes, the images of a video sequence or of a video stream are considered sequentially. Each image of the video sequence can be divided for example into slices (or “slice” in English). These slices can represent a portion of the image. Slices can also be divided into elements of fixed size, for example 16x16, 32x32, 64x64, or 128x128 pixels, for the future VVC standard. By way of example, the elements can be macroblocks, CTUs (coding tree unit), or even CTB (coding block unit). Each element can be further divided into smaller fixed size elements 4x4, 8x8, 16x16, 32x32, 64x64, or 128x128 pixels. These elements of smaller size can be for example blocks, CU (coding unit), CB (coding block), “prediction unit” (PU), “predict block” (PB), “transform unit” (TU), or even “transform blocks” (TB). One of the objectives of partitioning the image into elements, which themselves can be divided into elements, is to make it possible to adapt and refine the types of processing to be applied to each element of the image.

[0101] The term “type of processing” means one or more data processing operations carried out on the video data during encoding or decoding. For example, when encoding an image, the elements of the image can be processed according to the type of processing known as "intra prediction". This type of processing makes it possible to predict the elements, for example CU or CB, of an image to be encoded. Prediction aims to exploit the different correlations that exist between the pixels of the images of a video sequence. This correlation can be spatial called “intra prediction”, that is to say between pixels of the same image being encoded, or it can be temporal called “inter prediction”, that is to say between the pixels of an image being encoded and the pixels of the previously encoded images.

The neighboring pixels can be taken into account according to different configurations, and each configuration constitutes a specific coding parameter for the type of processing known as “intra prediction” for example. Thus, this encoding parameter can be a prediction mode in a certain direction.

[0103] For example, FIG. 1a illustrates an element 100, for example a CU or a PB or even an MB, treated from a type of treatment, for example a intra prediction, according to a given coding parameter, for example an intra prediction mode in a certain direction 103. According to this example, an element 100 of size 4x4 pixels can be surrounded by a set of pixels 101, 102 already encoded (respectively decoded ) belonging to neighboring elements.

[0104] FIG. 1b illustrates an example of the coding parameters available to encode the elements in intra, and in particular to encode the element 100 in the case of the future WC standard. A total of 67 encoding parameters (not fully shown in Figure 1a) may be available for the intra prediction of element 100.

[0105] From the coding parameters available for the type of processing known as intra prediction, the encoder can select the optimal coding parameter (that is to say the coding parameter making it possible to obtain the predicted element 100 closest to the element of the image), and predict all the pixels of the element 100 as a function of the pixels 101, 102. For an intra prediction according to a vertical prediction mode 103, the pixels used can be the pixels of pixel row 101.

[0106] The predicted element 100 is then compared to the original element of the image being processed to generate a residue representing the difference between the two elements. This residue can then be transformed, for example by DCT (discrete Fourier transform), then quantized before its encoding in the binary stream.

When decoding the intra-encoded element of an image to be restored, the decoder reconstructs the predicted element with the same encoding parameter used for the encoding of the element by the encoder and on the basis pixels 101 and 102 previously decoded. The residue is further added to the predicted element in order to bridge the existing differences between the predicted element and the original element. This operation can be repeated for all the elements of each frame of the video sequence.

According to another example, after the intra or inter encoding of an element, a so-called "offset" value can be added to a selection of an element, for example a CTU or CTB, during the processing. involving the application of an ODS filter. Indeed, after an intra or inter prediction followed by a quantization, the reconstructed element of the image being decoded may present distortions. To correct them, an offset value can be added to the value of each pixel of a selection of the element according to one of the SAO filter application modes (or SAO filter types). This value can be determined by the encoder during the encoding of the image of the sequence of images then signaled in the binary stream, for example at the level of the first syntactic element of each CTU for the HEVC standard or the future. WC. The signaling of this offset value can be performed against a predefined list of offset values. Depending on the compression standard, the number of values may be different. For example, in the case of the MPEG and WC standards, when the pixel encoding is carried out on 10 bits, the predefined list of offset value includes 32 offset values. Thus, the cost of signaling the selection of an offset value from 32 offset values for each element can be relatively large. In 8 bit, this list includes 8 values.

Thus, for each type of processing used during video encoding or decoding, for example for an intra prediction, an inter prediction or for the application of an SAO filter, a certain number of coding parameters can. be associated with each of these types of treatment.

[0110] During the encoding or decoding of an element for example, depending on the type of processing, it may be necessary for the encoder or the decoder to test all the encoding parameters to determine the optimal encoding parameter for a type of treatment given. The determination can be made on the basis of one or more metrics associated with this type of processing. For example, when the type of processing is an intra prediction, the metric is the distortion rate combined with the associated signaling rate within a criterion known as the rate-distortion cost. Thus, the encoder selects the encoding parameter with the lowest rate-distortion cost for the element being processed among all the available intra prediction modes.

[0111] Certain types of mechanisms, for example for MPEG-type compression standards, make it possible to reduce or eliminate the number of tests to be performed for each coding parameter. Certain other types of mechanisms make it possible to reduce the cost in signaling bits for each encoded element.

[0112] Despite the use of different types of mechanisms, these are insufficient, in particular in the context of under-use of all the coding parameters available and all the more so in view of new applications which are increasingly demanding. This is true whether the underuse is voluntary, due to a specific implementation, or related to the content and a given operating point of the encoder.

[0113] In fact, when encoding or decoding a video sequence, a compromise must be found between several aspects, such as, for example, the bit rate, the quality (distortion), the encoding time, and the consumption. of energy. This search for a compromise can depend on the application, for example the real-time encoding of a video stream (sport or live TV broadcast for example), or on the contrary of the encoding or decoding of a sequence. video intended to be stored on external memories, or for example according to a constraint of low consumption.

[0114] For example, from a bit rate-distortion point of view or overall bit rate cost, having a high number of coding parameters per type of processing can provide an overall gain on the efficiency of encoding or decoding, but this gain may vary depending for example on the type of image (I, P or B) or the content of the images. Thus in certain situations, a high number of coding parameters for a type of processing can become counterproductive due to the irrelevance of a large part of the coding parameters that can be used for a given image with regard to the additional cost of encoding linked to the signaling of an encoding parameter among more encoding parameters.

From an encoding / decoding time point of view and consumption of material resources with regard to the bit rate-quality compromise, the high number of encoding parameters for a given type of processing can become a very strong constraint for the real-time encoder and decoder implementation for example. [0116] FIG. 2 illustrates the method proposed according to one or more embodiments. This method can be implemented by any device comprising a processor and a memory, and configured for its implementation, such as for example, a video decoder, or any equipment configured for the decoding of a video sequence.

[0117] In step 11, the device 500 obtains first information from first encoded data contained in the binary stream. This binary stream can be encoded according for example to MPEG type standards, such as the HEVC standard or the future WC standard. For example, the first information can be signaled in the signaling part of the bit stream, for example the PPS. The first information encoded in the first data can be instructions intended for the decoder defining the management modes allowing the determination of lists of coding parameters according to different selection rules. The first data can be signaled at the level of the signaling part of the binary stream, for example the PPS, or the “slice headers”. This PPS field can contain initial information relating to one or more encoded images. Attaching multiple images to a single PPS containing the use of at least one list management mode can reduce overall signaling cost.

By way of example, the decoder can, on decoding the first encoded data, obtain an instruction for activating the method, for example “intra_adaptive_subset_enabled_flag”, and on decoding the instruction “intra_adaptive_subset_mode” determine the management mode to be used. . This management mode to be used can be the one that the encoder used to encode the elements of one or more images of the sequence of images.

[0119] According to one example, the instructions

"Intra_adaptive_subset_enabled_flag = 1" and "intra_adaptive_subset_mode = 2", indicate by the value 1 to the decoder to implement the process, and to use the management mode number 2 by the value 2.

[0120] In step 12, the device 500 determines a list of encoding parameters on the basis of the first information and at least one predefined list of encoding parameters. From, for example, the instructions defined by the first information encoded in the first encoded data which can be included in the signaling part of the binary stream, the decoder can determine, according to a management mode, a list of parameters. encoding selected from a predefined list of encoding parameters.

[0122] For example, the predefined list of encoding parameters can be a list of prediction modes (intra or inter). In another example, the predefined list of encoding parameters may be an "offset" list.

[0123] In step 13, the device obtains second information from second encoded data contained in the binary stream, the second information corresponding to an encoding parameter from among the encoding parameters of the list of encoding parameters.

[0124] Thus, the list of encoding parameters being able to be included strictly in the predefined list of encoding parameters, the list of encoding parameters may contain a reduced number of encoding parameters. The signaling of an encoding parameter, selected by the encoder, from the determined list of encoding parameters can allow the second data corresponding to the encoding of this signaling to be encoded on a reduced number of bits.

In step 14, the device determines an element of a restored image corresponding to an element of an image of the sequence of images by means of said encoding parameter corresponding to the second information item and of encoded data representative of the image element of the sequence.

[0126] Thus, the determination of an element of a restored image performed with an encoding parameter contained in a restricted list of encoding parameters can allow optimization of the signaling cost, but while maintaining relevant encoding parameters in one or more lists of encoding parameters.

The use of one or more coding parameter lists for determining elements of a restored image can be conditioned by the element coding characteristics of the image. Characteristics encoding can be the size of the elements, the type of intra and / or inter encoding, the quantization step, etc. The encoding characteristics relating to the elements of one or more images can be included in the first information. When encoding the elements of the image, the encoder can encode the selected encoding parameter allowing the determination of the element of the image reconstructed by the decoder,

[0128] When decoding the first data, the decoder can obtain the encoding characteristics of the picture elements, and thus determine the lists of encoding parameters specific to these encoding characteristics of the picture elements. Depending on their encoding characteristics, the encoded picture elements can be determined based on one or more encoding parameter lists meeting those characteristics.

[0129] For example, if the encoding characteristic is the size in pixels of the element. When encoding an image, a first determined list of encoding parameters can be used to encode an element having a size less than or equal to a first value. A second determined list of encoding parameter can be used to encode an element having an element size greater than this first value.

[0130] It may also be possible to define a range of size values for each list of encoding parameters. For example, a first list of encoding parameters can be used for image element sizes of 16x16 pixels, and the second list used for element sizes of the same image greater than 32x32 pixels.

[0131] The encoding characteristics of picture elements can include chromatic characteristics or luminosity characteristics of an element. Indeed, the pixels of each element of an image can be decomposed into two components. A first component relating to the luminosity characteristics is called “luma” and can define the luminosity of the pixel. A second component relating to the chromatic characteristics called “chroma” can define the color of the pixel, conventionally it consists of two components chroma U and chroma V. For example, for the same image, a first determined list of coding parameters can be used. exclusively for luma components, and a second determined list of encoding parameters can be used exclusively for chroma components.

Thus for an image element, the choice of an encoding parameter, for example a prediction mode, whether for the encoding or decoding of the element of this image is carried out according to a first list for chroma components and a second list for luma components.

[0133] The encoding characteristics of picture elements may include quantization steps.

[0134] The encoding characteristics of picture elements can include the type of encoding of the element or the type of encoding of the image to be encoded within a "group of picture" (GOP). . A first determined list of encoding parameters can be used exclusively for the encoding of intra-type images, that is to say for which the elements are encoded intra. A second determined list of encoding parameters can be used for encoding exclusively predictive type images (P), that is to say for which the elements can be encoded intra and inter. A third list can be used for image encoding of the bi-predictive type (B), that is to say for which the elements can be encoded intra and inter using several reference images.

[0135] According to another example, the same determined list of encoding parameters can be used for encoding the same type of encoding of picture elements not found in the same GOPs. For example, the same list of coding parameters used for two intra images located respectively in two separate GOPs.

In fact, during the encoding of a video sequence, in current compression schemes, for example of MPEG type, the images processed according to one or more types of processing associated with respective encoding parameters are encoded within a group. of images or "group of picture" (GOP). The GOP defines the order in which the images with intra (spatial) prediction and the images with inter (temporal) prediction can be arranged. GOPs are key elements in the compression of a video sequence, since they allow to group the images dependent on each other.

Each GOP can contain, depending on the type of structure of the GOP, a succession of encoded images of intra, predictive, and bidirectional type, and the first image of a GOP is generally an image of intra type in order to make it possible to start GOP decoding. When decoding a GOP of a video sequence, each frame is generated from the frames encoded in that GOP.

[0138] Figures 3a and 3b illustrate examples of two conventional structures of set of images or GOP ("group of pictures" in English) constituting a video sequence to be encoded.

The first structure is an IP type structure, that is to say an alternation of images of intra type and images of predictive type. Intra-type images are placed regularly, for example every 4 images in this case, in order to refresh the prediction and to allow access to the stream in particular. Thus, according to this example, the predictive type image 401 is predicted, according to the inter prediction, from the image data of the intra type image 400, by means of a motion compensation signaled by motion information based for example on a motion vector and a reference image.

The second GOP structure shown is an IBP structure, ie an alternation of images of intra type (I), images of predictive type (P), and images of bidirectional type (B). Thus in this type of structure, the predictive type image 411 is predicted from the image data of the intra type image 410, and the bidirectional type image 420 is predicted from the image data of the intra-type images 412 and predictive type 413, through the intermediary, as previously, of motion compensation. This type of structure has the advantage of allowing the bidirectional type images (B) to use a prediction between two images from opposite temporal directions and thus improve the efficiency of this prediction, and therefore the compression efficiency. video data. [0141] According to another example not shown, a GOP may have a pyramidal type structure, where each image of the GOP is further defined by a hierarchical level or level or temporaljd in English.

[0142] When decoding the GOP, the decoder decodes the first data in the signaling part, for example the PPS ("picture parameter set") of the binary stream, in order to obtain the first information allowing it to know the characteristics of encoding of the elements of the image of the sequence of images, for example the GOP structure comprising an image alternation of intra type, predictive type, or bidirectional type.

[0143] A first determined list of coding parameters can be used only for the encoding or decoding of one or more intra-type images, and for a size of elements greater than a defined value. A second determined list of coding parameters can be used for type P or B images.

[0144] It is possible that the same encoding parameter belongs to several lists of encoding parameters.

[0145] FIG. 4a shows a decoder implementing the method on the basis of an MPEG-type compression standard, for example HEVC or the future WC standard.

On receipt of a binary stream 506 generated by an encoder implementing the method, the decoder 500 decodes first data relating to first information, for example instructions corresponding to one or more management modes for determining d 'one or more lists of encoding parameters. These instructions defined for example in the PPS or the “slice headers” can be used for determining elements of an image, or for a set of images of the sequence of images, for example a GOP or images presenting the same encoding characteristics of image elements.

[0147] Thus, from the instructions defined in the first information, the decoder 500 determines one or more lists of encoding parameters on the basis of one or more predefined lists of encoding parameters that the decoder has in its memory. These determined lists of encoding parameters are identical to those that the encoder has determined for encoding the picture elements of the picture sequence. In fact, from the first information, the decoder can deduce the parameter setting of the encoder, and thus the various management modes which can be used.

[0148] When the decoder decodes an element, it determines the encoding characteristics of this element to deduce therefrom the determined list to be used among the determined lists. Then it decodes the second data relating to the second information. This second information provides information on the index of the encoding parameter to be used from the determined list of encoding parameters to be used for determining the element to be decoded. This process may be repeated for all image elements in the sequence of images to generate an output 507 of an image sequence.

[0149] In the case where the coding parameters are intra prediction modes, for an image to be decoded, from the first data encoded in the binary stream 506, the decoder determines one or more lists of intra prediction modes which are identical to those used for encoding this image. A list can be determined to be used for the determination of image elements encoded according to characteristics of type intra, that is to say elements of an image of type intra, and for elements of size 16x16 pixels. , then an element encoded according to these characteristics, that is to say encoded in an image of intra type (or intra image) and having a pixel size of 16x16 pixels, can then be predicted (determined) on the basis of this list of prediction modes. From the second encoded information the decoder identifies the index of the encoding parameter of the determined list to be used for the decoding of the element to be decoded.

[0150] Figure 4b illustrates an implementation of the method in one or more embodiments on the encoder side. Thus, an encoder 501 receives a sequence of images 502 to be encoded. The encoder also receives input data 503 allowing the encoding to be configured. These input data can contain instructions, which can indicate, among other things, the implementation of the method, the management mode to be used, the encoding characteristics. of image elements, and / or the use of encoding parameter lists as a function of encoding characteristics.

[0151] During the encoding process, the encoder 501 can then generate a binary stream comprising the encoded image elements 504. The binary stream 504 can also include the input data having served for the encoding of the elements. images and which can allow the decoder to correctly decode them. This input data can be encoded in the first data located in the signaling part of the bit stream, for example PPS.

[0152] In addition, the indices of the encoding parameters used for encoding the picture elements can be signaled by encoding in the second data located in the signaling portion of the bit stream 504.

Obtaining a predictive element by means of the optimal or relevant coding parameter makes it possible to calculate, as previously described for the calculation of a residue, which can then be transformed, for example by DCT, then quantified according to a step of quantification. After quantization, the data from the encoding can be encoded into the bitstream 504 called "bitstream".

[0154] FIG. 5 illustrates an example of a management mode implemented by the method for determining an element of an image being encoded or decoded. Thus, when the management mode is a so-called “explicit” management mode, the decoder 500 determines, from the instructions contained in the first information defining this management mode, one or more lists of coding parameters. The instructions may for example explicitly list a first set of encoding parameter indices. A second list may explicitly list a second set of encoding parameter indices. A third list may explicitly list a third set of encoding parameter indices, etc. The instructions may, in addition, specify encoding characteristics of elements for which these predefined lists of encoding parameter indices are to be used.

[0155] On reading the instructions defining the first set of encoding parameter indices, and from the predefined list 600 of encoding parameters L _P comprising N-1 encoding parameters, the decoder 500 determines 603 the first list LI_E1 by selection of the coding parameters according to the indices defined in the first set of indices. Here, 3 encoding parameter indices are defined in the instructions for the first list. The indices can refer to the predefined list L _P 600. For example the index 3 for the coding parameter p3, the index 2 for the coding parameter P2, and the index 5 for the coding parameter p ₅ . Thus, the first indexed list LI_E1 601 can be determined 603 and comprising a first index 0 corresponding to the coding parameter p3, an index 1 corresponding to the coding parameter P2, and an index 2 corresponding to the coding parameter ps.

According to the instructions included in the first information, the decoder 500 can determine 603 a second indexed list of coding parameters LI_E2 604 on the basis of the predefined list L _P 600. For example, the instructions can define a second set of coding parameter indices comprising 5 coding parameters p ₃ , P4, P7, Ps, P10 associated respectively with the indices 3, 4, 7, 8 and 10 in the predefined list of coding parameters. From these instructions, the decoder 500 can determine an indexed list LI_E2 comprising the coding parameters P3, P4, P7, Ps, P10 associated respectively with the indices 0, 1, 2, 3, 4.

The use of the first and / or second list for determining the elements of the image to be restored can depend on the encoding characteristics of the elements of the image which can be indicated in the instructions included in the first ones. information. For example, the first list LI_E1 can be used for the determination of elements encoded in intra (spatial prediction), and the second list LI_E2 can be used for the determination of elements encoded in inter (temporal prediction) (in particular for example when the encoding parameters are so-called “offset” values.

The determination of a multitude of determined lists of encoding parameters can allow, by constraining their use to the encoding characteristics of picture elements, a great adaptability of the possible coding parameters for the encoding of the elements of. an image of a sequence of images while maintaining low signaling cost. FIG. 6a illustrates an example of a mode of management of coding parameters by “Decimation”. Decoder 500 determines at least one indexed list 702 by selecting encoding parameters from the prefinished lists of encoding parameters 700 (eg, a list of 67 intra prediction modes).

From the information included in the first information, the decoder 500 obtains information on the fact that the so-called "decimation" management mode is activated and the values of the selection shift parameters k and of the decimation factor n. As indicated above, this information enables the decoder 500 to determine a list of encoding parameters by selecting the indices k + n.m from a predefined list of encoding parameters. The information can also comprise, in addition to or in replacement of k and n, a hierarchical level of the image to be decoded in the structure of the GOP defined by an integer L. In the case of FIG. 6a, the value of n is equal to 2 and the value of k is equal to 1.

[0161] Thus the decoder 500 determines the list LI_D1 indexed according to a new indexing 702 and where each index of this list LI_D1 is associated with an encoding parameter obtained following the selection of one encoding parameter out of two in the predefined parameter list. coding 700.

[0162] Then during the decoding of the elements of the image, the decoder 500 can use the lists determined by decimation according to the encoding characteristics of the element being decoded. For example, a first list 702 can be used exclusively for the decoding of elements having a certain level of brightness and included in images of intra type. A second indexed list (not shown) can be used exclusively for the decoding of elements having an element size greater than a value (for example greater than 8x8 pixels) included in type B or type P images.

When the first information includes a hierarchical level L, the decimation is carried out only for the images below (or above) a certain hierarchical level. It is also possible to define several hierarchical levels with for each level specific values of k and n. When the value L is included in the first information, the values k and n can be implicit, that is to say that on the basis of the value L the decoder 500 deduces a pair k, n without these being included in the first information.

[0164] FIG. 6b illustrates hierarchical levels in a GOP structure comprising 19 images 760, 761, 762. The structure of the GOP here comprises several hierarchical levels L0 to L5.

[0165] For each value of the hierarchical level L, the selection by decimation can provide for different values of k and n (for example the pairs (1; 2), (1; 4) and (1; 8)). The determined lists are then used for the decoding of the elements of each image of the GOP according to the hierarchical level L1 to L5 of the image in which the element to be decoded is located (for example the list obtained with the pair (1; 2 ) can be used for images of hierarchical levels L3 to L5, the list obtained with the pair (1; 4) can be used for images of hierarchical level L2 and the list obtained with the pair (1; 8) can be used for the images of hierarchical level L1, the predefined list being used for the images of hierarchical level L0.

[0166] The use of the lists obtained by decimation can be made only for the decoding of images of the same type. For example, with reference to FIG. 6b, only B type images 762 can be decoded on the basis of lists obtained by decimation. Considering that image 760 is an intra-type image and that image 761 is a predictive-type image.

[0167] According to the first information, the decimation can be carried out on only part of the predefined list. For example, when the encoding parameters are intra prediction modes, the decimation can be performed only on the angular modes. The first information can then contain additional information, for example the indices of the coding parameters to be kept before carrying out the decimation. It is also possible to define a value of k sufficient to spare the first coding parameters of the predetermined list.

[0168] For example, the first determined list LI_D1 then comprises 35 prediction modes. Indeed, this first list includes 2 modes not angular, namely DC and Planar, for 65 angular modes. The list LI_D1 then comprises the two non-angular modes and the angular modes resulting from the selection by decimation, ie after selection a total of 33 angular modes.

[0169] FIG. 7 illustrates an example of another management mode implemented by the method. Thus, when the management mode is a management mode by “Histogram”, the decoder 500 determines 850 at least one indexed list 810 or 870 as a function of a level of use frequencies of the coding parameters of the predefined list of. encoding parameters. This frequency of use may, for example, be representative of the use of an encoding parameter for determining the previous picture elements and / or the picture being decoded. According to one example, the frequency of use for each encoding parameter of a predefined list can be represented in the form of a histogram.

[0170] Thus, with reference to FIG. 7, a histogram 800 can be defined by an abscissa 840 of coding parameters, and according to an ordinate 830 indicating the frequency of use, for example in percentage.

[0171] The list of encoding parameters can be determined from a selection of encoding parameters having a frequency of use obeying selection rules.

The selection rules for this management mode can be defined for example by a selection factor h, and / or a selection threshold S. These selection parameters can be included in the first information allowing the decoder 500 to define the management mode.

Thus, the selection rules can be the selection of the h coding parameters from among the coding parameters of the histogram (h a positive integer) having the highest frequencies of use. The determined list of coding parameters can for example include a ranking in decreasing or increasing order of the frequency of use. For example, according to histogram 800, when the factor h is equal to 3, the list LI_H1 810 of determined coding parameter 850 comprises the three coding parameters p ₁₀ , p ₉ , and p ₆ presenting the most frequent use frequencies. higher among encoding parameters from the predefined list. This list is reordered so that the index 0 is associated with the encoding parameter p ₁₀ the index 1 is associated with the encoding parameter p ₉ and the index 2 is associated with the encoding parameter p ₆ .

The first information can then include information on the fact that the management mode called "histogram" is activated and the factor h.

The selection rules can also be the selection of coding parameters having frequencies of use greater than a threshold "S" 820. The determined list of coding parameters can for example include a ranking in decreasing or increasing order of frequency of use. For example, according to the histogram 800, with the threshold S 820, the list LI_H2 870 of coding parameters is determined 860 which comprises the coding parameters rb, P ₇ , R _Q , P _IO and p ₁₂ presenting the frequencies use of the encoding parameters in the predefined list. This list is reordered so that the index 0 is associated with the encoding parameter p ₁₀ , the index 1 is associated with the encoding parameter p ₉ , the index 2 is associated with the encoding parameter rb, the index 3 is associated with the coding parameter p, the index 4 is associated with the coding parameter pi ₂ .

The first information can then include information on the fact that the management mode called "histogram" is activated and the selection threshold "S".

[0177] It may be possible to combine the selection parameters. Thus, following the instructions defined in the configuration of the "histogram" management mode, it may be possible, for example, to apply a selection threshold followed by a selection factor.

[0178] FIG. 8 illustrates an example of another management mode implemented by the method. When the management mode is a management mode called for example “group”, the decoder 500 determines 901 at least one list of indexed coding parameters 902 comprising one or more so-called primary coding parameters surrounded by so-called secondary coding parameters according to the 'set I - [Ϊ (R) _K - m * :; KR) K + m *]. where i (p) k are the indices of the primary encoding parameters and m _A. (shift factor) are strictly positive integers specified for example in the instructions of the so-called “group” management mode included in the first information.

[0179] For example, in FIG. 8, the first information contains three primary coding parameters (P2, P7, and p,) and an offset factor (d) of 1. On the basis of these instructions the decoder 500 determines list 902. This contains the explicitly signaled primary encoding parameters and the secondary encoding parameters which are then pi, p ₃ , R _Q , P _S , P _M and p _{i + 1} .

[0180] FIG. 9 illustrates the method proposed according to one or more embodiments. This method can be implemented by any device comprising a processor and a memory, and configured for its implementation, such as, for example, a video encoder 501, or any equipment configured for encoding a video sequence.

[0181] In step 17, the device 501 determines an encoding setting for an image of the sequence of images (video sequence). For example, the encoding setting can be encoding input data, such as, for example, instructions specifying the management modes to be used, the encoding characteristics of elements, and / or the management modes. to be used according to the encoding characteristics.

[0182] In step 18, the device 501 determines a list of encoding parameters. The determination of the list of coding parameters can be carried out according to one of the previously indicated methods, for example, by "decimation", by "histogram", by explicit selection or even by "group". These different methods can also be combined, i.e. a first method (from among its methods) is applied to select from the predefined list of encoding parameters a first list of encoding parameters and a second method. (from among its methods) is applied to select from this first list of encoding parameters a second list of encoding parameters.

[0183] In step 19, the device 501 obtains a binary stream comprising first information as well as encoded data representative of the elements of the image. The first information can for example indicate the parameter setting used for the encoding of the image of the sequence of images.

[0185] Second information can be included in the encoded data representative of the elements. This second information indicates the encoding parameter used from the list determined to obtain an element of the image.

[0186] FIG. 10 illustrates an example of the architecture of a decoding device, and an example of the architecture of a device for encoding a video sequence for the implementation of the proposed method.

[0187] The device 500 comprises a controller 902, operably coupled to an input interface 901, to an output interface 906 and to a memory 903, which controls a coding parameter management module 905 and a control unit. decoding 904 for decoding images of a video sequence.

[0188] The input interface 901 is configured to receive as input a binary stream corresponding to a sequence of images encoded, according to the method for example, in a binary stream.

[0189] The controller 902 is configured to drive the coding parameter management module 905 and the decoding unit 904 for the implementation of one or more embodiments of the proposed method.

The coding parameter management module 905 can be configured for the implementation of one or more embodiments of the method for determining and using lists of coding parameters proposed. In particular, it can be configured to receive on the input interface a binary stream encoded according to the method and comprising first and second encoded data. On decoding this binary stream, the management mode can be configured to obtain first information from first encoded data contained in the binary stream, to determine a list of encoding parameters based on the first information and at least a predefined list of encoding parameters. The coding parameter management module 905 can furthermore be configured to obtain second information from second encoded data contained in the stream. binary, the second information corresponding to an encoding parameter among the encoding parameters of the list of encoding parameters. The memory 903 can be configured to store signaled information in the bit stream, such as first and second decoded information of the first and second encoded data.

[0191] The decoding unit 904 via the management module can be configured to determine an element of a restored image corresponding to an element of an image of the sequence of images by means of a parameter of encoding corresponding to the second information and encoded data representative of the element of the image of the sequence, and in which the size of the second encoded data depends on a number of encoding parameters in the list of encoding parameters.

[0192] The device 500 can be configured to provide at output 906 a decoded video data stream, for example a sequence of images.

[0193] The device 500 may be a computer, a computer network, an electronic component, or another device comprising a processor operably coupled to a memory, as well as, depending on the embodiment chosen, a storage unit. data, and other associated hardware elements such as a network interface and a media drive for reading and writing to removable storage media (not shown in Fig. 10). The removable storage medium can be, for example, a compact disc (CD), a digital video / versatile disc (DVD), a flash disc, a USB stick, etc. Depending on the embodiment, the memory, data storage unit, or removable storage medium contains instructions which, when executed by controller 902, cause controller 902 to perform or control interface parts of it. input 901, coding parameter management module 905, image decoding and / or data processing of the examples of implementation of the proposed method described herein. The controller 902 can be a component implementing a processor or a calculation unit for the decoding of one or more images of a sequence of images on the basis of at least one list of encoding parameters according to the proposed method and the control of units 901, 902, 903, 904, 905, 906 of device 500. [0194] In addition, the device 500 can be implemented in software form, as described above, or in hardware form, such as an application specific integrated circuit (ASIC), or in the form of a combination of hardware elements. and software, for example a software program intended to be loaded and executed on a component of FPGA (Field Programmable Gâte Array) type.

[0195] The device 501 comprises a controller 912, operatively coupled to an input interface 91 1, to an output interface 916 and to a memory 913, which controls a coding parameter management module 915 and a unit. encoding 914 for encoding images of a sequence of images.

[0196] The 911 input interface is configured to receive as input a sequence of images corresponding, for example, to a video sequence or a live distributed video stream.

[0197] The controller 912 is configured to drive the encoding parameter management module 915 and the encoding unit 914 for the implementation of one or more embodiments of the proposed method.

[0198] The coding parameter management module 915 can be configured for the implementation of one or more embodiments of the method for determining and using lists of coding parameters proposed. In particular, it can be configured for, an image of the sequence of images, to determine an encoding setting, for example from the input interface. The encoding parameter management module 915, further, can be configured to determine a list of encoding parameters based on the encoding setting and at least one predefined list of encoding parameters.

[0199] The input interface is configured to also receive input data comprising, for example a parameterization of the encoding of the sequence of images, for example instructions, which can signal, among other things, the setting. implementation of the method, the management mode to be used, the encoding characteristics of image elements of the sequence of images, and the way in which the use of the lists of encoding parameters of the management modes may depend of these encoding characteristics. [0200] Memory 913 can be configured to store input information, such as input data.

[0201] The encoding unit 914 can be configured to obtain a video binary stream comprising, encoded data representative of first information, the first information indicating the encoding setting, as well as encoded data representative of an element of the image obtained by means of an encoding parameter of the list of determined encoding parameters, the encoded data representative of the element comprises encoded data representative of second information, the size of the encoded data representative of the second information depends on a number of coding parameters in the list of coding parameters, the binary stream obtained being able to be sent to the device 500 for example.

[0202] The device 501 may be a computer, a computer network, an electronic component, or another device comprising a processor operably coupled to a memory, as well as, depending on the embodiment chosen, a storage unit. data, and other associated hardware elements such as a network interface and a media drive for reading and writing to removable storage media (not shown in Fig. 10). The removable storage medium can be, for example, a compact disc (CD), a digital video / versatile disc (DVD), a flash disc, a USB stick, etc. Depending on the embodiment, the memory, the data storage unit or the removable storage medium contains instructions which, when executed by the controller 912, cause this controller 912 to perform or control the interface parts of it. input 911, coding parameter management module 915, encoding of images and / or data processing of the examples of implementation of the proposed method described herein. The controller 912 can be a component implementing a processor or a computing unit for encoding one or more images of a video sequence on the basis of at least one list of encoding parameters according to the proposed method and the control. units 911, 912, 913, 914, 915, 916 of device 501. [0203] Furthermore, the device 501 can be implemented in software form, as described above, or in hardware form, such as an application specific integrated circuit (ASIC), or in the form of a combination of hardware elements. and software, for example a software program intended to be loaded and executed on a component of FPGA (Field Programmable Gâte Array) type.

The method is not limited to the examples of embodiments described above, only by way of example, but it encompasses all the variants that a person skilled in the art may envisage in the context of the claims above. after.

Claims

Claims [Claim 1] A method of decoding a video bit stream corresponding to a sequence of images, the method comprising: - Obtain first information from first encoded data contained in the binary stream; - determining a list of encoding parameters based on the first information and at least one predefined list of encoding parameters; obtaining second information from second encoded data contained in the binary stream, the second information corresponding to an encoding parameter from among the encoding parameters of the list of encoding parameters; determining an element of a restored image corresponding to an element of an image of the sequence of images by means of said coding parameter corresponding to the second information and of encoded data representative of the element of the image of the sequence ; wherein the size of the second encoded data depends on a number of encoding parameters in the encoding parameter list. [Claim 2] The method of claim 1, wherein the list of encoding parameters is determined by selecting encoding parameters from the predefined list of encoding parameters or by selecting from a plurality of predefined lists of encoding parameters. [Claim 3] A method according to any preceding claim, wherein the first encoded data and / or the second encoded data is included in a signaling portion of the bit stream. [Claim 4] A method according to any preceding claim, wherein the list of encoding parameters is a function of encoding characteristics of elements of the image. [Claim 5] A method according to claim 4, wherein the encoding characteristics of elements of the picture comprise an element size and / or an element quantization step and / or a chromatic type of element and / or an element brightness type and / or an intra encoding type element and / or an inter-element encoding type and / or a transform size and / or a transform type and / or a filter type. [Claim 6] A method according to any preceding claim, wherein the determination of the list of encoding parameters is by selecting from the predefined list of encoding parameters encoding parameters specified by the first information. [Claim 7] A method according to any one of the preceding claims, wherein the predefined list of encoding parameters is indexed from 0 to N-1, with N a number of encoding parameters in the predefined list and wherein the determination of the list of coding parameters is made by selection on the basis of the first information: - coding parameters of the predefined list of indices k + nm, with n strictly positive integer and k positive or zero integer and n + k less than or equal to N-1 and m between 0 and the default integer part of (Nk-1) / n; and - coding parameters from the predefined list of indices strictly less than k ’with k’ positive or zero integer less than or equal to k. [Claim 8] The method of claim 7 wherein n and m are determined based on a hierarchical level of the image relative to the images of the sequence of images. [Claim 9] A method according to any one of the preceding claims, in which the determination of the list of encoding parameters is carried out by selection on the basis of the first information of the encoding parameters of the predefined list as a function of frequency levels d 'use encoding parameters from the predefined list. [Claim 10] A method according to any preceding claim, wherein the predefined list of encoding parameters is indexed and wherein the determination of the list of encoding parameters is by selection based on the first information: - primary encoding parameters specified by the first information; and - secondary coding parameters such as an index of each secondary coding parameter is included in II _¾ [ί (r) _¾ - m _& ; î (p) _fc + m _& ] where i (p) k are the indices of the primary encoding parameters and m _¾ are strictly positive integers specified by the first information. [Claim 11] A method according to any preceding claim, wherein the predefined list of encoding parameters or the plurality of predefined lists are predefined lists of prediction modes or predefined lists of adaptive shift offset values of. a filter. [Claim 12] A method of encoding a video bit stream corresponding to a sequence of images, the method comprising: - for an image of the sequence of images, determine an encoding setting; - determine a list of encoding parameters based on the encoding setting and at least one predefined list of encoding parameters; - get a video binary stream including: -encoded data representative of first information, the first information indicating the encoding setting; - encoded data representative of an element of the image obtained by means of an encoding parameter of the list of determined encoding parameters, the encoded data representative of the element comprises encoded data representative of second information, the size encoded data representative of the second information depends on a number of encoding parameters in the list of encoding parameters. [Claim 13] A computer program comprising instructions for implementing the method according to one of claims 1 to 12, when this program is executed by a processor. [Claim 14] A device for decoding a video bit stream corresponding to a sequence of images, the device comprising: - a processor; and - non-transient computer support comprising instructions which, when executed by the processor, configures the device to: - Obtain first information from first encoded data contained in the binary stream; - determine a list of encoding parameters on the basis of the first information and at least one predefined list of encoding parameters; obtaining second information from second encoded data contained in the binary stream, the second information corresponding to an encoding parameter from among the encoding parameters of the list of encoding parameters; determining an element of a restored image corresponding to an element of an image of the sequence of images by means of said coding parameter corresponding to the second information and of encoded data representative of the element of the image of the sequence ; wherein the size of the second encoded data depends on a number of encoding parameters in the encoding parameter list. [Claim 15] A device for encoding a video bit stream corresponding to a sequence of images, the device comprising: - a processor; and - non-transient computer support comprising instructions which, when executed by the processor, configures the device to: - for an image of the sequence of images, determine an encoding setting; - determine a list of encoding parameters based on the encoding setting and at least one predefined list of encoding parameters; - get a video binary stream including: - encoded data representative of first information, the first information indicating the encoding setting; - encoded data representative of an element of the image obtained by means of an encoding parameter of the list of determined encoding parameters, the encoded data representative of the element comprises encoded data representative of second information, the size encoded data representative of the second information depends on a number of encoding parameters in the list of encoding parameters.