EP4226614A1 - Procédé et dispositif électronique de décodage d'un flux de données, et programme d'ordinateur associé - Google Patents
Procédé et dispositif électronique de décodage d'un flux de données, et programme d'ordinateur associéInfo
- Publication number
- EP4226614A1 EP4226614A1 EP21786467.7A EP21786467A EP4226614A1 EP 4226614 A1 EP4226614 A1 EP 4226614A1 EP 21786467 A EP21786467 A EP 21786467A EP 4226614 A1 EP4226614 A1 EP 4226614A1
- Authority
- EP
- European Patent Office
- Prior art keywords
- decoding
- neural network
- artificial neural
- context
- data
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T9/00—Image coding
- G06T9/002—Image coding using neural networks
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/90—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals using coding techniques not provided for in groups H04N19/10-H04N19/85, e.g. fractals
- H04N19/91—Entropy coding, e.g. variable length coding [VLC] or arithmetic coding
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/42—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation
- H04N19/436—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by implementation details or hardware specially adapted for video compression or decompression, e.g. dedicated software implementation using parallelised computational arrangements
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04N—PICTORIAL COMMUNICATION, e.g. TELEVISION
- H04N19/00—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals
- H04N19/70—Methods or arrangements for coding, decoding, compressing or decompressing digital video signals characterised by syntax aspects related to video coding, e.g. related to compression standards
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06T—IMAGE DATA PROCESSING OR GENERATION, IN GENERAL
- G06T2207/00—Indexing scheme for image analysis or image enhancement
- G06T2207/20—Special algorithmic details
- G06T2207/20084—Artificial neural networks [ANN]
Definitions
- the present invention relates to the technical field of data decoding.
- It relates in particular to a method and an electronic device for decoding a data stream, as well as an associated computer program.
- Entropy coding is used, particularly in the field of audio or video content coding, to optimally compress data by taking into account the appearance statistics of the different symbols in this data.
- CABAC coding for "Context-based adaptive binary arithmetic coding" is known in this context as described in the article "Context-based adaptive binary arithmetic coding in the H.264/AVC video compression standard', by D. Marpe, H. Schwarz, and T. Wiegand, in IEEE Transactions on Circuits and Systems for Video Technology, vol.13, no.7, pp. 620-636, July 2003.
- the entropy decoder is at each instant configured in a context which depends, in a manner predefined in the standard concerned, on the previously decoded syntax elements.
- the present invention proposes a method for decoding a sequence of bits, the method comprising the following steps:
- the decoding method can thus comprise, at each of a plurality of iterations (comprising a current iteration and a previous iteration), obtaining values decoded by entropy decoding (by means of the aforementioned entropy decoder) , the method for decoding the sequence of binary elements defined above can then comprise the following steps:
- the method may further comprise a step of applying the new decoded value at the input of the artificial neural network so as to produce, at the output of the artificial neural network, data representative of an audio or video content.
- the artificial neural network thus produces on the one hand the context index and on the other hand the data representative of the audio or video content.
- the sequence of bits may be included in a data stream.
- This data stream can then further comprise information indicative of a set of contexts that can be used within the entropy decoder.
- This information is for example indicative of a number of contexts that can be used within the entropy decoder.
- the method may further comprise a step of initializing each context usable within the entropy decoder, for example by means of parameterization data included in the data stream comprising the sequence of binary elements.
- the artificial neural network can be implemented by a processing unit.
- the method can then comprise a step of configuring the processing unit according to data included in the data stream comprising the sequence of binary elements.
- the entropy decoder can be parameterized in a predefined initial context as long as the artificial neural network does not produce a context index.
- the entropy decoding can thus be performed in this initial context to process the first elements of the sequence of binary elements and thus obtain first decoded values to be applied as input to the artificial neural network so that the artificial neural network produces an index of context.
- the artificial neural network is for example implemented by means of a parallelized processing unit designed to perform in parallel at a given instant a plurality of operations of the same type.
- the entropy decoder can be implemented by means of a separate processor from the parallelized processing unit.
- the invention also proposes a computer program comprising instructions executable by a processor and designed to implement a decoding method as described above when these instructions are executed by the processor.
- the invention finally proposes an electronic device for decoding a sequence of binary elements, comprising:
- an artificial neural network designed to receive previously decoded values as input and to produce a context index as output;
- control module designed to configure the entropy decoder in the context identified by the context index produced, so as to obtain a new decoded value at the output of the entropy decoder.
- the previously decoded values are for example decoded by entropy decoding during an iteration which precedes the iteration during which the new decoded value is obtained (/.e. produced by the entropy decoder).
- This electronic decoding device may comprise a synchronization mechanism capable of suspending an entropy decoding process by the entropy decoder as long as a new context index is not produced at the output of the artificial neural network.
- This electronic decoding device can also comprise a processing unit capable of implementing the artificial neural network.
- the control module can then be designed to configure the processing unit according to data included in the data stream comprising the sequence of bits, so that the processing unit can implement the network of artificial neurons as shown above.
- the electronic decoding device may also comprise a processor separate from the processing unit and designed to implement the entropy decoder.
- FIG. 1 shows a data processing assembly comprising several parts of an artificial neural network
- FIG. 1 schematically illustrates maps of characteristics used within the processing unit of Figure 1;
- FIG. 3 shows an electronic coding device used in the context of the invention
- FIG. 4 is a flowchart representing the steps of an encoding method implemented within the electronic encoding device of Figure 3;
- FIG. 5 shows the data stream produced by the electronic encoding device of Figure 3;
- FIG. 6 shows an example of an electronic decoding device according to the invention.
- FIG. 7 is a flowchart representing the steps of a decoding method implemented within the electronic decoding device of Figure 6.
- FIG. 1 represents a data processing assembly of which different parts are used either for the coding of an audio or video content, or for the decoding of coded data in order to render an audio or video content, as explained below.
- This set includes a coding artificial neural network 8, an entropic encoder 10, a context determining artificial neural network 40, an entropic decoder 30 and a decoding artificial neural network 28.
- the coding artificial neural network 8 is designed to receive as input (that is to say on an input layer) content data B forming a representation (here uncompressed) of an audio or video content.
- content data includes, for each pixel of each image of a sequence of images, data representing a luminance value of the pixel and data representing chrominance values of the pixels.
- the content data B applied at a given time to the input layer of the coding artificial neural network 8 can represent a block of an image, or a block of a component of an image (for example a block of 'a luminance or chrominance component of this image, or a block of a color component of this image), or an image from a video sequence, or a component of an image from a video sequence (for example a component luminance or chrominance, or a color component), or even a series of images of the video sequence.
- the coding artificial neural network 8 When these content data B are applied as input (i.e. on the input layer) of the coding artificial neural network 8, the coding artificial neural network 8 produces values V at the output, they also representative of the audio or video content.
- the representative values V produced at the output of the coding artificial neural network 8 however form a more compact representation than the corresponding content data B applied at the input of the coding artificial neural network 8.
- the number of nodes of the layer output of the artificial neural network coding 8 is lower (for example 4 times lower, even 8 times lower or 16 times lower) than the number of nodes of the input layer of the artificial neural network coding 8.
- the representative values V produced at the output of the artificial neural network 8 are for example organized into a sequence of feature maps F (or "feature maps" according to the Anglo-Saxon designation sometimes used), as schematically represented in FIG. 2.
- the coding artificial neural network 8 produces for example here N maps of characteristics F.
- Each feature map F has for example a two-dimensional structure (or matrix structure).
- each feature map F here forms a matrix with H rows and W columns.
- An element located at a given position in a given map of features F corresponds to the representative value V produced by an output node (or output layer node) of the coding artificial neural network 8, this output node being associated in a predefined manner at this given position and at this given map of characteristics F.
- the network of artificial neurons 8 produces as output (that is to say on its output layer) at a given instant all of the N maps of characteristics.
- different sets of content data B are applied at different times at the input (i.e.
- the network of artificial neurons 8 produces at each of these different instants at output (that is to say on its output layer) a corresponding map of characteristics F (the output nodes of the artificial neural network 8 being in this case associated respectively with the different positions of a single map of characteristics F).
- the representative values V produced at the output of the coding artificial neural network 8 can be organized into an ordered sequence of representative values V.
- the content data B applied to the input of the coding artificial neural network 8 represents a block of an image (or a block of a component of an image)
- the ordered sequence of representative values V produced at the output of the coding artificial neural network 8 is associated with this block.
- the different sequences of representative values successively produced by the artificial coding neural network 8 are thus respectively associated with the different blocks of the image (or with the different blocks of the relevant component of the image).
- the representative values V produced at the output of the coding artificial neural network 8 are placed within a data structure with several dimensions (for example M dimensions). Each element of this structure is then identified by its position within the structure, namely in the aforementioned example by means of an M-tuple of coordinates.
- a representative value V produced by a given output node (that is to say by a given node of the output layer) of the coding artificial neural network 8 then forms an element of the structure designated by a position within of the structure associated in a predefined manner with this output node (that is to say by coordinates associated in a predefined manner with this output node).
- the representative values V produced at the output (that is to say on the output layer) of the coding artificial neural network 8 are applied on the one hand to the input of the entropic coder 10 and on the other hand to the input (c i.e. on the input layer) of the context determination artificial neural network 40.
- the context determination artificial neural network 40 thus receives as input (that is to say on its input layer) representative values V (corresponding for example to a block of an image being coded, or to a block of a component of the image being coded) and consequently produces at output a context index C.
- the context index C is here produced on an output node of the artificial neural network for determining context 40.
- the context determination artificial neural network 40 presents only this single output node.
- the context determination artificial neural network 40 could produce as output a plurality of context indices C each associated respectively with at least one representative value V (for example with a set of representative values V).
- the context indices C produced at the output of the context determination artificial neural network 40 can be respectively associated with the different maps of characteristics F (each containing a set of representative values V).
- the context index C produced by the context determination artificial neural network 40 is applied to the entropy coder 10 and to the entropy decoder 30.
- the entropic coder 10 is designed to code several statistical sources each corresponding to a particular probability of appearance of the symbols to be coded. To do this, the entropy coder 10 can be parameterized in a particular context, associated with a given statistical source and in which the entropy coder 10 produces an optimal entropy code if the symbols actually coded (here the representative values V) respect the expected probability for this statistical source.
- K the number of contexts in which the entropy coder 10 can be parameterized during the entropy coding of the representative values V (that is to say the number of different statistical sources which can be processed in the signal formed by representative values V).
- K 160.
- the entropy coder 10 is here a coder of the CABAC type (for "Context-based adaptive binary arithmetic coding”). Alternatively, it could be another type of entropy coder, for example a Huffman type coder, an arithmetic coder or an LZW coder (for "Lempel-Ziv-Welch”).
- the representative values V received at the input of the entropy coder 10 are ordered in a predefined manner for entropy coding by the entropy coder 10.
- the different F feature maps are processed in the order of this sequence and, within each F feature map, the elements (i.e. the representative values V) are taken into account in a predefined (scanning) order.
- the entropy coder 10 carries out the entropy coding of the ordered representative values V received as input, while being parameterized in the context designated by the context index C produced at the output of the context determination artificial neural network 40.
- the entropy coder 10 performs the entropy coding of the ordered representative values V received as input , while being parameterized (at each instant) in the context designated by the context index C associated with the feature map F during entropy coding.
- the entropic encoder 10 then produces at output a sequence of bits Fnn.
- this sequence of binary elements Fnn corresponds to the compressed data stream representing the audio or video content (data stream generated at the output of the electronic coding device 2 described later with reference to FIG. 3 and intended to the electronic decoding device 20 described below with reference to FIG. 6).
- the sequence of binary elements Fnn is applied at the input of the entropy decoder 30, the entropy decoder 30 being parameterized in the context designated by the context index C produced at the output of the context determination artificial neural network C.
- the entropy decoder 30 is designed to perform inverse entropy decoding of the entropy coding performed by the entropy encoder 10 mentioned above.
- the entropy decoder 30 is therefore here a CABAC type entropy decoder (for "Context-based adaptive binary arithmetic coding").
- CABAC type entropy decoder for "Context-based adaptive binary arithmetic coding”
- the entropy decoder 30 therefore produces at output representative values V identical to those applied at the input of the entropy coder 10. (It should be recalled in this respect that entropy coding is lossless coding.)
- the representative values V produced at the output of the entropy decoder 30 are applied at the input (that is to say on an input layer) of the decoding artificial neural network 28.
- the respective allocation of the representative values V to the input nodes (or nodes of the input layer) of the decoding artificial neural network 28 is predefined. (It should also be noted that the output layer of the coding artificial neural network 8 corresponds to the input layer of the decoding artificial neural network 28. Indeed, the use of entropic coding allows better data compression , but does not modify this data.)
- the decoding artificial neural network 28 When the decoding artificial neural network 28 receives the representative values V as input, the decoding artificial neural network 28 produces at output (that is to say on an output layer) a representation I of the content adapted to a reproduction on an audio or video reproduction device.
- the artificial neural network 28 thus produces as output (that is to say at the level of its output layer) at least one representation matrix I of an image block (or of a block of an image component or, alternatively, of an image or of an image component).
- Such a set of data can be optimized, during a training phase of the various artificial neural networks 8, 28, 40 as described now, for a particular type of content and/or a particular rate-distortion trade-off.
- a sequence of learning audio or video content (for example here a series of learning videos) is selected. This is a set of content representative of the type of content that it is desired to compress with this data processing set.
- Each content (here each video) of the learning sequence can then be applied (as content data B) as input to the coding artificial neural network 8, which makes it possible to produce each time (as explained below above) a sequence of binary elements Fnn (at the output of the entropic encoder 10) and a representation I to be displayed (at the output of the decoding artificial neural network 28).
- a cost function is used to numerically evaluate the performance of the data processing set in its current configuration.
- a cost function is for example rate-distortion cost such as R+ ⁇ .D, where D is the distortion (quadratic error) between the rendered content (using the representation I) and the initial content (represented by the content data B), R is the bit rate (actual or estimated) of the compressed stream (i.e. of the sequence of bits Fnn), and ⁇ is a parameter supplied by the user, making it possible to choose the compromise between compression and quality.
- This cost function is used within a gradient backpropagation learning algorithm to scale the weights assigned to the neurons of the artificial neural networks 8, 28, 40 so as to minimize the cost function.
- weights assigned to the neurons of the artificial neural networks 8, 28, 40 when the minimum cost is considered reached define these artificial neural networks 8, 28, 40, and therefore the data processing set, as they will be used in the following.
- FIG. 3 thus represents an electronic coding device 2 using the coding artificial neural network 8, the context determination artificial neural network 40 and the entropic coder 10.
- This electronic coding device 2 comprises a processor 4 (for example a microprocessor) and a parallelized processing unit 6, for example a graphics processing unit (or GPU for "Graphical Processing Unit') or a tensor processing unit (or TPU for "Tensor Processing Unit”).
- a processor 4 for example a microprocessor
- a parallelized processing unit 6 for example a graphics processing unit (or GPU for "Graphical Processing Unit') or a tensor processing unit (or TPU for "Tensor Processing Unit”).
- the processor 4 is programmed (for example by means of computer program instructions executable by the processor 4 and stored in a memory - not shown - associated with the processor 4) to implement a control module 5 and the entropic encoder 10 already mentioned.
- control module 5 receives data P, B representing an audio or video content to be compressed, here format data P and content data B.
- content data B are of the same nature as those mentioned above in the context of the description of Figure 1 and will not be described again.
- the format data P indicates characteristics of the format of representation of the audio or video content, for example for a video content the dimensions (in pixels) of the images, the frame rate, the bit depth of the luminance information and the depth in bits of the chrominance information.
- the parallelized processing unit 6 is designed to implement the coding artificial neural network 8 and the context determination artificial neural network 40 (both forming part of a data processing set such as that of the FIG. 1) after having been configured by the processor 4 (for example by the control module 5). To do this, the parallelized processing unit 6 is designed to perform in parallel at a given instant a plurality of operations of the same type.
- the “global coding network” 9 denotes the artificial neural network formed by the artificial coding neural network 8 and by the artificial context determination neural network 40.
- the parallelized processing unit 6 is thus designed to implement this global coding network 9 (after having been configured by the processor 4 as already indicated).
- the method of FIG. 4 begins with a step E2 of selecting a data processing set from among a plurality of data processing sets in accordance with what has been described above with reference to FIG.
- these different data processing sets can have the same general structure (as shown in Figure 1), but the different artificial neural networks 8, 28, 40 are defined by varying weights (associated with the neurons) from one data processing set to another (the optimization criteria being different from one data processing set to another).
- the data processing set can for example be selected from among data processing sets for which the decoding artificial neural network 28 and the context determination artificial neural network 40 (together forming a global decoding network as explained later) are available for an electronic decoding device (such as the electronic decoding device 20 shown in FIG. 6 and described later).
- the electronic coding device may optionally receive beforehand (from the electronic decoding device or from a dedicated server) a list of artificial neural networks accessible by this electronic decoding device.
- the data processing assembly can also be selected according to the intended application (indicated for example by a user by means of a user interface (not shown) of the electronic coding device 2). For example, if the targeted application is a videoconference, the selected data processing set allows low latency decoding. In other applications, the selected data processing set might allow random access decoding.
- an image of the video sequence is for example represented by coded data which can be sent and decoded immediately; the data can then be sent in the order in which the images of the video are displayed, which in this case guarantees a latency of one image between the coding and the decoding.
- the coded data relating respectively to a plurality of images are sent in an order different from the display order of these images, which makes it possible to increase the compression .
- Images coded without reference to the other images can then be coded regularly, which makes it possible to start the decoding of the video sequence from several places in the coded stream.
- the data processing set can also be selected in order to obtain the best possible compression-distortion compromise.
- the different criteria for selecting the data processing set can optionally be combined.
- control module 5 proceeds in step E4 to configure the parallelized processing unit 6 so that the parallelized processing unit 6 can implement the coding provided in this data processing set.
- This step E4 comprises in particular the instantiation, within the parallelized processing unit 6, of the global coding network 9 comprising the coding artificial neural network 8 and the context determination artificial neural network 40 of the selected data processing set.
- This instantiation may include the following steps:
- the following steps aim at the coding (that is to say the preparation) of the data stream intended for the electronic decoding device (for example the electronic decoding device 20 described below with reference to FIG. 6).
- the method thus notably comprises a step E6 of encoding a first header part Fc which comprises data characteristic of the format of representation of the audio or video content (here for example data linked to the format of the current video sequence coding).
- These data forming the first header part Fc indicate for example the dimensions (in pixels) of the images, the frame rate, the bit depth of the luminance information and the bit depth of the chrominance information. These data are for example constructed on the basis of the data of format P mentioned above (after a possible reformatting).
- the control module 5 proceeds in step E8 to the coding of a second header part comprising data R indicative of the global decoding network comprising the artificial neural network for decoding 28 and the artificial neural network for determining of context 40 which are part of the data processing set selected in step E2.
- these indicative data R can comprise an identifier of the global decoding network.
- Such an identifier designates (among a plurality of global decoding networks, for example among all the global decoding networks available for the electronic decoding device) the global decoding network corresponding to the global coding network 9 mentioned above, global network of decoding which must therefore be used for decoding the representative values V.
- a global decoding network comprising on the one hand a network of artificial decoding neurons corresponding to the network of artificial coding neurons 8 included in the global coding network 9 , and on the other hand the context determination artificial neural network 40 included in the global coding network 9.
- such an identifier defines by convention (shared in particular by the electronic coding device and the electronic decoding device) this global decoding network, for example within the set of global decoding networks available for (or accessible by ) the electronic decoding device.
- the electronic coding device 2 can optionally receiving beforehand (from the electronic decoding device or from a dedicated server) a list of artificial neural networks accessible by the electronic decoding device.
- these indicative data R can comprise descriptive data of the global decoding network.
- the global decoding network (comprising the decoding artificial neural network 28 and the context determination artificial neural network 40 which are part of the data processing set selected in step E2) is for example coded (c that is to say represented) by these descriptive data (or coding data of the decoding artificial neural network) in accordance with a standard such as the MPEG-7 part 17 standard or in a format such as the JSON format.
- the indicative data R can also be made for the indicative data R to include an indicator indicating whether the global decoding network is part of a predetermined set of artificial neural networks (in which case the first possibility of embodiment mentioned above is used) or whether the network global decoding is encoded in the data stream, that is to say represented by means of the aforementioned descriptive data (in which case the second embodiment mentioned above is used).
- the method of FIG. 4 continues with a step E10 of determining the possibility for the electronic decoding device to implement the decoding process using the aforementioned global decoding network.
- the control module 5 determines for example this possibility by determining (possibly by means of prior exchanges between the electronic coding device 2 and the electronic decoding device) whether the electronic decoding device comprises a module adapted to implement this process decoding or software adapted to the implementation of this process of decoding by the electronic decoding device when this software is executed by a processor of the electronic decoding device.
- control module 5 determines that it is possible for the electronic decoding device to implement the decoding process, the method continues at step E14 described below.
- step E12 If the control module 5 determines that it is not possible for the electronic decoding device to implement the decoding process, the method performs step E12 described below (before going to step E14 ).
- step E12 could be made on another criterion, for example according to a dedicated indicator stored within the electronic coding device 2 (and possibly adjustable by the user via a user interface of the electronic coding device 2) or according to a choice of the user (obtained for example via a user interface of the electronic coding device 2).
- the control module 5 codes in the data stream at step E12 a third header part containing a computer program Exe (or code) executable by a processor of the electronic decoding device. (The use of the Exe computer program within the electronic decoding device is described below with reference to Figure 7.)
- the computer program is for example chosen within a library according to information relating to the hardware configuration of the electronic decoding device (information received by example during prior exchanges between the electronic encoding device 2 and the electronic decoding device).
- the method of FIG. 4 then continues with steps of encoding data representative of the configuration of the entropy coder 10 used in the data processing assembly selected in step E2 (and therefore of the entropy decoder 30 of this same assembly ).
- the method of FIG. 4 firstly comprises a step E14 of coding a fourth header part comprising information 11 indicative of the set of contexts used for the entropy coding.
- the information I1 is indicative of the number K of contexts used for the entropy coding.
- the method of FIG. 4 then comprises a step E16 of coding a fifth part of the header comprising, for each context used for the entropy coding, a data item linit for parameterizing the context concerned.
- the parameterization data linit associated with a given context is initialization data for this context, as described for example in Recommendation ITLI-T H.265, part "9.3.2.2 Initialization process for context variables".
- the parametrization datum associated with a given context may be datum indicative of the probability model used for the context concerned during the entropy coding.
- the method of FIG. 4 continues with a step E18 of initializing the entropic encoder 10 (by the control module 5) by means of the aforementioned parametrization data relating to the different contexts.
- This type of initialization is described in the document already mentioned (ITU-T Recommendation H.265, part "9.3.2.2 Initialization process for context variables").
- the method of FIG. 4 then comprises a step E20 of implementing the coding process, that is to say here a step of applying the content data B at the input of the global coding network 9 (or in other words a step of activating the global coding network 9 by taking the content data B as input). (The content data B is therefore then applied as input to the coding artificial neural network 8.)
- Step E20 thus makes it possible to produce (here at the output of the global coding network 9) the representative values V and the context index C.
- the representative values V are produced at the output of the artificial coding neural network 8; these representative values V are applied as input to the context determination artificial neural network 40 so that this context determination artificial neural network 40 produces the context index C as output.
- the method of FIG. 4 then comprises a step E22 of entropy coding of the representative values V by the entropy coder 10, the entropy coder 10 being parameterized (possibly via the control module 5) in the context defined by the index of context C produced at the output of the global coding network 9 (here precisely at the output of the context determination artificial neural network 40).
- the entropy coding of the representative values V by the entropic encoder 10 is carried out by parameterizing the encoder entropy 10 in the context defined by the context index C associated with the set containing the representative value V during entropy coding.
- the entropy coder 10 thus produces at output a sequence Fnn of binary elements representing, in compressed form, the audio or video content.
- the step E22 may in some cases include a sub-step (prior to the entropic coding strictly speaking) of binarization of the representative values V, as described in the article "Context-based adaptive binary arithmetic coding in the H.264/A VC video compression standard' cited above.
- the objective of this binarization step is to convert a representative value V which can take a large number of values into a series of binary elements, each binary element being coded by entropy coding (and in this case , a context is associated with the coding of each binary element).
- step E20 allows the processing of only part of the audio or video content to be compressed (for example when step E20 processes a block, or a component, or a image of a video sequence to be compressed), it is possible to repeat the implementation of steps E20 (to obtain values representative of the successive parts of the content) and E22 (to carry out the entropy coding of these representative values).
- the processor 4 can thus construct in step E24 the complete data stream comprising the header Fet and the sequence of bits Fnn.
- the complete data stream is constructed so that the Fet header and the Fnn sequence of bits are individually identifiable.
- the header Fet contains a start indicator of the sequence of bits Fnn in the complete data stream.
- This indicator is for example the location, in bits, of the start of the sequence of binary elements Fnn from the start of the complete data stream. (That is, the header in this case has a predetermined fixed length.)
- identifying the header Fet and the sequence of bits Fnn can be envisaged as a variant, such as for example a marker (i.e. i.e. a combination of bits used to indicate the start of the sequence of bits Fnn and whose use is prohibited in the rest of the data stream, or at least in the header Fet).
- a marker i.e. i.e. a combination of bits used to indicate the start of the sequence of bits Fnn and whose use is prohibited in the rest of the data stream, or at least in the header Fet.
- the data stream constructed in step E24 can be encapsulated in transmission formats known per se, such as the "Packet-Transport System” format or the "Byte-Stream” format.
- the data is coded by identifiable packets and transmitted over a communication network.
- the network can easily identify data boundaries (pictures, groups of pictures and here header Fet and sequence of bits Fnn), using the packet identification information provided by the network layer.
- step E24 In the "Byte-Stream" format, there are no specific packets and the construction of step E24 must make it possible to identify the boundaries of the relevant data (such as boundaries between parts of the stream corresponding to each image, and here between header Fet and sequence of bits Fnn) using additional means, such as the use of network abstraction layer units (or NAL units for “Network Abstraction Layer”), where unique combinations of bits (such as 0x00000001 ) are used to identify data boundaries).
- network abstraction layer units or NAL units for “Network Abstraction Layer”
- step E24 The complete data stream constructed in step E24 can then be transmitted in step E26 to the electronic decoding device 20 described below (by means of communication not shown and/or through at least one network of communication), or stored within the electronic coding device 2 (for subsequent transmission or, as a variant, subsequent decoding, for example within the electronic coding device itself, which is in this case designed to further implement the method decoding 20 described below with reference to Figure 6).
- This data stream thus comprises, as represented in FIG. 5, the header Fet and the sequence of bits Fnn.
- the Fet header includes:
- first part Fc which comprises data characteristic of the representation format of the audio or video content
- second part which comprises data R indicative of the global decoding network (which comprises a network of artificial decoding neurons and a network of artificial neurons for determining the context);
- FIG. 6 represents an electronic decoding device 20 using the entropy decoder 30, the artificial neural network for determining context 40 and the artificial neural network for decoding 28 (these elements having been introduced above with reference to FIG. 1 ).
- This electronic decoding device 20 comprises a reception unit 21, a processor 24 (for example a microprocessor) and a parallelized processing unit 26, for example a graphics processing unit (or GPU for "Graphical Processing Unit") or a unit tensor processing (or TPU for "Tensor Processing Unit”).
- a processor 24 for example a microprocessor
- a parallelized processing unit 26 for example a graphics processing unit (or GPU for "Graphical Processing Unit") or a unit tensor processing (or TPU for "Tensor Processing Unit”).
- the reception unit 21 is for example a communication circuit (such as a radiofrequency communication circuit) and makes it possible to receive data (and in particular here the data stream described above) from an external electronic device, such as than the electronic coding device 2, and to communicate these data to the processor 24 (to which the reception unit 21 is for example connected by a bus).
- a communication circuit such as a radiofrequency communication circuit
- the electronic decoding device 20 also comprises a storage unit 22, for example a memory (possibly a rewritable nonvolatile memory) or a hard disk.
- a storage unit 22 for example a memory (possibly a rewritable nonvolatile memory) or a hard disk.
- the storage unit 22 is represented in FIG. 5 as a separate element from the processor 24, the storage unit storage 22 could alternatively be integrated into (i.e. included in) processor 24.
- the processor 24 is in this case designed to successively execute a plurality of instructions of a computer program stored for example in the storage unit 22.
- control module 25 having in particular the functionalities described below.
- some of the functions of the control module 25 could be implemented due to the execution, by the processor 24, of instructions identified within the header Fet at step E52 as described below. below.
- Another part of the instructions stored in the storage unit 22 allow, when they are executed by the processor 24, to implement the entropy decoder 30 already mentioned.
- the entropy decoder 30 could be implemented due to the execution, by the processor 24, of instructions identified within the header Fet at step E52 as described below.
- the parallelized processing unit 26 is designed to implement the context determination artificial neural network 40 and the decoding artificial neural network 28 after having been configured by the processor 24 (here precisely by the control module 25) . To do this, the parallelized processing unit 26 is designed to perform in parallel at a given instant a plurality of operations of the same type.
- the context determination artificial neural network 40 and the decoding artificial neural network 28 together form an artificial neural network referred to here as the “global decoding network” and referenced 29 in FIG. 6.
- the processor 24 receives (here via the reception unit 21) the data stream comprising the header Fet and the sequence of bits Fnn.
- the decoding artificial neural network 28 is used in the context of processing data obtained by entropy decoding (by means of the entropy decoder 30) of the sequence of binary elements Fnn, this data processing aimed at obtaining audio or video content corresponding to the initial audio or video content B.
- the storage unit 22 can store a plurality of sets of parameters, each set of parameters defining a global decoding network (comprising a network of artificial neural context determination and a network of artificial decoding neural).
- the processor 24 can in this case configure the parallelized processing unit 26 by means of a particular set of parameters among these sets of parameters so that the parallelized processing unit 26 can then implement implements the artificial neural network (that is to say here the global decoding network) defined by this particular set of parameters.
- the storage unit 22 can in particular store a first set of parameters defining a first network of artificial neurons forming a random access decoder and/or a second set of parameters defining a second network of artificial neurons forming a low latency decoder.
- the electronic decoding device 20 in this case holds in advance decoding possibilities both for situations where it is desired to obtain random access to the content and for situations where it is desired to display the content without delay.
- the method of FIG. 7 begins a step E50 of reception (by the electronic decoding device 20, and precisely here by the reception unit 21) of the data stream comprising the header Fet and the sequence of binary elements Fnn.
- Receiver unit 21 transmits the received data stream to processor 24 for processing by control module 25.
- the control module 25 then proceeds to a step E52 of identifying the header Fet and the sequence of bits Fnn within the data stream received, for example by means of the start of sequence indicator binary elements (already mentioned during the description of step E24).
- the control module 25 can also identify in step E52 the different parts of the header Fet (as described above with reference to FIG. 5).
- control module 25 can launch at step E54 the execution of these executable instructions in order to implement at least some of the steps (described below) for processing the header data (and possibly entropy decoding).
- These instructions can be executed by the processor 24 or, as a variant, by a virtual machine instantiated within the electronic decoding device 20.
- the method of FIG. 7 continues with a step E56 of decoding data Fc characteristic of the format for representing the audio or video content so as to obtain characteristics of this format.
- the decoding of the Fc data makes it possible to obtain the dimensions (in pixels) of the images and/or the frame rate and/or the bit depth of the luminance information and/or the bit depth of the chrominance information.
- the control module 25 then proceeds to a step E58 of decoding the data R indicative of the global decoding network to be used.
- these data R are an identifier designating the global decoding network 28, for example within a predetermined set of artificial neural networks.
- This predetermined set is for example the set of global decoding networks accessible by the electronic decoding device 20, namely the set of global decoding networks for which the electronic decoding device 20 stores a set of parameters defining the network of artificial neurons concerned (as indicated above) or can have access to this set of parameters by connection to remote electronic equipment such as a server (as explained below).
- the control module 25 can in this case read, for example in the storage unit 22, a set of parameters associated with the decoded identifier (this set of parameters defining the global decoding network identified by the decoded identifier).
- control module 25 can send a request for a set of parameters intended for a remote server (this request including for example the decoded identifier) and receive in response the set of parameters defining the artificial neural network (forming here the global decoding network) identified by the decoded identifier.
- the set of parameters may in practice comprise certain parameters defining the decoding artificial neural network 28 and other parameters defining the context determination artificial neural network 40.
- the data R is descriptive data Rc of the global decoding network 29.
- these descriptive data are for example coded in accordance with a standard such as the MPEG-7 part 17 standard or in a format such as the JSON format.
- the decoding of these descriptive data makes it possible to obtain the parameters defining the global decoding network 29 to be used, comprising the artificial neural network for determining context 40 and the artificial neural network for decoding 28 (to which the data obtained by entropy decoding from the sequence of bits Fnn, as explained below).
- Such parameters may in practice comprise certain parameters defining the decoding artificial neural network 28 and other parameters defining the context determination artificial neural network 40.
- the use of the aforementioned first possibility or the second possibility depends on an indicator also included in the R data, as already indicated.
- the decoding of the data R indicative of the global decoding network to be used makes it possible (here to the control module 25) to determine in particular the characteristics of the artificial neural network of decoding 28.
- the control module 25 thus determines the number N of characteristic maps expected at the input of the decoding artificial neural network 28 and the dimensions H, W of these characteristic maps.
- the input layer of the decoding artificial neural network 28 corresponding to the output layer of the coding artificial neural network 8 as explained above, each element of a map of characteristics F is associated in a predetermined manner with an input node (or input layer node) of the decoding artificial neural network.
- the number and the dimensions of the maps of characteristics F are therefore linked to the characteristics of the decoding artificial neural network 28, as well as to certain header data such as the aforementioned data Fc (comprising in particular the dimensions of the image) .
- control module 25 then proceeds in step E60 to configure the parallelized processing unit 26 by means of the parameters defining the global decoding network 29 (parameters obtained in step E58), so that the processing unit parallelized processing 26 can implement this global decoding network 29 (comprising the context determination artificial neural network 40 and the decoding artificial neural network 28).
- This configuration step E60 notably comprises the instantiation of the global decoding network 29 (and therefore the instantiation of the artificial neural network for determining context 40 and of the artificial neural network for decoding 28) within the processing unit parallelized 26, here using the parameters obtained in step E58.
- This instantiation may include the following steps:
- the configuration step E60 can further comprise the application of predefined (initial) values (stored for example in the storage unit 22) on the input layer of the global decoding network 29 so that the global decoding network decoding 29 is activated and thus produces at output (precisely at the output of the context determination artificial neural network 40) an initial context index C.
- control module 25 then proceeds in step E64 to decoding the parameterization data Init relating respectively to the different contexts of the set of contexts used (this set being determined thanks to the information item 11 decoded in step E62).
- the control module 25 can then implement a step E66 for initializing each usable context within the entropy decoder 30 by means of the parameterization data linit of the context concerned (decoded in step E64).
- each context is initialized with the probability model defined by the parameterization datum linit relating to this context.
- control module 25 configures in step E66 each context usable by the entropy decoder 30 with the probability model defined by the relative parameterization data linit to this context.
- the control module 25 then applies (step E70) at the input of the entropy decoder 30 the sequence of binary elements Fnn (received via the reception unit 21), while setting the parameters of the entropy decoder 30 in the context identified by the index of context C produced at the output of the global decoding network 29 (that is to say here produced at the output of the context determination artificial neural network 40).
- the context index C is for example the aforementioned initial context index C, produced due to the early activation of global decoding network 29 during step E60 as explained above.
- the entropy decoder 30 can be parameterized in a predefined initial context, possibly stored in the storage unit 22 (step E60 does not include in this case the application of predefined values to the global decoding network 29 for early activation of the latter).
- the entropy decoder 30 is therefore parameterized in the context identified by the context index C produced by the global decoding network 29 when representative values V, decoded (by entropy decoding) during the previous passage to step E70, have been applied at the input of the global decoding network 29 (during the previous going to step E72).
- the entropy decoder 30 thus produces in step E70 new representative values V (by entropy decoding of the sequence of binary elements Fnn).
- a synchronization mechanism can be provided between the global decoding network 29 and the entropy coder 30 (in order to guarantee that the context index C to be used is available at the moment when the entropy decoding of the corresponding representative value V takes place ).
- This synchronization mechanism comprises for example the suspension of the entropy decoding (by the entropy decoder 30) as long as a new context index C is not available at the output of the global decoding network 29 (that is to say precisely at the output of the context determination artificial neural network 40).
- an intermediate variable C′ is stored in memory (for example in a register of the processor 24 or in the storage unit 22) and updated by the global decoding network 29 (i.e. i.e. precisely by the context-determining artificial neural network 40).
- the global decoding network 29 i.e. i.e. precisely by the context-determining artificial neural network 40.
- an entropy decoding of one (or a predetermined number of) representative value(s) V is carried out with the context defined by the intermediate variable C' (identical to the context index C produced by the global decoding network 29), then the entropic decoding is again suspended (until a new update of the intermediate variable C′ following the production of a new context index C at the output of the global decoding network 29, that is to say at the output of the context determination artificial neural network 40).
- the synchronization mechanism consists in making the progress of the entropic decoding (by the decoder entropy 30) of the production of the context index C by the global decoding network 29 (that is to say precisely by the context determination artificial neural network 40).
- information for example the context index C itself or, as a variant, dedicated synchronization information
- the global decoding network 29 for example from the artificial neural network context determination 40
- the entropy decoder 30 when the part of the neural network (for example the layer of the neural network) which supplies the context index C to be used is activated.
- the decoding entropy of the representative values V by the entropy decoder 30 is achieved by parametrizing the entropy decoder 30 at each instant in the context defined by the context index C associated with the set containing the representative value V to be obtained by entropy decoding at this instant .
- the entropy decoder 30 allows the successive (entropic) decoding of the different characteristic maps F and the control module 25 can therefore (at each instant) parameterize the entropy decoder 30 in the context defined by the context index C associated with the map of characteristics F during entropy decoding.
- the processor 24 (here directly at the output of the entropy decoder 30 or, as a variant, via the control module 25) can then apply (/.e. present) in step E72 the representative values V to the neural network (global decoding network) 29 implemented by the parallelized processing unit 26 so that on the one hand these representative values V are processed by a decoding process using at least in part the artificial neural decoding network 28 and on the other hand that a (new) context index C is produced at the output of the artificial neural network for determining context C.
- the decoding artificial neural network 28 receives the representative values V as input and produces as output a representation I of the coded content suitable for reproduction on an audio or video reproduction device.
- the representative values V are applied to the input layer of the decoding artificial neural network 28 and the output layer of the decoding artificial neural network 28 produces the representation I of the aforementioned encoded content.
- the artificial neural network 28 thus produces as output (that is to say at the level of its output layer) at least one representation matrix I of an image.
- the decoding artificial neural network 28 can receive as input at least some of the data produced at the output of the neural network artificial decoding devices 28 during the processing of previous data (here of previous representative values V), corresponding for example to the previous block or to the previous image. In this case, a step E74 of reinjecting data produced at the output of the decoding artificial neural network 28 into the input of the decoding artificial neural network 28 is carried out.
- the control module 25 determines in step E76 whether the processing of the sequence of bits Fnn is finished.
- step E70 the method loops to step E70 for entropy decoding of the following part of the sequence of binary elements Fnn and application of other representative values V (produced by this entropy decoding) to the network of artificial decoding neurons 28.
Landscapes
- Engineering & Computer Science (AREA)
- Multimedia (AREA)
- Signal Processing (AREA)
- Theoretical Computer Science (AREA)
- Computing Systems (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Compression Or Coding Systems Of Tv Signals (AREA)
- Compression, Expansion, Code Conversion, And Decoders (AREA)
Abstract
Description
Claims
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| FR2010214A FR3114933B1 (fr) | 2020-10-06 | 2020-10-06 | Procédé et dispositif électronique de décodage d’un flux de données, et programme d’ordinateur associé |
| PCT/EP2021/077157 WO2022073885A1 (fr) | 2020-10-06 | 2021-10-01 | Procédé et dispositif électronique de décodage d'un flux de données, et programme d'ordinateur associé |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| EP4226614A1 true EP4226614A1 (fr) | 2023-08-16 |
Family
ID=73793423
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| EP21786467.7A Pending EP4226614A1 (fr) | 2020-10-06 | 2021-10-01 | Procédé et dispositif électronique de décodage d'un flux de données, et programme d'ordinateur associé |
Country Status (7)
| Country | Link |
|---|---|
| US (2) | US12445659B2 (fr) |
| EP (1) | EP4226614A1 (fr) |
| JP (1) | JP2023545042A (fr) |
| KR (1) | KR20230081719A (fr) |
| CN (1) | CN116601954A (fr) |
| FR (1) | FR3114933B1 (fr) |
| WO (1) | WO2022073885A1 (fr) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| FR3124342B1 (fr) * | 2021-06-17 | 2024-01-12 | Fond B Com | Procédés et dispositifs de décodage d’une partie au moins d’un flux de données, programme d’ordinateur et flux de données associés |
Family Cites Families (13)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9313514B2 (en) * | 2010-10-01 | 2016-04-12 | Sharp Kabushiki Kaisha | Methods and systems for entropy coder initialization |
| US10963782B2 (en) * | 2016-11-04 | 2021-03-30 | Salesforce.Com, Inc. | Dynamic coattention network for question answering |
| JP6956803B2 (ja) * | 2017-05-26 | 2021-11-02 | グーグル エルエルシーGoogle LLC | ニューラルネットワークを使用するタイル型画像圧縮 |
| WO2019009449A1 (fr) * | 2017-07-06 | 2019-01-10 | 삼성전자 주식회사 | Procédé et dispositif de codage/décodage d'image |
| JP6994868B2 (ja) * | 2017-08-09 | 2022-01-14 | パナソニック インテレクチュアル プロパティ コーポレーション オブ アメリカ | 符号化装置、復号装置、符号化方法、および復号方法 |
| EP3562162A1 (fr) * | 2018-04-27 | 2019-10-30 | InterDigital VC Holdings, Inc. | Procédé et appareil de codage et de décodage vidéo basés sur la mise en uvre d'un réseau neuronal de cabac |
| US10886943B2 (en) * | 2019-03-18 | 2021-01-05 | Samsung Electronics Co., Ltd | Method and apparatus for variable rate compression with a conditional autoencoder |
| KR20250117730A (ko) * | 2019-03-18 | 2025-08-05 | 프라운호퍼 게젤샤프트 쭈르 푀르데룽 데어 안겐반텐 포르슝 에. 베. | 신경망의 매개변수를 압축하기 위한 방법 및 장치 |
| CN110728726B (zh) * | 2019-10-24 | 2022-09-23 | 湖南大学 | 一种基于用户交互与深度神经网络的图像压缩方法 |
| FR3112662B1 (fr) | 2020-07-17 | 2024-12-13 | Fond B Com | Procédé et dispositif électronique de décodage d’un flux de données, et flux de données associé |
| FR3112661B1 (fr) | 2020-07-17 | 2024-02-02 | Fond B Com | Procédés de décodage d’un flux de données, dispositifs et flux de données associés |
| FR3114717B1 (fr) | 2020-09-30 | 2023-10-13 | Fond B Com | Procédé et dispositif électronique de décodage d’un flux de données, programme d’ordinateur et flux de données associés |
| CN113658130B (zh) * | 2021-08-16 | 2023-07-28 | 福州大学 | 基于双重孪生网络的无参考屏幕内容图像质量评估方法 |
-
2020
- 2020-10-06 FR FR2010214A patent/FR3114933B1/fr active Active
-
2021
- 2021-10-01 US US18/248,023 patent/US12445659B2/en active Active
- 2021-10-01 CN CN202180081925.1A patent/CN116601954A/zh active Pending
- 2021-10-01 WO PCT/EP2021/077157 patent/WO2022073885A1/fr not_active Ceased
- 2021-10-01 JP JP2023521083A patent/JP2023545042A/ja active Pending
- 2021-10-01 EP EP21786467.7A patent/EP4226614A1/fr active Pending
- 2021-10-01 KR KR1020237015366A patent/KR20230081719A/ko active Pending
-
2025
- 2025-09-18 US US19/332,585 patent/US20260019650A1/en active Pending
Also Published As
| Publication number | Publication date |
|---|---|
| CN116601954A (zh) | 2023-08-15 |
| US20260019650A1 (en) | 2026-01-15 |
| WO2022073885A1 (fr) | 2022-04-14 |
| US12445659B2 (en) | 2025-10-14 |
| KR20230081719A (ko) | 2023-06-07 |
| JP2023545042A (ja) | 2023-10-26 |
| FR3114933A1 (fr) | 2022-04-08 |
| FR3114933B1 (fr) | 2023-10-13 |
| US20230379506A1 (en) | 2023-11-23 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| EP4183130B1 (fr) | Decodage video utilisant un reseau de neurones | |
| WO2022069437A1 (fr) | Procédé et dispositif électronique de décodage d'un flux de données, programme d'ordinateur et flux de données associés | |
| EP2700226B1 (fr) | Procedes et appareils de production et de traitement de representations de scenes multimedias | |
| WO2022013247A1 (fr) | Decodage video parallelise utilisant un reseau de neurones | |
| WO2024121109A1 (fr) | Procédé et dispositif de codage et décodage d'images | |
| WO2022073885A1 (fr) | Procédé et dispositif électronique de décodage d'un flux de données, et programme d'ordinateur associé | |
| EP2368367A1 (fr) | Système et procédé interactif pour la transmission sur un réseau bas débit d'images clefs sélectionnées dans un flux video | |
| JP2024520961A (ja) | データストリームの少なくとも一部を復号化するための方法と装置、コンピュータプログラム、及び関連するデータストリーム | |
| FR2923970A1 (fr) | Procede et dispositif de formation, de transfert et de reception de paquets de transport encapsulant des donnees representatives d'une sequence d'images | |
| WO2024260628A1 (fr) | Procédé et dispositif de codage et décodage de séquences d'images | |
| WO2024260629A1 (fr) | Procédé et dispositif de codage et décodage de séquences d'images | |
| WO2026003293A1 (fr) | Procédé et dispositif de codage et décodage d'un signal | |
| WO2024121107A1 (fr) | Procédé et dispositif de codage et décodage d'images. | |
| WO2026003294A1 (fr) | Procédé et dispositif de codage et décodage d'un signal | |
| FR3143246A1 (fr) | Procédé et dispositif de codage et décodage d’images. | |
| FR3143247A1 (fr) | Procédé et dispositif de codage et décodage de séquences d’images. | |
| WO2016097556A1 (fr) | Procédé de codage d'une image numérique, procédé de décodage, dispositifs, terminal d'utilisateur et programmes d'ordinateurs associes |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: UNKNOWN |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: THE INTERNATIONAL PUBLICATION HAS BEEN MADE |
|
| PUAI | Public reference made under article 153(3) epc to a published international application that has entered the european phase |
Free format text: ORIGINAL CODE: 0009012 |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: REQUEST FOR EXAMINATION WAS MADE |
|
| 17P | Request for examination filed |
Effective date: 20230406 |
|
| AK | Designated contracting states |
Kind code of ref document: A1 Designated state(s): AL AT BE BG CH CY CZ DE DK EE ES FI FR GB GR HR HU IE IS IT LI LT LU LV MC MK MT NL NO PL PT RO RS SE SI SK SM TR |
|
| DAV | Request for validation of the european patent (deleted) | ||
| DAX | Request for extension of the european patent (deleted) | ||
| RAP1 | Party data changed (applicant data changed or rights of an application transferred) |
Owner name: ORANGE |
|
| STAA | Information on the status of an ep patent application or granted ep patent |
Free format text: STATUS: EXAMINATION IS IN PROGRESS |