Abstract: A method for processing a multimedia content previously digitally encoded into an audio-video stream and transmitted from a distribution platform for the purposes of on-demand transmission to a plurality of receiving terminals includes the following steps: applying a demultiplexing function to the audio-video stream, said function being arranged to generate at least one group of data streams; applying a transformation to each data stream; segmenting each data stream; storing the at least one group of segmented data streams in a memory; carrying out, only in response to a request issued by a receiving terminal among the plurality of receiving terminals, the following steps: selecting, on the basis of the request, at least one data stream of at least one group of stored data streams; and transmitting the at least one selected data stream to the receiving terminal that issued the request.

Abstract: A method for distributing an audiovisual content to a terminal is disclosed. The content is received by the terminal in the form of a succession of consecutive segments, each segment being distributed to the terminal following a transmission of a request by the terminal and being obtained by an application of a workflow to a portion of the content.

Abstract: An elementary module of a workflow of an audiovisual content distribution system, each content being received by a terminal in the form of a succession of segments, each segment being distributed to the terminal following transmission of a request by terminal and being obtained by an application of a workflow to a portion of content. The elementary module executes a processing of a predefined type and comprises: a variable plurality of processing units available for executing said processing; a scaling module, able to determine, using a first model, a number of processing units to be allocated for an implementation of a set of processing operations requested of said elementary module; and a load management module able to choose, using a second model, for each processing operation requested, one processing unit among processing units allocated by the scaling module for performing processing, each model being a neural network of the deep learning type.

Abstract: An elementary module of a workflow of a system for distributing audiovisual contents is disclosed. Content is received as a succession of segments which are distributed following a transmission of a request and are obtained by an application of a workflow to a portion of a content. The elementary module executes a processing operation of a predefined type of the workflow and comprises: a variable plurality of processing units available for executing the processing operation of the predefined type, the elementary module regularly obtaining information representing each processing unit added to or removed from the plurality, an allocation module able to manage allocations or disallocations of processing units of the plurality according to requests for first data received; and a load management module able to choose, for each request received, at least one processing unit among the processing units allocated for processing second data in order to obtain first data requested.

Abstract: An elementary module of a workflow of an audiovisual content distribution system, each content being received by a terminal in the form of a succession of segments, each segment being distributed to the terminal following transmission of a request by terminal and being obtained by an application of a workflow to a portion of content. The elementary module executes a processing of a predefined type and comprises: a variable plurality of processing units available for executing said processing; a scaling module, able to determine, using a first model, a number of processing units to be allocated for an implementation of a set of processing operations requested of said elementary module; and a load management module able to choose, using a second model, for each processing operation requested, one processing unit among processing units allocated by the scaling module for performing processing, each model being a neural network of the deep learning type.

Abstract: An elementary module of a workflow of a system for distributing audiovisual contents is disclosed. Content is received as a succession of segments which are distributed following a transmission of a request and are obtained by an application of a workflow to a portion of a content. The elementary module executes a processing operation of a predefined type of the workflow and comprises: a variable plurality of processing units available for executing the processing operation of the predefined type, the elementary module regularly obtaining information representing each processing unit added to or removed from the plurality, an allocation module able to manage allocations or disallocations of processing units of the plurality according to requests for first data received; and a load management module able to choose, for each request received, at least one processing unit among the processing units allocated for processing second data in order to obtain first data requested.

Abstract: A method for distributing an audiovisual content to a terminal is disclosed. The content is received by the terminal in the form of a succession of consecutive segments, each segment being distributed to the terminal following a transmission of a request by the terminal and being obtained by an application of a workflow to a portion of the content.

Abstract: A method for processing a multimedia content previously digitally encoded into an audio-video stream and transmitted from a distribution platform for the purposes of on-demand transmission to a plurality of receiving terminals includes the following steps: applying a demultiplexing function to the audio-video stream, said function being arranged to generate at least one group of data streams; applying a transformation to each data stream; segmenting each data stream; storing the at least one group of segmented data streams in a memory; carrying out, only in response to a request issued by a receiving terminal among the plurality of receiving terminals, the following steps: selecting, on the basis of the request, at least one data stream of at least one group of stored data streams; and transmitting the at least one selected data stream to the receiving terminal that issued the request.

Abstract: The method includes determining at least one candidate motion vector associated with a neighboring block of said current block. A prediction motion vector is determined from the candidate motion vector, and the current block is coded from the prediction motion vector. The prediction motion vector is determined according by determining, for the candidate motion vector, a corrective motion vector so as to minimize a distortion calculated between the neighoboring block successively coded and reconstructed and a prediction block motion compensated by the at least one candidate motion vector modified by the corrective motion vector. The prediction motion vector is then determined from the candidate motion vector modified by the corrective motion vector.

Abstract: A spatial prediction method of a block of pixels of an image, called current block, using a plurality of spatial prediction modes based on neighboring pixels of the current block is disclosed. The method comprises determining a prediction block of the current block according to each one of said spatial prediction modes and selecting one of said prediction blocks according to a predetermined criterion. According to the invention, before carrying out the preceding steps, it is detected, among said spatial prediction modes, whether some of them are redundant. If such redundant modes are detected, one of them is replaced by an additional mode, distinct from the other spatial prediction modes.

Abstract: A method for decoding a stream of coded data to reconstruct a current block of a sequence of images is disclosed. The method comprises the steps of: reconstructing for the current block a first item of motion data from a stream of coded data, identifying a first prediction block with the first item of motion data, determining a second item of motion data by template matching, identifying a second prediction block with the second item of motion data, reconstructing the current block from the first and second prediction blocks. The step of determination of the second item of motion data is a function of the first distortion calculated between the first prediction block and the second prediction block.

Abstract: A method for coding a current block using a parametric motion model whose parameters are defined from the motion data of the causal neighborhood is disclosed. Specifically, the method for coding the current block comprises the following steps: determine a prediction block of the current block from a motion model defined by a set of parameters, the model associating a motion vector pointing at a pixel of a prediction block with each pixel of the current block. calculate a residual block between the current block and the prediction block, and code the residual block. According to the invention, the parameters of the motion model are determined from the motion data associated with the causal neighborhood blocks of the current block. Therefore, they can be deducted from parts of images already coded and do not need to be incorporated in the image data stream.

Abstract: The invention proposes a method for reconstructing a self-similar textured region of an image. Said method comprises determining pixels of a part of the self-similar textured region by copying sample pixels from a sample part of the self-similar textured region, the sample pixels being selected using a neighborhood matching, wherein a size of neighborhoods used for matching is selected based on an analysis of descriptors computed from coefficients of OCT transform of differently sized blocks of the sample part. The analysis of descriptors computed from coefficients of DCT transform of differently sized blocks of the sample part allows for determining the neighborhood size close to a feature size of the texture.

Abstract: A method for coding a block of pixels is disclosed. The method comprises the following steps: determining, for each pixel of the current block, a prediction pixel by thresholding, with a current threshold value, of coefficients resulting from a transformation applied on a window covering at least the pixel of the current block and by inverse transformation applied to the thresholded coefficients, extracting from the current block a prediction block formed of prediction pixels to generate a residue block, and coding said residue block, According to the invention, the current threshold value is determined or coded from neighboring reconstructed pixels of the current block.

Abstract: A method for predicting a current block belonging to a current image with a view to the coding or the reconstruction of the current block. The method includes determining for the current block a reference block from at least one motion data previously associated with the current block. The fulfillment of a predefined condition is is verified. The predefined condition being representative of a temporal variation between the image data of the reference block and the image data of a prediction block of the reference block previously determined from at least one motion data associated with the reference block. The motion data is modified according to the result of the verification, and a final prediction block for the current block is generated from the motion data of the current block.

Abstract: The invention relates to a method for deriving motion data for a macroblock divided in elementary blocks of a high resolution picture, called high layer macroblock, from motion data of macroblocks of a low resolution picture, called base layer macroblock.

Abstract: The invention is made in the field of image block encoding and reconstruction. A method for reconstructing a current block of an image is described, the current block being encoded using predictive encoding of segments resulting from geometry adaptive block partitioning of said current block. After decoding encoded data comprising at least one flag and motion vectors, wherein blocks referenced by the motion vectors are used for prediction of segments of the current block, the at least one flag and an encoding order of the motion vectors in the encoded data is used for determining which of the referenced blocks exhibits a geometry that matches the geometry of the current block and for determining a segment of the determined block, wherein prior to determining the segment, the determined block is geometry adaptive block partitioned into segments from which the segment is determined for predicting one of the segments of the current block.

Abstract: A method for coding a block of pixels is described. The coding method comprises: determining a prediction block for the pixel block from a motion vector; calculating a residue between the pixel block and the prediction block; and code the residue. The determination of the prediction block comprises the following steps: determining an intermediate prediction block of size strictly greater than the size of the pixel block; transforming the intermediate prediction block into a first block transformed with a first transform; and transforming the first transformed block into a second block transformed with a second transform of inverse the first transform whose basis functions are shifted by at least a part of each component of the motion vector, the prediction block being extracted from the second transformed block.

Abstract: A high dynamic picture transmission system is provided. The transmission system includes a coding unit configured to generate a standard bitstream and at least a second bitstream. The standard bitstream coding the pictures such that the luminance of each pixel is coded with a standard dynamic, and the second bitstream contains the information necessary to reconstruct the luminance of high dynamic pictures from the coded luminance with the standard dynamic contained in the standard bitstream.

Abstract: A spatial prediction method of a block of pixels of an image, called current block, using a plurality of spatial prediction modes based on neighbouring pixels of the current block is disclosed. The method comprises determining a prediction block of the current block according to each one of said spatial prediction modes and selecting one of said prediction blocks according to a predetermined criterion. According to the invention, before carrying out the preceding steps, it is detected, among said spatial prediction modes, whether some of them are redundant. If such redundant modes are detected, one of them is replaced by an additional mode, distinct from the other spatial prediction modes.