A method and wearable camera for encoding video captured by a wearable camera determines a centre of rotation for an image frame to be encoded. The centre of rotation relates to a rotation of the wearable camera at the time of capturing the video and the image frame comprises multiple groups of pixels. Furthermore, compression levels are set for the multiple groups of pixels of the image frame. The compression levels for the multiple groups of pixels of the image frame are set such that a level of compression increases with a radial distance from the centre of rotation. The image frame is encoded using the compression levels.

A system and methods for a CODEC driving a real-time light field display for multi-dimensional video streaming, interactive gaming and other light field display applications is provided applying a layered scene decomposition strategy. Multi-dimensional scene data is divided into a plurality of data layers of increasing depths as the distance between a given layer and the plane of the display increases. Data layers are sampled using a plenoptic sampling scheme and rendered using hybrid rendering, such as perspective and oblique rendering, to encode light fields corresponding to each data layer. The resulting compressed, (layered) core representation of the multi-dimensional scene data is produced at predictable rates, reconstructed and merged at the light field display in real-time by applying view synthesis protocols, including edge adaptive interpolation, to reconstruct pixel arrays in stages (e.g. columns then rows) from reference elemental images.

A system and methods for a CODEC driving a real-time light field display for multi-dimensional video streaming, interactive gaming and other light field display applications is provided applying a layered scene decomposition strategy. Multi-dimensional scene data is divided into a plurality of data layers of increasing depths as the distance between a given layer and the plane of the display increases. Data layers are sampled using a plenoptic sampling scheme and rendered using hybrid rendering, such as perspective and oblique rendering, to encode light fields corresponding to each data layer. The resulting compressed, (layered) core representation of the multi-dimensional scene data is produced at predictable rates, reconstructed and merged at the light field display in real-time by applying view synthesis protocols, including edge adaptive interpolation, to reconstruct pixel arrays in stages (e.g. columns then rows) from reference elemental images.

To represent a 3D scene, the MPI format uses a set of fronto-parallel planes. Different from MPI, the current MIV standard accepts a 3D scene represented as sequence input pairs of texture and depth pictures as input. To enable transmission of an MPI cube via the MIV-V3C standard, in one embodiment, an MPI cube is divided into empty regions and local MPI partitions that contain 3D objects. Each partition in the MPI cube can be projected to one or more patches. For a patch, the geometry is generated as well as the texture attribute and alpha attributes, and the alpha attributes may be represented as a peak and a width of an impulse. In another embodiment, an MPI RGBA layer of the MPI is cut into sub-images. Each sub-image may correspond to a patch, and the RGB and alpha information of the sub-image are assigned to the patch.

A method, system and computer program product for segmenting generic foreground objects in images and videos. For segmenting generic foreground objects in videos, an appearance stream of an image in a video frame is processed using a first deep neural network. Furthermore, a motion stream of an optical flow image in the video frame is processed using a second deep neural network. The appearance and motion streams are then joined to combine complementary appearance and motion information to perform segmentation of generic objects in the video frame. Generic foreground objects are segmented in images by training a convolutional deep neural network to estimate a likelihood that a pixel in an image belongs to a foreground object. After receiving the image, the likelihood that the pixel in the image is part of the foreground object as opposed to background is then determined using the trained convolutional deep neural network.

A system includes at least one imaging sensor and a processor. The processor is configured to acquire detected data describing an environment of an autonomous vehicle using the imaging sensor; derive reference data which describes the environment from a predefined map; compute difference data representing a difference between the detected data and the reference data; and transfer the difference data, wherein an image computed based on the difference data and the reference data represents the detected data. Other embodiments are also described.

A method of coding at least one image comprising the steps of splitting the image into a plurality of blocks, of grouping said blocks into a predetermined number of subsets of blocks, of coding each of said subsets of blocks in parallel, the blocks of a subset considered being coded according to a predetermined sequential order of traversal. The coding step comprises, for a current block of a subset considered, the sub-step of predictive coding of said current block with respect to at least one previously coded and decoded block, and the sub-step of entropy coding of said current block on the basis of at least one probability of appearance of a symbol.

A method of coding at least one image comprising the steps of splitting the image into a plurality of blocks, of grouping said blocks into a predetermined number of subsets of blocks, of coding each of said subsets of blocks in parallel, the blocks of a subset considered being coded according to a predetermined sequential order of traversal. The coding step comprises, for a current block of a subset considered, the sub-step of predictive coding of said current block with respect to at least one previously coded and decoded block, and the sub-step of entropy coding of said current block on the basis of at least one probability of appearance of a symbol.

A method and wearable camera for encoding video captured by a wearable camera determines a centre of rotation for an image frame to be encoded. The centre of rotation relates to a rotation of the wearable camera at the time of capturing the video and the image frame comprises multiple groups of pixels. Furthermore, compression levels are set for the multiple groups of pixels of the image frame. The compression levels for the multiple groups of pixels of the image frame are set such that a level of compression increases with a radial distance from the centre of rotation. The image frame is encoded using the compression levels.

One or more methods and systems are provided for encoding, decoding and playback of a video content in a client-server architecture. The invention proposes a video encoding and decoding method that includes identification of activities in the video content, identification of corresponding API's with related parameters corresponding to activity and storing those API's along with base frame and object frame in the database. In this invention, animation API functions are created for unknown/random activities. The playback involves decoding the data, which is a set of instructions to play the animation with given objects and base frames, and animating object frame over base frame using said API functions.