Base anchored models and inference for the compression and upsampling of video and multiview imagery

A method for coding displacement fields within a video sequence comprising video frames, wherein the video frames are assigned to Groups of Pictures known as GOPs, the method comprising:

• coding a base displacement model for a GOP, the base displacement model describing a displacement that carries each location in a designated base frame of the GOP to a corresponding location in each other frame of the GOP, and other displacement relationships between the frames of the GOP are inferred from the base displacement model,
• wherein the base displacement model is described via a deformable base mesh whose nodes are assigned one displacement vector for each non-base frame in the GOP, and
• wherein breaks in the base displacement model are covered by a subset of elements of the base mesh, said subset of elements being break-induced mesh elements, said break-induced mesh elements being located between nodes, where the nodes of the break-induced mesh elements appear in co-located pairs, one node in each pair carrying displacement vectors associated with one side of the break, while the other carries said displacement vectors associated with the other side of the break.

A method for displacement compensated prediction of certain image frames from other frames, for frames in a video and/or Multiview sequence, wherein the frames are assigned to groups of pictures (GOPs), comprising:

• providing a base displacement model for each GOP, describing a displacement that carries each location in a designated base frame of a GOP to a corresponding location in each other frame of the GOP, said base displacement model being used to infer displacement relationships between the frames of the GOP, and the inferred displacement field at a prediction target frame being used to predict that frame from one or more other frames in the GOP,
• wherein the base displacement model is described via a deformable base mesh whose nodes are assigned one displacement vector for each non-base frame in the GOP, and
• wherein breaks in the base displacement model are covered by a subset of elements of the base mesh, said subset of elements being break-induced mesh elements, said break-induced mesh elements being located between nodes, where the nodes of the break-induced mesh elements appear in co-located pairs, one node in each pair carrying displacement vectors associated with one side of the break, while the other carries said displacement vectors associated with the other side of the break.