Method and apparatus for encoding and decoding light field based image, and corresponding computer program product

A device for encoding at least one block of pixels of a view belonging to a matrix of views obtained from light-field data associated with a scene, wherein said device comprises a processor configured to:

• obtain a predicted block of pixels, and for each pixel to predict of said block of pixels:
• obtain, from said matrix of views, at least one epipolar plane image (EPI) to which said at least one pixel to predict belongs,
• determine, among a set of bidirectional prediction modes, at least one optimal bidirectional prediction mode from a set of previous reconstructed pixels neighbouring said at least one pixel to predict in said at least one epipolar plane image,
• obtain a prediction value of said at least one pixel to predict by performing an interpolation using said at least one optimal bidirectional prediction mode,
• provide at least one group of pixels to predict within said block of pixels, said group of pixels comprising at least two pixels of a same line, of a same column or of a set of at least two pixels presenting an angular orientation with respect to a line or a column of said block of pixels, and
• wherein determining said optimal bidirectional prediction mode comprises:
• for at least one pixel to predict of said group of pixels, apply each predetermined bidirectional prediction mode of said set of bidirectional prediction modes to said set of previous reconstructed pixels neighbouring said at least one pixel to predict in said at least one epipolar plane image to obtain an energy level for each bidirectional prediction mode,
• select a same optimal bidirectional prediction mode for all pixels to predict belonging to said group of pixels, said optimal bidirectional prediction mode corresponding to the bidirectional prediction mode of which the energy level is the argument of the minimum of energy levels obtained for at least one pixel to predict of said group of pixels,
• determine a residual error corresponding to the difference between said at least one block of pixels and said predicted block of pixels,
• encode said residual error associated with said block of pixels.

A device for decoding a signal representing at least one block of pixels of a view belonging to a matrix of views obtained from light-field data associated with a scene, wherein said device comprises a processor configured to:

• decode a residual error associated to said block of pixels,
• predict said at least one block of pixels, to obtain a predicted block of pixels, and for at least one pixel to predict of said block of pixels:
• obtain, from said matrix of views, at least one epipolar plane image (EPI) to which said at least one pixel to predict belongs,
• determine, among a set of bidirectional prediction modes, at least one optimal bidirectional prediction mode from a set of previous reconstructed pixels neighbouring said at least one pixel to predict in said at least one epipolar plane image,
• obtain a prediction value of said at least one pixel to predict by performing an interpolation using said at least one optimal bidirectional prediction mode,
• provide at least one group of pixels to predict within said block of pixels, said group of pixels comprising at least two pixels of a same line, of a same column or of a set of at least two pixels presenting an angular orientation with respect to a line or a column of said block of pixels, and
• wherein determining said optimal bidirectional prediction mode comprises:
• for at least one pixel to predict of said group of pixels, apply each predetermined bidirectional prediction mode of said set of bidirectional prediction modes to said set of previous reconstructed pixels neighbouring said at least one pixel to predict in said at least one epipolar plane image to obtain an energy level for each bidirectional prediction mode,
• select a same optimal bidirectional prediction mode for all pixels to predict belonging to said group of pixels, said optimal bidirectional prediction mode corresponding to the bidirectional prediction mode of which the energy level is the argument of the minimum of energy levels obtained for at least one pixel to predict of said group of pixels,
• reconstruct said at least one block of pixels by adding said residual error to said predicted block of pixels.