Flexible tile partitions
An encoder, a decoder, and methods for partitioning a picture from a sequence of video pictures into a layout having a plurality of flexible tiles or segments is disclosed. Each tile or segment comprises a single rectangular or square region. The encoder generates a partition structure and encodes the tiles according to the partition structure. The encoder also generates a bitstream comprising a plurality of coded picture segments and information indicating the partition structure used to partition the picture into the plurality of flexible tiles or segments, and sends the bitstream to the decoder. Upon receipt, the decoder uses the coded picture segments and information in the bitstream to decode the plurality of coded picture segments.
1 . A method of decoding a picture, the method implemented by a decoder and comprising:
receiving a bitstream comprising a plurality of coded tiles and information associated with the plurality of coded tiles;
deriving, from the information in the bitstream, a spatial partition structure that divides the picture into tiles:
wherein each tile corresponds to one of the plurality of coded tiles and comprises at least one unit;
wherein the information comprises both a height and a width of a first tile; and
wherein the information indicates that the decoder should copy a height, a width, or both the height and the width of a tile of the spatial partition structure from an adjacent tile of the spatial partition structure; and
decoding the plurality of coded tiles received in the bitstream in accordance with the spatial partition structure, wherein derivation of an intra prediction mode for a first unit of the first tile depends on an intra prediction mode for a second unit of the first tile and is independent of intra prediction modes for units of any other tile of the spatial partition structure.
2 . The method of claim 1 , wherein the deriving, from the information in the bitstream, further comprises deriving a quantization parameter value for the first unit based on a quantization parameter value for the second unit, wherein derivation of the quantization parameter value for the first unit is independent of quantization parameter values for units of any other tile of the spatial partition structure.
3 . The method of claim 1 , wherein deriving the spatial partition structure comprises receiving an indication that an encoder is using flexible partitioning.
4 . The method of claim 1 , wherein decoding the plurality of coded tiles comprises decoding each tile according to a predefined decoding order.
5 . The method of claim 1 , wherein each unit of the spatial partition structure is a coding tree unit (CTU).
6 . The method of claim 1 , wherein the information comprises an integer number specifying the height or the width with respect to a unit of the spatial partition structure.
7 . The method of claim 1 , wherein each unit of the spatial partition structure is a region of the picture, and the information comprises a single size for all the units of the spatial partition structure.
8 . The method of claim 1 , wherein the spatial partition structure divides the picture by arranging corners of the tile on a grid with a start corner and an end corner on the grid defined for each tile.
9 . A method of encoding a picture, the method implemented by an encoder and comprising:
defining a spatial partition structure that divides the picture into a plurality of tiles, wherein each tile comprises at least one unit;
encoding the plurality of tiles in accordance with the spatial partition structure to generate a plurality of coded tiles, wherein each coded tile corresponds to one of the tiles of the spatial partition structure and is independent, such that derivation of an intra prediction mode for a first unit of a first tile depends on an intra prediction mode for a second unit of the first tile and is independent of intra prediction modes for units of any other tiles of the spatial partition structure;
generating a bitstream comprising the plurality of coded tiles and information indicating the spatial partition structure that divides the picture into the plurality of tiles, wherein:
the information comprises both a height and a width of the first tile; and
the information indicates that a decoder should copy a height, a width, or both the height and the width of a tile of the spatial partition structure from an adjacent tile of the spatial partition structure; and
transmitting the bitstream.
10 . The method of claim 9 , wherein generating a bitstream further comprises including an indication in the bitstream that an encoder is using flexible partitioning.
11 . The method of claim 9 , further comprising determining a size for a unit in the spatial partition structure, wherein a unit defines a region of the picture.
12 . The method of claim 9 , wherein each unit of the spatial partition structure is a coding tree unit (CTU).
13 . The method of claim 9 , wherein each unit of the spatial partition structure is a region of the picture, and the information comprises a single size for all the units of the spatial partition structure.
14 . The method of claim 9 , wherein the spatial partition structure divides the tiles of the spatial partition structure by arranging corners of the tile on a grid with a start corner and an end corner on the grid defined for each tile.
15 . A decoder comprising:
receiver circuitry configured to receive a bitstream comprising a plurality of coded tiles and information associated with the plurality of coded tiles;
a memory comprising executable instructions; and
processing circuitry operatively connected to the receiver circuitry and the memory, and configured to execute the executable instructions to:
derive, from the information in the bitstream, a spatial partition structure that divides a picture into tiles:
wherein each tile corresponds to one of the plurality of coded tiles and comprises at least one unit;
wherein the information comprises both a height and a width of a first tile; and
wherein the information indicates that the decoder should copy a height, a width, or both the height and the width of a tile of the spatial partition structure from an adjacent tile of the spatial partition structure; and
decode the plurality of coded tiles received in the bitstream in accordance with the spatial partition structure, wherein derivation of any intra prediction mode for a first unit of the first tile depends on an intra prediction mode for a second unit of the first tile and is independent of intra prediction modes for units of any other tiles of the spatial partition structure.
16 . The decoder of claim 15 , wherein the decoder is comprised in a viewing device for displaying decoded tiles.
17 . A non-transitory computer-readable medium having executable instructions stored thereon that, when executed by a processing circuit in a decoder, causes the decoder to:
receive a bitstream comprising a plurality of coded tiles and information associated with the plurality of coded tiles;
derive, from the information in the bitstream, a spatial partition structure that divides a picture into tiles:
wherein each tile corresponds to one of the plurality of coded tiles and comprises at least one unit;
wherein the information comprises both a height and a width of a first tile; and
wherein the information indicates that the decoder should copy a height, a width, or both the height and the width of a tile of the spatial partition structure from an adjacent tile of the spatial partition structure; and
decode the plurality of coded tiles received in the bitstream in accordance with the spatial partition structure, wherein derivation of any intra prediction mode for a first unit of the first tile depends on an intra prediction mode for a second unit of the first tile and is independent of intra prediction modes for units of any other tile of the spatial partition structure.
18 . An encoder comprising:
a memory comprising executable instructions;
processing circuitry configured to execute the instructions to:
define a spatial partition structure that divides a picture into a plurality of tiles, wherein each tile comprises at least one unit;
encode the plurality of tiles in accordance with the spatial partition structure to generate a plurality of coded tiles, wherein each coded tile corresponds to one of the tiles of the spatial partition structure, and is independent such that derivation of an intra prediction mode for a first unit of a first tile depends on an intra prediction mode for a second unit of the first tile and is independent of intra prediction modes for units of any other tile of the spatial partition structure; and
generate a bitstream comprising the plurality of coded tiles and information indicating the spatial partition structure that divides the picture into the plurality of tiles, wherein:
the information comprises both a height and a width of the first tile; and
the information indicates that a decoder should copy a height, a width, or both the height and the width of a tile of the spatial partition structure from an adjacent tile of the spatial partition structure.
19 . A non-transitory computer-readable medium having executable instructions stored thereon that, when executed by a processing circuit in an encoder, causes the encoder to:
define a spatial partition structure that divides a picture into a plurality of tiles, wherein each tile comprises at least one unit;
encode the plurality of tiles in accordance with the spatial partition structure to generate a plurality of coded tiles, wherein each coded tile corresponds to one of the tiles of the spatial partition structure, and is independent such that derivation of an intra prediction mode for a first unit of a first tile depends on an intra prediction mode for a second unit of the first tile and is independent of an intra prediction modes for units of any other tiles of the spatial partition structure; and
generate a bitstream comprising the plurality of coded tiles and information indicating the spatial partition structure that divides the picture into the plurality of tiles, wherein:
the information comprises both a height and a width of the first tile; and
the information indicates that a decoder should copy a height, a width, or both the height and the width of a tile of the spatial partition structure from an adjacent tile of the spatial partition structure.