Back to results
US 12,120,360 B2Granted
Tile signaling in video coding
AssigneeHuawei Technologies Co., Ltd.
InventorsYe-Kui Wang, Fnu Hendry, Jianle Chen
Life of this patent
Estimating the term…
Abstract
A video coding mechanism is disclosed. The mechanism includes receiving a bitstream including image data divided into one or more slices. The one or more slices are further divided into one or more tiles. The mechanism further includes determining that a first slice contains a single tile based on a first flag. The mechanism further includes decoding the first slice to generate a reconstructed image.
Claims20
A method implemented by a decoder, the method comprising:
- receiving a bitstream including a picture divided into one or more slices, wherein the one or more slices are further divided into one or more tiles;
- determining a number of the tiles in a first slice based on a first parameter;
- determining whether the picture comprises only one tile based on a first flag;
- determining a second parameter that specifies a tile index of a tile in the first slice that contains a first coding tree unit (CTU), wherein coding tree block (CTB) addresses for the one or more slices are included in a range from zero to picture size in luma coding tree blocks (PicSizeInCtbsY)−1;
- determining a number of entry points in the first slice based on the number of the tiles, wherein an entry point offset minus one (entry_point_offset_minus1[i]) plus one syntax element specifies each ith entry point offset in bytes based on the number of entry points, wherein the bitstream is a sub-bitstream of a coded bitstream, and wherein a header of the first slice has not been rewritten due to the bitstream lacking at least one tile of the picture from the coded bitstream;
- and
- decoding the first slice to decode the picture based on the first parameter and the second parameter.
4 more elements
A method implemented in an encoder, the method comprising:
- partitioning a picture into one or more slices, wherein the one or more slices are further partitioned into one or more tiles;
- determining a number of tiles in a first slice of the picture after partitioning;
- encoding a first parameter to indicate the number of the tiles in the first slice;
- encoding a first flag indicating whether the picture comprises only one tile;
- determining a second parameter that specifies a tile index of a tile in the first slice that contains a first coding tree unit (CTU), wherein coding tree block (CTB) addresses for the one or more slices are included in a range from zero to picture size in luma coding tree blocks (PicSizeInCtbsY)−1;
- encoding the first slice into a bitstream, wherein a number of entry points in the first slice is based on the number of the tiles, and wherein an entry point offset minus one (entry_point_offset_minus1[i]) plus one syntax element specifies each ith entry point offset in bytes based on the number of entry points, wherein the bitstream is a sub-bitstream of a coded bitstream, and wherein a header of the first slice has not been rewritten due to the bitstream lacking at least one tile of the picture from the coded bitstream;
- and
- transmitting the bitstream containing the first parameter to support decoding the first slice.
5 more elements
A video coding device comprising:
- a receiver configured to receive a bitstream including a picture divided into one or more slices, wherein the one or more slices are further divided into one or more tiles;
- and
- a processor configured to:
- determine a number of the tiles in a first slice based on a first parameter;
- determine whether the picture comprises only one tile based on a first flag;
- determine a second parameter that specifies a tile index of a tile in the first slice that contains a first coding tree unit (CTU), wherein coding tree block (CTB) addresses for the one or more slices are included in a range from zero to picture size in luma coding tree blocks (PicSizeInCtbsY)−1;
- determine a number of entry points in the first slice based on the number of the tiles, wherein an entry point offset minus one (entry_point_offset_minus1[i]) plus one syntax element specifies each ith entry point offset in bytes based on the number of entry points, wherein the bitstream is a sub-bitstream of a coded bitstream, and wherein a header of the first slice has not been rewritten due to the bitstream lacking at least one tile of the picture from the coded bitstream;
- and
- decode the first slice to decode the picture based on the first parameter and the second parameter.
6 more elements
Dependencies are inferred from the claim text.