Video Tile Group Signaling for Raster and Rectangular Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding systems struggle to efficiently signal and utilize both raster scan and rectangular tile groups, leading to inefficient resource usage and complex syntax structures, particularly in applications like virtual reality and region of interest coding.
Innovation Solution
Implementing a flag in the bitstream to indicate whether the tile group is raster scan or rectangular, along with signaling the first and last tile of the group, allows the decoder to determine tile inclusion, reducing the need for a full list of tiles and simplifying the signaling process.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a flag is added to indicate tile group type, then decoder functionality and adaptability are improved, but bitstream complexity increases
Solution Approach 1:
The tile group configuration is segmented into two distinct types (raster scan tile group and rectangular tile group), each with specific syntax elements. The flag `rectangular_tile_group_flag` segments the configuration process, allowing the decoder to selectively parse different syntax elements based on the flag value, thereby improving adaptability without uniformly increasing complexity for all cases.
Solution Approach 2:
The tile group configuration is made dynamic through the flag mechanism, allowing the bitstream to adaptively indicate whether rectangular or raster scan tile groups are used. This dynamic signaling enables the decoder to adjust its parsing behavior based on the actual tile group type, enhancing versatility while managing complexity through conditional processing.
2Adaptability or versatility
If rectangular tile groups are used to support sub-picture extraction, then functionality for VR and region of interest coding is improved, but syntax structure complexity increases
Solution Approach 1:
The syntax structure is segmented into two distinct configurations: one for raster scan tile groups and another for rectangular tile groups. The `rectangular_tile_group_flag` acts as a delimiter that segments the syntax elements into two sets, allowing the decoder to process only the relevant syntax for the current tile group type, thereby supporting sub-picture extraction without uniformly increasing syntax complexity.
Solution Approach 2:
Different syntax elements are applied locally to different tile group types. The rectangular tile group configuration includes additional syntax elements (`rectangular_tile_group_flag`, `tile_group_rectangular_flag`) that are locally relevant only when rectangular tile groups are used, while raster scan tile groups use different syntax. This local quality approach enables specialized functionality for VR and region of interest coding without complicating the entire syntax structure.
3Productivity
If multiple tile group schemes are supported, then coding efficiency and resource usage are improved, but encoder and decoder complexity increases
Solution Approach 1:
The encoder and decoder implement dynamic tile group configuration through the `rectangular_tileGroup_flag` mechanism. The flag dynamically indicates whether rectangular or raster scan tile groups are used, allowing the encoder to select the appropriate coding scheme based on the video content and application requirements. This dynamic approach enables multiple tile group schemes to be supported with a single unified syntax structure, improving coding efficiency while managing complexity through conditional processing rather than requiring separate handling for each scheme.
Data Source
AI summary
A video coding mechanism is disclosed. The mechanism includes partitioning a picture into a plurality of tiles. A number of the tiles are included in a tile group. A flag is also encoded into a parameter set of a bitstream. The flag is set to a first value when the tile group is a raster scan tile group and a second value when the tile group is a rectangular tile group. The tiles are encoded into the bitstream based on the tile group. The bitstream is stored for communication toward a decoder.