Reference Picture Resampling Rules for P-Slice Motion Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding standards face challenges in efficiently managing the conversion between coded representations of video and pixel values, particularly in handling slice types, temporal motion vector prediction, and subpicture structures, which affect bandwidth utilization and decoding efficiency.
Innovation Solution
Implementing a series of format rules and syntax elements that govern the conversion process between video slices and bitstreams, including rules for slice types, temporal motion vector prediction, and subpicture handling, to enhance compatibility and decoding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If reference picture resampling (RPR) is enabled for temporal motion vector prediction, then prediction accuracy is improved, but processing complexity and bandwidth requirements increase
Solution Approach 1:
The patent applies local quality by selectively enabling RPR only for specific slice types (P slices) and specific reference pictures (collocated pictures), rather than uniformly applying it to all cases. This allows the system to achieve improved prediction accuracy where needed while avoiding unnecessary processing complexity in other areas.
Solution Approach 2:
The patent introduces dynamic control through syntax elements that allow the encoder to adaptively enable or disable RPR based on picture characteristics, slice types, and reference picture lists. This dynamic approach allows the system to optimize between prediction accuracy and processing complexity on a case-by-case basis.
2Area of stationary object
If multiple subpictures are included in a video picture, then spatial utilization is improved, but syntax element signaling complexity increases
Solution Approach 1:
The patent implements segmentation by dividing a video picture into multiple subpictures, each with its own slice structure. This allows independent processing and optimization of different spatial regions, improving overall spatial utilization while maintaining manageable complexity through modular organization.
Solution Approach 2:
The patent creates a universal syntax structure that can handle both single-subpicture and multi-subpicture cases through conditional syntax elements. The same framework adapts to different numbers of subpictures, reducing the need for separate signaling mechanisms and thereby controlling complexity.
3Productivity
If slice type determines inheritance of picture header information, then decoding efficiency is improved, but format rule complexity increases
Solution Approach 1:
The patent applies preliminary action by pre-defining inheritance rules for different slice types (P slices, B slices, etc.) in the format specifications. During decoding, the decoder can efficiently determine which picture header information to inherit based on the slice type without complex runtime analysis, as the rules are established in advance.
Solution Approach 2:
The patent uses parameter changes by varying the inheritance behavior based on slice type parameters. Different slice types have different inheritance characteristics, allowing the system to optimize decoding efficiency for each slice type while maintaining a unified format rule structure.
Data Source
AI summary
Methods, apparatus, systems for video processing, including video encoding or video decoding are described. One example method includes performing a conversion between a video comprising a video picture that includes a video slice and a bitstream of the video. The bitstream conforms to a format rule specifying that, in case a slice type of the video slice is type P and temporal motion vector prediction is enabled, usage of reference picture resampling (RPR) for a reference picture in a collocated reference picture list is disabled. The reference picture is indicated by a reference index of a collocated picture of the video slice used for the temporal motion vector prediction.