Inter-layer Reference Picture Set Decoding for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding systems face challenges in efficiently representing and decoding high-quality digital media, particularly in managing inter-layer reference picture sets and reference picture lists, which affects storage, transmission, and playback efficiency.
Innovation Solution
The method involves signaling the maximum number of sub-layers for inter-layer prediction and using a temporal identifier to determine whether an inter-layer picture should be added to the inter-layer reference picture set, thereby optimizing the reference picture list construction and reducing the decoded picture buffer size.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If inter-layer reference picture sets are expanded to include more pictures for better prediction quality, then video decoding quality is improved, but memory buffer size and system complexity increase
Solution Approach 1:
The patent introduces a new parameter max_sublayer_for_ilp_plus1 that dynamically controls the temporal identifier threshold for inter-layer prediction. By adjusting this parameter, the system can adaptively limit the number of reference pictures in the inter-layer reference picture set, thereby controlling buffer size while maintaining prediction quality where beneficial
Solution Approach 2:
The patent implements dynamic control of reference picture set composition based on temporal identifier values. The decoding process dynamically evaluates each inter-layer picture's temporal identifier against the max_sublayer_for_ilp_plus1 parameter to determine inclusion in the reference picture set, allowing flexible adaptation to different coding scenarios rather than using a fixed reference picture set size
2Measurement precision
If more inter-layer pictures are added to the reference picture set for better prediction, then prediction accuracy is improved, but device complexity increases
Solution Approach 1:
The patent simplifies reference picture list construction by introducing a clear threshold parameter max_sublayer_for_ilp_plus1. This parameter provides a straightforward criterion for including or excluding inter-layer pictures, replacing complex decision logic with a simple temporal identifier comparison that reduces computational complexity while maintaining prediction accuracy for suitable pictures
3Measurement precision
If temporal identifier checking is performed for all inter-layer pictures, then reference picture set accuracy is improved, but processing time increases
Solution Approach 1:
The patent performs preliminary filtering of inter-layer pictures by comparing their temporal identifiers against the max_sublayer_for_ilp_plus1 parameter before full decoding processing. This preliminary check eliminates unsuitable reference pictures early in the process, preventing wasted processing time on pictures that would not be included in the reference picture set anyway, thus reducing overall decoding time while maintaining accuracy
Data Source
AI summary
A method for video coding is described. Signaling of a maximum number of sub-layers for inter-layer prediction is obtained. A sub-layer non-reference picture is also obtained. It is determined whether a value of a temporal identifier of the sub-layer non-reference picture is greater than the maximum number of sub-layers for inter-layer prediction minus 1. The sub-layer non-reference picture is marked as “unused for reference” if the value of the temporal identifier of the sub-layer non-reference picture is greater than the maximum number of sub-layers for inter-layer prediction minus 1. In some cases a sub-layer non-reference picture is also obtained. It is determined whether a value of a temporal identifier of the sub-layer non-reference picture is greater than the maximum number of sub-layers for inter-layer prediction. The sub-layer non-reference picture is marked as “unused for reference” if the value of the temporal identifier of the sub-layer non-reference picture is greater than the maximum number of sub-layers for inter-layer prediction.