Adaptive Loop Filter Selection by Temporal Layer for Video Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding systems face challenges in efficiently managing the computational complexity of adaptive loop filters (ALFs) across different temporal layers, leading to suboptimal performance and resource utilization.
Innovation Solution
An adaptive loop filter (ALF) procedure is selectively applied based on the temporal layer of a frame, using reduced computational complexity techniques such as subset pixel selection and gradient calculation, to optimize ALF operations in video coding systems.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If a full ALF procedure is applied to all frames regardless of temporal layer, then video quality is maintained, but computational complexity increases
Solution Approach 1:
The patent applies different ALF processing qualities to different temporal layers. Frames in lower temporal layers (higher quality) receive full ALF processing, while frames in higher temporal layers (lower quality) receive simplified or skipped ALF processing. This local differentiation of processing quality matches the varying importance and quality requirements of different temporal layers, resolving the contradiction between maintaining video quality and reducing computational complexity.
Solution Approach 2:
The patent dynamically selects the ALF processing procedure based on the temporal layer of each frame. The decoder determines the temporal layer level and adaptively chooses between full ALF, simplified ALF, or skipping ALF entirely. This dynamic adaptation allows the system to optimize computational complexity while maintaining appropriate video quality for each temporal layer, rather than applying a static processing approach to all frames.
2Reliability
If ALF is applied to higher temporal layers, then video quality is maintained, but decoding speed decreases
Solution Approach 1:
The patent recognizes that higher temporal layers contain smoother, less complex content that requires less filtering. By applying simplified or skipped ALF processing specifically to these higher temporal layers, the system maintains adequate video quality for these regions while significantly improving decoding speed. This local quality adjustment resolves the contradiction between maintaining quality and improving productivity.
Solution Approach 2:
The patent applies partial ALF processing to higher temporal layers rather than full processing. Since these layers contain smoother content, the full ALF processing would be excessive. By applying only necessary filtering (or skipping it entirely for the smoothest layers), the system achieves adequate quality with improved decoding speed, resolving the contradiction between quality maintenance and productivity enhancement.
3Device complexity
If subset pixel selection is used for gradient calculation, then computational complexity is reduced, but filtering precision may decrease
Solution Approach 1:
The patent uses subset pixel selection for gradient calculation in higher temporal layers, applying partial processing rather than full processing. Since these layers contain smoother content, calculating gradients for all pixels would be excessive. By selecting a subset of pixels (e.g., every other pixel or only boundary pixels), the system reduces computational complexity while maintaining sufficient filtering precision for the smoother content characteristics of higher temporal layers.
Solution Approach 2:
The patent applies different gradient calculation strategies to different temporal layers. Lower temporal layers use full pixel gradient calculation for high precision, while higher temporal layers use subset pixel selection. This local differentiation of calculation precision matches the varying content complexity across temporal layers, resolving the contradiction between computational complexity and filtering precision.
Data Source
AI summary
Systems, methods and instrumentalities are disclosed for adaptively selecting an adaptive loop filter (ALF) procedure for a frame based on which temporal layer the frame is in. ALF procedures may vary in computational complexity. One or more frames including the current frame may be in a temporal layer of a coding scheme. The decoder may determine the current frame's temporal layer level within the coding scheme. The decoder may select an ALF procedure based on the current frame's temporal layer level. If the current frame's temporal layer level is higher within the coding scheme than some other temporal layer levels, an ALF procedure that is less computationally complex may be selected for the current frame. Then the decoder may perform the selected ALF procedure on the current frame.