Partition-Based Video Filters for Buffer and Quality Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional video compression techniques require significant buffer capacity and can introduce visual quality issues due to the application of adaptive loop filters across block boundaries, leading to inefficiencies in processing and potential artifacts like blockiness.
Innovation Solution
The implementation of partition-based filters, such as partial filters and mirror padding, which reduce buffer requirements and minimize visual artifacts by selectively applying filter coefficients and mirroring pixel values to eliminate the need for extensive line buffers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If conventional adaptive loop filters are applied across block boundaries, then video quality is improved, but buffer capacity requirements increase significantly
Solution Approach 1:
The patent divides the filtering operation into two distinct stages: an odd-line filter applied first, followed by an even-line filter. This segmentation allows each filter to operate with reduced buffer requirements compared to a single full adaptive loop filter, while collectively achieving the desired video quality improvement. The odd-line filter processes every other line, and the even-line filter processes the remaining lines, enabling progressive filtering with minimal buffering.
Solution Approach 2:
The patent applies the odd-line filter before the even-line filter in a sequential manner. This preliminary action on odd lines provides intermediate filtering results that the even-line filter can build upon, allowing the system to achieve complete filtering效果 with reduced buffer capacity compared to applying a single filter all at once.
2Manufacturing precision
If conventional adaptive loop filters are applied across block boundaries, then video quality is improved, but visual artifacts like blockiness increase
Solution Approach 1:
By segmenting the filtering into odd-line and even-line filters applied sequentially, the patent reduces the strength and scope of individual filtering operations across block boundaries. This prevents any single filter from creating strong blockiness artifacts, while the combined effect of both filters still achieves smooth video quality improvement.
Solution Approach 2:
Each line filter (odd and even) applies partial filtering action rather than complete filtering across all lines. This partial action reduces the risk of creating blockiness artifacts at block boundaries, while the sequential application of both filters collectively achieves the desired overall filtering效果 without excessive local processing that would cause artifacts.
3Quantity of substance
If partition-based filters with mirror padding are used, then buffer requirements are reduced, but processing complexity increases
Solution Approach 1:
The patent segments the complex adaptive loop filtering task into two simpler, identical odd-line and even-line filter operations. This segmentation reduces the buffer requirements for each individual filter while distributing the processing complexity across two sequential stages, making the overall system more manageable and efficient.
4Quantity of substance
If partition-based filters with mirror padding are used, then buffer requirements are reduced, but processing time may increase
Solution Approach 1:
The odd-line filter performs preliminary filtering action on alternate lines first, producing intermediate results that the even-line filter then processes. This preliminary action enables progressive filtering without requiring large buffers to hold all data, and the sequential processing can be efficiently implemented to minimize overall processing time while reducing buffer requirements.
Data Source
AI summary
In general, techniques are described for applying partition-based filters when coding video data. A device comprising at least one processor may be configured to implement the techniques. The processor selects a filter to apply near a boundary of a first portion of the video data and determines at least one of the plurality of filter coefficients of the selected filter for which the video data will not be available to be filtered. Based on the determination, the processor determines a partial filter that does not include the at least one of the plurality of filter coefficients for which the video data will not be available to be filtered. The processor renormalizes the plurality of filter coefficients included within the partial filter and applies the renormalized partial filter near the boundary of the first portion of the video data to generate a filtered first portion of the video data.


