Video Encoding Subsampling Reference Pixels
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing video encoding and decoding methods, such as HEVC, face challenges in reducing memory access granularity and bandwidth, leading to bottlenecks in encoding/decoding operations due to the large number of reference pixels required for parameter derivation, especially in worst-case conditions for images with different formats like YUV444.
Innovation Solution
The proposed solution involves a video encoding and decoding apparatus that subsamples reference pixels for intra-frame prediction, reducing the number of reference pixels by half, specifically for luminance and chrominance blocks, and performing subsampling only when the smallest coding unit is selected or at all times regardless of the coding unit size, to enhance prediction efficiency and reduce redundancy between color components.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of reference pixels is increased to improve prediction accuracy, then prediction accuracy is improved, but memory access bandwidth increases and becomes a bottleneck
Solution Approach 1:
The patent extracts only the necessary reference pixels from the complete set of reference pixels. By identifying and removing redundant reference pixels through subsampling, the system maintains prediction accuracy using only the essential pixels, thereby reducing memory access bandwidth requirements.
Solution Approach 2:
The patent applies partial action by using a subset of reference pixels rather than all available reference pixels. The subsampling process selects only the necessary portion of reference pixels needed for accurate prediction, avoiding the excessive use of all reference pixels and thus reducing memory access overhead.
2Productivity
If the number of reference pixels is reduced to decrease memory access bandwidth, then memory access efficiency is improved, but prediction accuracy deteriorates
Solution Approach 1:
The patent changes the parameter of reference pixel selection by introducing subsampling with specific sampling patterns and ratios. This parameter change allows the system to use fewer reference pixels while maintaining prediction accuracy by optimizing which pixels are selected through structured subsampling approaches.
3Quantity of substance
If subsampling is applied to reduce reference pixels, then the number of reference pixels is reduced, but processing complexity increases
Solution Approach 1:
The patent segments the reference pixel selection process into distinct stages: full-set usage for larger coding units and subsampled usage for smaller coding units. This segmentation allows the system to apply different processing strategies based on coding unit size, reducing overall complexity by avoiding unnecessary subsampling in cases where it is not beneficial.
Solution Approach 2:
The patent implements dynamic reference pixel processing by adapting the subsampling approach based on coding unit size. The system dynamically switches between using all reference pixels and using subsampled reference pixels, optimizing processing complexity for different scenarios while maintaining the benefit of reduced reference pixel quantity where applicable.
Data Source
AI summary
An intra prediction unit that performs intra-frame prediction includes a luminance reference pixel acquisition unit and a chrominance reference pixel acquisition unit. When the coding unit is set to a smallest CU block, the luminance reference pixel acquisition unit subsamples reference pixels arranged at integer pixel locates neighboring a luminance block that corresponds to a chrominance prediction target block, and acquires the pixel values of the reference pixels after the subsampling. Furthermore, when the coding unit is set to a smallest CU block, the chrominance reference pixel acquisition unit subsamples reference pixels arranged at integer pixel locates neighboring the chrominance prediction target block, and acquires the pixel values of the reference pixels after the subsampling. Such an arrangement is capable of reducing the number of reference pixels used to reduce the redundancy between the color components.