Video Encoder Coding Unit Partitioning via Intra-Prediction Cost Storage
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face challenges in efficiently partitioning coding units in video encoders, particularly in maximizing coding efficiency and adapting to various content, applications, and devices.
Innovation Solution
The method involves performing intra-prediction on each permitted coding unit (CU) in a CU hierarchy of a largest coding unit (LCU) to determine intra-prediction coding costs, storing these costs, and then performing inter-prediction, prediction mode selection, and CU partition selection to determine CU partitioning for encoding, utilizing the stored intra-prediction costs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If traditional video coding standards (H.264/AVC) are used with fixed 16×16 macroblocks, then compatibility is maintained, but coding efficiency is limited and cannot adapt to larger resolutions and higher frame rates
Solution Approach 1:
The patent implements dynamic coding unit size selection where the encoder can adaptively choose from multiple CU sizes (8×8, 16×16, 32×32, 64×64) based on the actual video content characteristics. This dynamic adaptation allows the system to optimize coding efficiency for different resolutions and frame rates without being locked into a fixed block size like H.264/AVC
Solution Approach 2:
The patent changes the fundamental parameter of block size from fixed (16×16) to variable (8×8 to 64×64). By allowing the coding unit size to change based on content complexity and resolution requirements, the system achieves both improved coding efficiency and adaptability to various video parameters
2Productivity
If larger coding units (LCUs) are used in HEVC, then coding efficiency is improved, but the complexity of partitioning and mode selection increases
Solution Approach 1:
The patent segments the LCU into multiple smaller CUs of different sizes using quadtree partitioning. This segmentation allows the encoder to process larger LCUs while maintaining manageable complexity by dividing them into smaller units that can be processed independently, reducing the overall computational burden while preserving the benefits of large-block coding
Solution Approach 2:
The patent performs preliminary intra-prediction cost calculation for all permitted CUs before final mode selection. By pre-calculating and storing these costs, the encoder avoids redundant computations during the final decision-making process, reducing real-time complexity while maintaining optimal partitioning choices
3Adaptability or versatility
If recursive quadtree partitioning is applied to maximize coding efficiency, then adaptation to content is improved, but processing time and computational cost increase
Solution Approach 1:
The patent performs preliminary intra-prediction cost calculations for all possible CU partitions and stores these costs in memory. This pre-computation allows the final mode selection process to simply compare stored values rather than performing complex calculations in real-time, significantly reducing processing time while maintaining optimal content adaptation
Solution Approach 2:
The patent creates a hierarchy of cost information by copying and storing intra-prediction costs at different CU levels. This copied information is then used during mode selection without requiring re-computation, effectively decoupling the content adaptation process from the final decision-making process and reducing overall processing time
Data Source
AI summary
A method for coding unit partitioning in a video encoder is provided that includes performing intra-prediction on each permitted coding unit (CU) in a CU hierarchy of a largest coding unit (LCU) to determine an intra-prediction coding cost for each permitted CU, storing the intra-prediction coding cost for each intra-predicted CU in memory, and performing inter-prediction, prediction mode selection, and CU partition selection on each permitted CU in the CU hierarchy to determine a CU partitioning for encoding the LCU, wherein the stored intra-prediction coding costs for the CUs are used.