Optical Flow Prediction Refinement for Unified BDOF and PROF Coding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing video coding technologies, such as VVC, face inefficiencies in motion compensation due to non-unified bit-depth representations and workflows between bi-directional optical flow (BDOF) and prediction refinement with optical flow (PROF) for affine mode, leading to suboptimal hardware implementations and limited coding efficiency.
Innovation Solution
Harmonize the designs of BDOF and PROF by aligning bit-depth representations and gradient calculations, allowing shared pipeline designs and unified gradient derivation methods to enhance motion refinement precision and efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If separate bit-depth representations and workflows are used for BDOF and PROF, then each process can be independently optimized, but hardware implementation complexity increases and coding efficiency is limited
Solution Approach 1:
The patent merges the bit-depth representation and workflow of BDOF and PROF into a unified structure. Both processes now use the same bit-depth control mechanism and gradient calculation workflow, eliminating the need for separate independent optimization paths while reducing hardware complexity through shared components
Solution Approach 2:
The patent creates a universal bit-depth control mechanism that serves both BDOF and PROF processes. The same gradient calculation workflow and shift operation system are made multi-functional, allowing a single hardware implementation to handle both optical flow refinement processes efficiently
2Adaptability or versatility
If non-unified bit-depth representations are used for BDOF and PROF, then each process maintains its own workflow, but motion refinement precision is limited
Solution Approach 1:
The patent changes the bit-depth parameter representation to be unified across both BDOF and PROF processes. By standardizing the bit-depth control mechanism and using consistent shift operations, the precision of motion refinement is improved while maintaining adaptability through configurable parameters that can be adjusted based on specific coding conditions
3Adaptability or versatility
If separate gradient calculation methods are used for BDOF and PROF, then each process can be independently optimized, but computational efficiency decreases
Solution Approach 1:
The patent combines the gradient calculation methods of BDOF and PROF into a single unified approach. By deriving gradients from padded prediction samples in the same manner for both processes, computational efficiency is improved through shared calculation logic while maintaining the ability to independently optimize each process through parameter control
Data Source
AI summary
Methods, apparatuses, and non-transitory computer-readable storage mediums are provided for coding a video signal. A method includes obtaining a first reference picture and a second reference picture associated with a video block; obtaining first prediction samples of the video block from the first reference picture; obtaining second prediction samples of the video block from the second reference picture; obtaining padded prediction samples, and obtaining horizontal and vertical gradient values of the first prediction samples and the second prediction samples based on the padded prediction samples; obtaining motion refinements for samples in the video block based on the horizontal and vertical gradient values; and obtaining bi-prediction samples of the video block based on the motion refinements.