POC-Conditioned DMVR and BDOF for Image Inter-Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images and videos, particularly in immersive media formats, necessitates a high-efficiency image and video compression technology to reduce transmission and storage costs while maintaining performance.
Innovation Solution
Implementing decoder-side motion vector refinement (DMVR) and bi-directional optical flow (BDOF) to enhance inter prediction, with conditions for applying these methods based on Picture Order Count (POC) differences, thereby optimizing motion information refinement during motion compensation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If decoder-side motion vector refinement (DMVR) and bi-directional optical flow (BDOF) are applied to enhance inter prediction, then image/video compression efficiency is improved, but calculation complexity increases
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the application of DMVR and BDOF based on POC difference conditions. When the first POC difference equals the second POC difference, BDOF is applied to refine motion vectors; when they differ, only DMVR is used. This conditional parameter adjustment optimizes compression efficiency while controlling calculation complexity by avoiding unnecessary computations.
2Measurement precision
If motion vector refinement is applied to all blocks, then prediction accuracy is improved, but processing time increases
Solution Approach 1:
The patent implements partial action by selectively applying BDOF only to blocks where the first POC difference equals the second POC difference, rather than applying motion vector refinement to all blocks. This conditional application ensures prediction accuracy is improved where needed while avoiding unnecessary processing time consumption in blocks where such refinement is not beneficial.
3Manufacturing precision
If high-resolution and high-quality image/video compression is implemented, then image quality is improved, but transmission and storage costs increase
Solution Approach 1:
The patent addresses this contradiction by dynamically changing compression parameters based on POC differences. By conditionally applying BDOF refinement only when first POC difference equals second POC difference, the method achieves high image quality where necessary while reducing overall computational burden, thereby lowering transmission and storage costs for high-resolution content.
Data Source
AI summary
An image decoding method according to the present document is characterized in that: decoder-side motion vector refinement (DMVR) is applied to derive refined L0 and L1 motion vectors on the basis of L0 and L1 motion vectors of the current block; prediction samples of the current block are derived on the basis of L0 prediction samples derived on the basis of the refined L0 motion vector and L1 prediction samples derived on the basis of the refined L1 motion vector; and recovery samples of the current block are generated on the basis of the prediction samples, wherein the L0 prediction samples are derived on the basis of an L0 reference picture and the refined L0 motion vector, the L1 prediction samples are derived on the basis of an L1 reference picture and the refined L1 motion vector, and in the step for deriving the prediction samples, whether to apply a bi-directional optical flow (BDOF) process is determined on the basis of whether a first POC difference between the current picture and the L0 reference picture and a second POC difference between the current picture and the L1 reference picture are the same.