DMVR Inter Prediction Refinement for Lower-Complexity Video Decoding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The increasing demand for high-resolution and high-quality images and videos, along with immersive media formats, necessitates a high-efficiency image and video compression technology to reduce transmission and storage costs while maintaining quality.
Innovation Solution
Implementing decoder-side motion vector refinement (DMVR) and bi-directional optical flow (BDOF) with application conditions to enhance inter prediction efficiency, reducing calculation complexity and improving overall coding efficiency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If high-resolution and high-quality images and videos are transmitted, then image and video quality is improved, but transmission costs and storage costs are increased
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting motion vector refinement operations based on picture characteristics and coding conditions. The system changes parameters such as refinement precision and computational complexity based on the specific requirements of different video sequences, achieving high quality compression without excessive transmission and storage costs.
Solution Approach 2:
The patent implements partial action by selectively applying motion vector refinement only to specific blocks or regions where it is most beneficial, rather than uniformly processing the entire image. This partial refinement approach maintains image quality while significantly reducing the overall computational burden and associated costs.
2Measurement precision
If decoder-side motion vector refinement (DMVR) is applied to enhance prediction accuracy, then inter prediction efficiency is improved, but calculation complexity is increased
Solution Approach 1:
The patent applies dynamics by making the motion vector refinement process adaptive and configurable. The system dynamically adjusts the refinement strength and computational complexity based on picture characteristics, coding conditions, and available resources. This allows the system to optimize between prediction accuracy and calculation complexity in real-time.
Solution Approach 2:
The patent implements local quality by applying different levels of motion vector refinement to different regions of the image based on their specific requirements. Areas with complex motion patterns receive more refined prediction, while regions with simple motion receive less processing, thereby optimizing the balance between accuracy and complexity.
3Measurement precision
If bi-directional optical flow (BDOF) is applied to refine motion information, then prediction performance is enhanced, but processing time is increased
Solution Approach 1:
The patent applies periodic action by implementing multi-stage motion vector refinement processes that operate in sequential phases. The BDOF algorithm is applied in periodic intervals or at specific stages of the coding process rather than continuously, maintaining prediction performance while managing processing time through staged computation.
Solution Approach 2:
The patent implements preliminary action by performing motion information refinement in advance during encoding, so that the refined motion vectors are available for subsequent decoding operations. This preliminary processing of motion information can reduce the computational burden during actual prediction and rendering operations.
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An image decoding method according to the present document comprises the steps of: determining whether or not an application condition of decoder-side motion vector refinement (DMVR) for applying motion vector refinement for a current block is satisfied; deriving a minimum sum of absolute differences (SAD) on the basis of L0 and L1 motion vectors of the current block if the application condition of the DMVR is satisfied; deriving refined L0 and L1 motion vectors of the current block on the basis of the minimum SAD; deriving prediction samples of the current block on the basis of the refined L0 and L1 motion vectors; and generating reconstructed samples of the current block on the basis of the prediction samples. With respect to whether or not the application condition of the DMVR is satisfied, the application condition of the DMVR is determined to be satisfied if a prediction mode, in which inter-prediction and intra-prediction are combined, is not applied to the current block.