JMVD Scaling Factor Derivation for Non-Linear Motion Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies, such as AV1 and HEVC, assume linear motion between reference frames, which may not accurately represent non-linear motion, leading to inefficiencies in motion vector difference (JMVD) coding.
Innovation Solution
A method for video coding that involves obtaining a coding block, determining if JMVD is used, deriving a list of scaling factors, generating prediction blocks, and signaling a selection of scaling factors based on a cost criterion to improve JMVD prediction accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If linear motion assumption is used in JMVD coding, then device complexity is reduced, but manufacturing precision (prediction accuracy) deteriorates
Solution Approach 1:
The patent applies dynamics by making the motion model adaptive rather than static. It introduces multiple motion modes (linear motion, non-linear motion with control points, and optical flow) that can be dynamically selected based on the characteristics of the video content. The system dynamically adjusts the complexity of the motion compensation algorithm to match the actual motion patterns in different regions, thereby resolving the contradiction between computational simplicity and prediction accuracy.
Solution Approach 2:
The patent changes parameters by introducing control points and adjusting motion vector precision dynamically. Instead of using a fixed linear motion model, the system modifies motion parameters (control point positions, motion vector resolutions) based on content characteristics. This allows the coding system to adapt its complexity level and achieve higher prediction accuracy for non-linear motions while maintaining efficiency for simpler scenes.
2Manufacturing precision
If non-linear motion modeling is implemented, then manufacturing precision (prediction accuracy) is improved, but device complexity increases
Solution Approach 1:
The patent applies segmentation by dividing the coding block into multiple regions with different motion characteristics. Control points are strategically placed to define sub-regions, each potentially using different motion models. This segmentation allows complex non-linear motion to be modeled accurately in regions where it occurs, while simpler regions continue to use efficient linear models, thus managing overall complexity while improving precision where needed.
Solution Approach 2:
The patent implements partial action by applying complex non-linear motion modeling only to specific regions or blocks where it is actually needed, rather than uniformly across the entire video stream. The system selectively enables control points and advanced motion compensation for blocks exhibiting non-linear motion patterns, leaving other blocks to use simpler models. This partial application reduces overall computational complexity while maintaining prediction accuracy in critical areas.
Data Source
AI summary
A method for video coding includes obtaining a coding block of video data, determining whether a joint coding of motion vector difference (JMVD) is used for predicting the coding block, obtaining, based on determining that the JMVD is used for predicting the coding block, a list of scaling factors and the JMVD from the video bitstream, deriving a predicted scaling factor based on the list of scaling factors, deriving a motion vector difference (MVD) for a reference list based on at least the joint motion vector difference and the predicted scaling factor, and reconstructing the coding block based on the derived MVD.