Base Anchored Displacement Model for Video Compression
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing video coding technologies face inefficiencies in describing motion between frames, leading to redundant and opportunistic motion descriptions that are not physically or temporally consistent, which hampers accurate inversion and composition of motion fields across time.
Innovation Solution
The method involves assigning video frames to Groups of Pictures (GOPs) with a base displacement model that describes displacement fields anchored at a base frame, allowing all displacement information within a GOP to be derived from this base model, reducing the need for auxiliary prediction modes and enabling more geometrically consistent displacement information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional target-frame centric motion modeling is used, then motion prediction for target frames is achieved, but temporal consistency and geometric accuracy of motion fields deteriorate
Solution Approach 1:
The patent inverts the conventional motion modeling approach by anchoring motion at reference frames instead of target frames. Motion vectors are attached to reference frames pointing to prediction target locations, enabling temporal reasoning and composition of motion fields across time while maintaining geometric accuracy.
Solution Approach 2:
The patent segments motion fields into piecewise smooth components anchored at reference frames, allowing independent estimation and composition of motion relationships. This segmentation enables sparse motion fields that are temporally consistent and geometrically accurate without requiring exhaustive block-based modes.
2Productivity
If block-based motion compensation is used, then compression is achieved, but temporal reasoning and geometric consistency of motion fields are lost
Solution Approach 1:
The patent transitions from static block-based motion compensation to dynamic piecewise smooth motion fields that can be composed and inverted across time. This dynamic approach maintains compression efficiency while enabling temporal reasoning and geometric consistency through the anchored reference frame structure.
3Measurement precision
If multiple auxiliary prediction modes are used, then prediction accuracy is improved, but computational complexity and device requirements increase
Solution Approach 1:
The patent creates a universal motion modeling framework where reference-frame anchored motion vectors serve multiple functions: prediction, temporal reasoning, and geometric transformation. This multi-functionality eliminates the need for multiple auxiliary prediction modes while maintaining or improving prediction accuracy.
Data Source
AI summary
A method of representing displacement information between the frames of a video and/or multiview sequence, comprising the steps of assigning a plurality of the frames to a Group of Pictures (GOPs), providing a base displacement model for each GOP, the base displacement model describing a displacement field that carries each location in a designated base frame of the GOP to a corresponding location in each other the frame of the GOP, and inferring other displacement relationships between the frames of the GOP from the base displacement model. Embodiments include a piecewise smooth displacement field, deformable mesh, reverse displacement field, double mapping, synthesize optical blur and foreground-background discrimination process.