Sub-block Motion Prediction Refinement for Video Coding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video coding systems face challenges in efficiently compressing digital video signals, particularly in handling rapid changes in illuminance and achieving optimal prediction techniques for bi-directional motion compensated prediction.

Innovation Solution

The method involves obtaining a sub-block based motion prediction signal and refining it using spatial gradients or motion vector difference values to generate a refined motion prediction signal, which is then used for decoding the current block.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If traditional block-based motion prediction is used, then device complexity is reduced, but measurement precision of motion prediction deteriorates

Engineering Contradiction:

Improvemotion prediction accuracyVSAvoidprediction system complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The current block is divided into multiple sub-blocks, and motion prediction is performed separately for each sub-block using sub-block based motion vectors. This segmentation allows for more precise local motion representation while maintaining manageable computational complexity through standardized processing procedures for each sub-block.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different motion prediction parameters are applied to different sub-blocks within the current block based on their local motion characteristics. Each sub-block can have its own motion vector and prediction mode, allowing the system to adapt to local variations in motion patterns and improve overall prediction accuracy.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If motion prediction refinement is applied, then measurement precision improves, but loss of time increases

Engineering Contradiction:

Improvemotion prediction accuracyVSAvoiddecoding time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

Motion vector differences are calculated and stored in advance during the encoding phase. These pre-computed differences are then directly utilized during decoding to refine motion predictions without requiring complex real-time calculations, thereby reducing decoding time while maintaining high prediction accuracy.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The refinement process uses copied and reused data structures, such as copying motion vector difference values from reference blocks and reusing gradient calculation results across adjacent sub-blocks. This approach minimizes redundant computations and accelerates the refinement process during decoding.

Inventive Principle:

Principle #26Copying

3Manufacturing precision

If spatial gradient calculations are performed, then manufacturing precision of prediction signal improves, but use of energy increases

Engineering Contradiction:

Improveprediction signal accuracyVSAvoidcomputational energy consumption

Core Design Contradiction:

Manufacturing precisionVSUse of energy by moving object

Solution Approach 1:

Spatial gradient calculations are performed selectively only for sub-blocks that require refinement, rather than uniformly across the entire picture. The system applies gradient-based refinement only where motion complexity warrants it, reducing overall computational energy consumption while maintaining high prediction accuracy in critical regions.

Inventive Principle:

Principle #16Partial or excessive action

4Measurement precision

If sub-block based prediction is used, then measurement precision improves, but device complexity increases

Engineering Contradiction:

Improvemotion vector precisionVSAvoidprocessing complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

A unified set of processing procedures and data structures is designed to handle all sub-blocks consistently, regardless of their position or size. The same gradient calculation methods, motion vector difference computation approaches, and refinement algorithms are applied universally across different sub-block types, simplifying the overall system architecture despite the increased precision requirements.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS20250039435A1Systems, apparatus and methods for inter prediction refinement with optical flow

Publication Date: 2025.01.30 INTERDIGITAL VC HOLDINGS INC

AI summary

Method, apparatus and systems are disclosed. In one embodiment, a method of decoding includes obtaining a sub-block based motion prediction signal for a current block of the video; obtaining one or more spatial gradients of the sub-block based motion prediction signal or one or more motion vector difference values; obtaining a refinement signal for the current block based on the one or more obtained spatial gradients or the one or more obtained motion vector difference values; obtaining a refined motion prediction signal for the current block based on the sub-block based motion prediction signal and the refinement signal; and decoding the current block based on the refined motion prediction signal.