Affine Motion Model Offsets for Lower-Complexity Video Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face inefficiencies in predicting motion vectors, particularly in complex video scenarios, leading to suboptimal compression ratios and increased data requirements.

Innovation Solution

Implementing an affine motion model with prediction offsets, utilizing a plurality of offset indices to transform and reconstruct video blocks based on an affine model, allowing for more accurate motion vector derivation and improved compression efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If conventional affine motion model coding is used, then motion compensation accuracy is improved, but computational complexity and processing time increase

Engineering Contradiction:
Improvemotion compensation accuracyVSAvoidcomputational complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies partial action by using only a subset of the full affine motion model parameters. Instead of coding all six parameters (a11, a12, a21, a22, v1, v2), the system uses a reduced set of parameters (e.g., 4-parameter or 3-parameter models) that provide sufficient motion compensation accuracy for the given video content, thereby reducing computational complexity and processing requirements.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent changes the parameters of the affine motion model by introducing simplified parameter sets. The system selectively applies different parameter configurations based on block characteristics, using fewer parameters when full affine modeling is not necessary, thus achieving a balance between accuracy and complexity.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If conventional affine motion model coding is used, then motion compensation accuracy is improved, but encoding time increases

Engineering Contradiction:
Improvemotion compensation accuracyVSAvoidencoding time
Core Design Contradiction:
Measurement precisionVSLoss of time

Solution Approach 1:

The patent applies partial action by using only a subset of the full affine motion model parameters. Instead of coding all six parameters (a11, a12, a21, a22, v1, v2), the system uses a reduced set of parameters (e.g., 4-parameter or 3-parameter models) that provide sufficient motion compensation accuracy for the given video content, thereby reducing computational complexity and processing requirements.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent applies preliminary action by performing early decisions about which motion model complexity to use. The system evaluates block characteristics and motion patterns beforehand to determine the appropriate parameter set, avoiding unnecessary computations during the main encoding process and thus reducing overall encoding time.

Inventive Principle:
Principle #10Preliminary action

3Device complexity

If reduced complexity affine motion modeling is used, then computational complexity is reduced, but coding accuracy deteriorates

Engineering Contradiction:
Improvecomputational complexityVSAvoidcoding accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent changes the parameters of the affine motion model by introducing simplified parameter sets. The system selectively applies different parameter configurations based on block characteristics, using fewer parameters when full affine modeling is not necessary, thus achieving a balance between accuracy and complexity.

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent applies local quality by adapting the motion model complexity to local block characteristics. Different regions of the video are encoded with different levels of motion model sophistication based on their specific motion patterns, ensuring high accuracy where needed while reducing complexity where simpler models suffice.

Inventive Principle:
Principle #3Local quality

4Measurement precision

If standard affine motion model is applied to all blocks, then coding accuracy is maintained, but processing efficiency decreases

Engineering Contradiction:
Improvecoding accuracyVSAvoidprocessing efficiency
Core Design Contradiction:
Measurement precisionVSProductivity

Solution Approach 1:

The patent applies partial action by using only a subset of the full affine motion model parameters. Instead of coding all six parameters (a11, a12, a21, a22, v1, v2), the system uses a reduced set of parameters (e.g., 4-parameter or 3-parameter models) that provide sufficient motion compensation accuracy for the given video content, thereby reducing computational complexity and processing requirements.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The patent applies dynamics by making the motion model complexity adaptive rather than static. The system dynamically selects between different parameter configurations based on real-time analysis of block characteristics, motion patterns, and content complexity, thereby optimizing processing efficiency while maintaining coding accuracy where required.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP3788785B1Techniques for simplified affine motion model coding with prediction offsets
Publication Date: 2026.05.06 TENCENT AMERICA LLC
  • EP3788785B1 patent drawingFigure 1
  • EP3788785B1 patent drawingFigure 2
  • EP3788785B1 patent drawingFigure 3

AI summary

In a method for video decoding in a decoder, prediction information of a block is decoded in a current picture from a coded video bitstream. The prediction information includes a plurality of offset indices for prediction offsets associated with an affine model in an inter prediction mode. The plurality of offset indices includes at least one of a distance offset index, an offset direction index, a delta scaling index, and a delta rotation index. Further, parameters of the affine model are determined based on the plurality of offset indices. Each of the plurality of the offset indices including a respective pre-defined mapping table that includes indexes and corresponding offset values. The parameters of the affine model are used to transform between the block and a reference block in a reference picture that has been reconstructed. At least a sample of the block is reconstructed according to the affine model.