Video Coefficient Context Modeling for Adaptive Entropy Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding methods use fixed local patterns to calculate context models for entropy coding, which can lead to inefficient compression due to inaccurate probability estimation of transformed coefficients.

Innovation Solution

Adaptive context modeling that considers the distance from a current block, uses cross-bin dependency, and allows for flexible control of local template size and number of context models to improve entropy coding accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Device complexity

If a fixed local pattern is used to calculate the context model, then the calculation process is simple, but the probability estimation accuracy of transformed coefficients deteriorates

Engineering Contradiction:
Improvecalculation process complexityVSAvoidprobability estimation accuracy
Core Design Contradiction:
Device complexityVSMeasurement precision

Solution Approach 1:

The patent applies dynamics by transitioning from a fixed local pattern to a dynamic adaptive pattern selection mechanism. The context model calculation adapts to different block characteristics by selecting from multiple candidate patterns based on gradient calculations and variance comparisons, allowing the system to optimize probability estimation accuracy while maintaining manageable complexity through structured adaptation.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes parameters by introducing multiple candidate patterns with different spatial extents and orientations. Instead of using a single fixed pattern, the system evaluates multiple patterns with varying parameters (size, shape, orientation) and selects the most appropriate one based on local image characteristics, thereby improving probability estimation accuracy.

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If multiple candidate patterns with different spatial extents are used, then the probability estimation accuracy improves, but the calculation complexity increases

Engineering Contradiction:
Improveprobability estimation accuracyVSAvoidcalculation complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent segments the pattern selection process into distinct stages: gradient calculation for each candidate pattern, variance computation, and selective application based on threshold comparisons. This segmentation allows the system to evaluate multiple patterns efficiently by breaking down the complex selection process into manageable steps, improving accuracy without overwhelming computational burden.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent implements partial action by not fully evaluating all possible patterns in all cases. Instead, it uses gradient-based pre-screening to identify promising candidate patterns, then applies full evaluation only to those that meet certain criteria. This selective approach achieves high probability estimation accuracy while avoiding the full computational cost of evaluating every possible pattern.

Inventive Principle:
Principle #16Partial or excessive action

3Productivity

If adaptive pattern selection based on gradient calculation is used, then the compression efficiency improves, but the computational load increases

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcomputational load
Core Design Contradiction:
ProductivityVSPower

Solution Approach 1:

The patent applies preliminary action by calculating gradients and evaluating candidate patterns before the actual entropy coding process. By pre-determining the optimal pattern based on gradient analysis and variance comparisons, the system prepares the context model in advance, enabling more efficient compression during the actual encoding phase without repeating complex calculations.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The system performs self-service by automatically selecting the appropriate pattern based on local image characteristics without requiring external intervention or manual configuration. The gradient calculation and variance-based selection mechanism enables the context model to adapt autonomously to different block types, improving compression efficiency while the structured self-adaptation keeps computational load manageable.

Inventive Principle:
Principle #25Self-service

Data Source

PatentEP3844961B1Coefficient context modeling in video coding
Publication Date: 2026.02.18 HULU LLC
  • EP3844961B1 patent drawingFigure 1
  • EP3844961B1 patent drawingFigure 2A
  • EP3844961B1 patent drawingFigure 2B

AI summary

In some embodiments, a method determines a plurality of classes of bins that are used to determine a context model for entropy coding of a current block in a video. The method calculates a first value for a first class of bins in the plurality of classes of bins and calculates a second value for a second class of bins in the plurality of classes of bins. The first value for the first class of bins is weighted by a first weight to generate a weighted first value and the second value for the second class of bins is weighted by a second weight to generate a weighted second value. The method then selects a context model based on the first weighted value and the second weighted value.