Neighbor Sample Filtering for Flexible Video Block Partitioning

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video coding technologies face inefficiencies due to misalignment between reconstructed samples in the current block and reference samples caused by in-loop filtering, leading to suboptimal coding accuracy and efficiency.

Innovation Solution

Implement methods for video encoding and decoding that determine whether to apply a filter to neighboring reconstructed areas of the current block, predict the current block based on filtered neighboring samples, and signal control flags and prediction modes in the video bitstream, while also allowing flexible partitioning of video areas for improved coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If in-loop filtering is applied to reconstructed samples, then coding efficiency is improved, but misalignment between reconstructed samples and reference samples occurs reducing coding accuracy

Engineering Contradiction:

Improvecoding efficiencyVSAvoidcoding accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent applies different filtering strategies to different regions: template regions use filtering to improve efficiency, while reference sample regions use selective filtering (only when boundary strength indicates no misalignment) to maintain accuracy. This local differentiation resolves the contradiction between efficiency improvement and accuracy preservation.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The patent performs preliminary assessment of boundary strength before applying filtering to reference samples. By evaluating alignment conditions in advance, the system avoids unnecessary filtering that would cause misalignment, thus maintaining coding accuracy while still achieving efficiency gains where appropriate.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If filtering is applied to neighboring reconstructed areas, then prediction accuracy is improved, but additional processing complexity is introduced

Engineering Contradiction:

Improveprediction accuracyVSAvoidprocessing complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent changes the boundary strength parameter to control filtering application. By using boundary strength as a decision criterion, the system simplifies the complexity-management approach while maintaining prediction accuracy - filtering is applied only when boundary strength indicates favorable conditions, avoiding unnecessary processing.

Inventive Principle:

Principle #35Parameter changes

3Adaptability or versatility

If control flags and prediction modes are signaled in the bitstream, then coding flexibility is improved, but bitstream overhead increases

Engineering Contradiction:

Improvecoding flexibilityVSAvoidbitstream overhead

Core Design Contradiction:

Adaptability or versatilityVSLoss of information

Solution Approach 1:

The patent signals control flags and prediction modes partially - only when boundary strength conditions are met or when specific scenarios require enhanced flexibility. This partial signaling approach maintains necessary coding flexibility while avoiding unnecessary bitstream overhead from always signaling all parameters.

Inventive Principle:

Principle #16Partial or excessive action

Data Source

PatentUS20250365408A1Signaling and filtering of template and flexible partition split

Publication Date: 2025.11.27 TENCENT AMERICA LLC

AI summary

Methods and apparatuses for video decoding and video encoding and methods of processing visual media data are provided. A method for video decoding includes receiving coded information of a current block in a current picture. Whether to apply a filter to a neighboring reconstructed area of the current block is determined. The neighboring reconstructed area is adjacent to the current block and includes reconstructed samples in the current picture. The filter is applied to the neighboring reconstructed area of the current block. The current block is reconstructed according to the filtered neighboring reconstructed area.