Intra Prediction Boundary Filtering in QTBT Block Structures

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video compression techniques face challenges in efficiently processing next-generation video contents with high spatial resolution, high frame rate, and high dimensionality, requiring enhanced coding tools to manage increased memory storage, memory access rate, and processing power.

Innovation Solution

A method for adaptively filtering boundary pixels of prediction blocks in a quad-tree binary-tree (QTBT) block partitioning structure during intra-picture prediction, where filtering decisions are based on the width and height of non-square blocks and specific intra prediction modes, optimizing the filtering process for improved compression performance.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If block boundary filtering is applied to all blocks regardless of shape, then filtering consistency is maintained, but processing complexity increases for non-square blocks in QTBT structure

Engineering Contradiction:
Improvefiltering consistencyVSAvoidprocessing complexity
Core Design Contradiction:
ReliabilityVSDevice complexity

Solution Approach 1:

The patent applies different filtering strategies to different parts of the block structure. For non-square blocks in QTBT, the filtering is adapted based on the specific dimensions and orientation of each block, rather than applying a uniform filtering approach to all blocks. This local adaptation reduces processing complexity while maintaining necessary filtering quality.

Inventive Principle:
Principle #3Local quality

Solution Approach 2:

The filtering process is made dynamic by adapting to the block characteristics. The patent determines filtering parameters based on the actual block width, height, and orientation in the QTBT structure, allowing the filtering behavior to change dynamically according to the specific block being processed rather than using fixed filtering rules.

Inventive Principle:
Principle #15Dynamics

2Productivity

If adaptive filtering is applied based on block dimensions, then processing efficiency improves, but filtering uniformity deteriorates

Engineering Contradiction:
Improveprocessing efficiencyVSAvoidfiltering uniformity
Core Design Contradiction:
ProductivityVSStability of the object's composition

Solution Approach 1:

The patent changes filtering parameters based on block dimensions. For non-square blocks, the filtering strength, range, or method is adjusted according to the specific width and height of the block, optimizing processing efficiency for each block type while maintaining overall system consistency through controlled parameter variations.

Inventive Principle:
Principle #35Parameter changes

3Manufacturing precision

If filtering is applied to both top and left boundaries of non-square blocks, then boundary quality is improved, but processing load increases

Engineering Contradiction:
Improveboundary qualityVSAvoidprocessing load
Core Design Contradiction:
Manufacturing precisionVSPower

Solution Approach 1:

The patent applies filtering selectively to boundaries based on their specific needs. For non-square blocks, the filtering is applied differently to top and left boundaries according to their respective dimensions and orientations, improving boundary quality where necessary while avoiding unnecessary processing on boundaries that require less filtering.

Inventive Principle:
Principle #3Local quality

Data Source

PatentEP3509299B1Image encoding/decoding method and device therefor
Publication Date: 2024.05.01 ROSEDALE DYNAMICS LLC
  • EP3509299B1 patent drawingFigure 1
  • EP3509299B1 patent drawingFigure 2
  • EP3509299B1 patent drawingFigure 3

AI summary

Disclosed are an image coding/decoding method and an apparatus therefor. Specifically, a method for decoding an image by a decoding apparatus may include: deriving an intra prediction mode of a current block; configuring reference samples with samples neighboring to an upper end and samples neighboring to a left side of the current block; generating a prediction block of the current block by using the reference sample; and filtering a sample adjacent to a top boundary and/or a left boundary of the current block among the prediction samples, in which when the current block is a non-square block, whether to filter the sample adjacent to the top boundary of the current block may be determined by considering only a width of the current block and whether to filter the sample adjacent to the left boundary of the current block may be determined by considering only a height of the current block.