Chroma Block Prediction with Reconstructed Luma Filtering

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image encoding/decoding methods struggle with efficient prediction of chroma blocks, particularly in high-resolution and high-quality images, leading to increased data volume and transmission/storage costs.

Innovation Solution

The method involves deriving filter coefficients for chroma block prediction using reconstructed luma blocks, applying filters based on luma and chroma component linearity, and performing prediction in sub-block units, with filter parameters determined by similarity between sub-blocks.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If high-resolution and high-quality image data is transmitted or stored using existing methods, then image quality is improved, but transmission and storage expenses increase due to increased data volume

Engineering Contradiction:
Improveimage qualityVSAvoiddata volume
Core Design Contradiction:
Measurement precisionVSQuantity of substance

Solution Approach 1:

The image data is divided into luma (Y) and chroma (CbCr) components, allowing independent processing and prediction. The chroma prediction is further segmented into sub-blocks for more precise local prediction, reducing the overall data volume while maintaining high image quality

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The luma block serves as an intermediary to predict chroma blocks. By using the already-encoded luma information as a reference for chroma prediction, the system reduces chroma data volume without requiring separate high-resolution chroma data, thus lowering transmission and storage expenses while maintaining image quality

Inventive Principle:
Principle #24Intermediary (Mediator)

2Productivity

If chroma block prediction is performed using existing methods, then encoding speed is maintained, but prediction accuracy is insufficient leading to larger data volume

Engineering Contradiction:
Improveencoding speedVSAvoidprediction accuracy
Core Design Contradiction:
ProductivityVSMeasurement precision

Solution Approach 1:

The chroma block prediction is divided into sub-blocks, allowing independent filter coefficient derivation for each sub-block. This segmentation enables more accurate local prediction while maintaining encoding efficiency through template matching and gradient-based methods that can be computed rapidly

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Filter coefficients are dynamically derived based on local image characteristics such as gradients and template matching results. By adapting the prediction parameters to local variations in the image content, the system achieves higher prediction accuracy without significantly increasing encoding complexity or reducing encoding speed

Inventive Principle:
Principle #35Parameter changes

Data Source

PatentEP4589946A1Image encoding/decoding method and recording medium storing bitstream
Publication Date: 2025.07.23 KT CORP
  • EP4589946A1 patent drawingFigure 1
  • EP4589946A1 patent drawingFigure 2
  • EP4589946A1 patent drawingFigure 3

AI summary

An image encoding/decoding method according to the present invention comprises the steps of: deriving filter coefficients for prediction of a chroma block; and on the basis of the filter coefficients, deriving a prediction sample for the chroma block by filtering a plurality of luma reconstructed samples within a luma block corresponding to the chroma block.