Image Block Decoding with Reduced-Dimension Non-Separable Transform

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image compression technologies face challenges in efficiently compressing high-resolution and high-quality images due to limitations in transform methods, particularly in handling non-separable primary transforms.

Innovation Solution

The method involves partitioning a current block into sub-blocks based on odd or even horizontal and vertical coordinates, using a non-separable primary inverse transform kernel with a reduced dimension, and determining this kernel based on encoding parameters such as block size, inter prediction information, and intra prediction mode.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If non-separable primary transform is used for image compression, then transform performance is improved, but device complexity increases

Engineering Contradiction:

Improvetransform performanceVSAvoiddevice complexity

Core Design Contradiction:

Manufacturing precisionVSDevice complexity

Solution Approach 1:

The current block is divided into multiple sub-blocks, and the non-separable primary inverse transform is applied separately to each sub-block. This segmentation approach enables the use of complex transform kernels while managing computational complexity through localized processing, resolving the contradiction between improved transform performance and increased device complexity.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different transform kernels are selected and applied to different sub-blocks based on local characteristics such as prediction mode and block size. This local adaptation optimizes transform performance for each region while avoiding the application of overly complex transforms uniformly across the entire block, thus balancing performance improvement with complexity management.

Inventive Principle:

Principle #3Local quality

2Productivity

If non-separable primary inverse transform kernel of reduced dimension is used, then encoding efficiency is improved, but transform precision may be compromised

Engineering Contradiction:

Improveencoding efficiencyVSAvoidtransform precision

Core Design Contradiction:

ProductivityVSManufacturing precision

Solution Approach 1:

The transform kernel dimension is adjusted based on encoding parameters such as block size and prediction mode. By dynamically changing the kernel dimension parameter, the system achieves reduced computational complexity and improved encoding efficiency while maintaining sufficient transform precision through adaptive selection of appropriate kernel sizes for different block types.

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The transform kernel is made dynamic and adaptive rather than fixed. The system dynamically selects from multiple kernel options based on local block characteristics, allowing the transform precision to be optimized for each specific case while maintaining overall encoding efficiency. This dynamic approach resolves the contradiction by making precision requirements match the actual local needs.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS20260107017A1Image encoding/decoding method and device, and recording medium on which bitstream is stored

Publication Date: 2026.04.16 LG ELECTRONICS INC

AI summary

An image decoding method and device according to the present disclosure may induce transform coefficients of a current block from a bitstream, perform inverse quantization and inverse transformation on the transform coefficients of the current block to induce residual samples of the current block, and restore the current block on the basis of the residual samples of the current block.