NSPT Kernel Selection for Efficient Image Block Decoding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing image compression technologies face challenges in efficiently compressing high-resolution and high-quality images, particularly in determining and signaling non-separable transform kernels for improved encoding efficiency.

Innovation Solution

The method and apparatus utilize a non-separable primary transform (NSPT) kernel of reduced dimension, applied based on an encoding parameter, to derive and encode transform coefficients for current blocks, allowing for improved encoding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If non-separable primary transform kernels are used to improve transform performance, then encoding efficiency is improved, but device complexity and signaling overhead increase

Engineering Contradiction:

Improveencoding efficiencyVSAvoiddevice complexity

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The patent applies parameter changes by using reduced-dimension NSPT kernels (e.g., 4x4, 8x8) instead of full-dimension kernels, and by selectively applying NSPT based on block size parameters. This reduces computational complexity while maintaining encoding efficiency benefits for appropriate block types

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent segments the transform kernel selection process by creating different NSPT kernel sets for different block size groups (e.g., 4x4, 8x8, 16x16 blocks use different kernels). This segmentation reduces overall complexity by only computing and signaling kernels relevant to each block type

Inventive Principle:

Principle #1Segmentation

2Productivity

If non-separable primary transform kernels are used to improve transform performance, then encoding efficiency is improved, but signaling overhead increases

Engineering Contradiction:

Improveencoding efficiencyVSAvoidsignaling overhead

Core Design Contradiction:

ProductivityVSLoss of information

Solution Approach 1:

The patent changes the parameter of kernel dimension from full-size to reduced-size (e.g., using 4x4 kernels instead of larger kernels), which significantly reduces the number of coefficients that need to be signaled in the bitstream while maintaining transform effectiveness

Inventive Principle:

Principle #35Parameter changes

Solution Approach 2:

The patent implements dynamic kernel selection where the NSPT kernel set is determined based on encoding parameters such as block size and prediction mode. This dynamic adaptation allows the system to use reduced-dimension kernels only when beneficial, minimizing signaling overhead while maintaining encoding efficiency

Inventive Principle:

Principle #15Dynamics

Data Source

PatentEP4727123A1Image encoding/decoding method and apparatus, and recording medium storing bitstreams

Publication Date: 2026.04.15 LG ELECTRONICS INC

Figure 1
Figure 2
Figure 3

AI summary

An image decoding method and apparatus, according to the present disclosure, may derive transform coefficients of a current block from a bitstream, determine a non-separable primary transform (NSPT) set for the current block, determine an NSPT kernel of the current block from the NSPT set, derive residual samples by performing NSPT on the transform coefficients of the current block on the basis of the NSPT kernel, and reconstruct the current block on the basis of the residual samples. The NSPT may be applied on the basis that the size of the current block belongs to a group of one or more block sizes to which the NSPT may be applied.