Image Decoding Using Block-Size-Specific Inverse Quantization

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

The existing HEVC and VVC encoding methods lack effective quantization control for zeroed-out orthogonal transform coefficients, leading to suboptimal subjective image quality due to the inability to adjust quantization based on frequency components.

Innovation Solution

An image encoding and decoding apparatus that utilizes different quantization matrices for different block sizes within a single image, allowing for separate quantization and inverse quantization of orthogonal transform coefficients, thereby enabling improved subjective image quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If zeroing out technique is used to reduce code amount, then encoding efficiency is improved, but quantization control based on frequency components is lost

Engineering Contradiction:
Improveencoding efficiencyVSAvoidquantization control capability
Core Design Contradiction:
ProductivityVSAdaptability or versatility

Solution Approach 1:

The patent segments the quantization process by introducing separate quantization matrices for different block sizes (e.g., 4×4, 8×8, 16×16, 32×32). This allows frequency-component-specific quantization control for each block size, resolving the contradiction by maintaining both compression efficiency through zeroing out and adaptability through size-specific quantization matrices.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by assigning different quantization matrices to different block sizes within the same image. Each block size receives customized quantization parameters that match its frequency characteristics, enabling localized optimization of both compression efficiency and image quality without compromising the other.

Inventive Principle:
Principle #3Local quality

2Device complexity

If same quantization matrix is used for all block sizes, then device complexity is reduced, but subjective image quality deteriorates

Engineering Contradiction:
Improvequantization matrix managementVSAvoidsubjective image quality
Core Design Contradiction:
Device complexityVSManufacturing precision

Solution Approach 1:

The patent segments the single quantization matrix into multiple size-specific quantization matrices (4×4, 8×8, 16×16, 32×32). This segmentation enables optimized quantization for each block size, significantly improving subjective image quality while managing complexity through a systematic organization of quantization parameters by block size.

Inventive Principle:
Principle #1Segmentation

3Loss of time

If quantization is performed without block size consideration, then processing time is reduced, but encoding efficiency deteriorates

Engineering Contradiction:
Improveprocessing timeVSAvoidencoding efficiency
Core Design Contradiction:
Loss of timeVSProductivity

Solution Approach 1:

The patent applies preliminary action by pre-defining multiple quantization matrices corresponding to different block sizes before the actual encoding process. This allows the encoder to quickly select and apply the appropriate quantization matrix based on block size without time-consuming calculations during encoding, thus maintaining both speed and efficiency.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS12526414B2Image decoding apparatus, image decoding method, and storage medium
Publication Date: 2026.01.13 CANON KK
  • US12526414B2 patent drawing
  • US12526414B2 patent drawing
  • US12526414B2 patent drawing

AI summary

Decoding corresponding to a first array of quantized coefficients including an N×M array corresponding to a first block and data corresponding to a second array including an N×M array corresponding to a second block. Deriving a first array of orthogonal transform coefficients from the first array of quantized coefficients by using at least a first quantization matrix of an N×M array of elements, and derives a second array of orthogonal transform coefficients from the second array of by using at least a second quantization matrix of an N×M array of elements. Performing inverse orthogonal transform on the first array of orthogonal transform coefficients to generate a P×Q array of pixels of first prediction residuals, and performs inverse orthogonal transform on the second array of orthogonal transform coefficients to generate an N×M array of pixels of second prediction residuals.