Encoder Quantization Matrix Selection for Video Coding Efficiency

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face challenges in optimizing coding efficiency and reducing processing complexity while maintaining subjective image quality, especially when dealing with increasing amounts of digital video data.

Innovation Solution

An encoder and decoder system that dynamically determines whether to use a quantization matrix based on the presence of orthogonal and secondary transforms, allowing for efficient quantization of transform coefficients and reducing processing complexity without compromising subjective image quality.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If quantization matrix is used for all transform coefficients, then coding efficiency is improved, but device complexity and processing overhead increase

Engineering Contradiction:
Improvecoding efficiencyVSAvoidprocessing complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies different quantization strategies to different regions of the transform coefficient block. Specifically, it uses a first quantization matrix for a first region and a second quantization matrix for a second region, allowing optimization of coding efficiency in specific areas without unnecessarily increasing complexity across the entire block. This selective application of quantization matrices resolves the contradiction by making the processing complexity localized rather than universal.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If quantization matrix is applied to all blocks, then subjective image quality is maintained, but encoding speed decreases

Engineering Contradiction:
Improvesubjective image qualityVSAvoidencoding speed
Core Design Contradiction:
Manufacturing precisionVSSpeed

Solution Approach 1:

The patent divides the transform coefficient block into multiple regions and applies quantization matrices selectively to different regions based on their characteristics. This segmentation allows the encoder to maintain image quality in critical regions while skipping quantization matrix application in regions where it would not significantly impact quality, thereby improving encoding speed without sacrificing overall subjective image quality.

Inventive Principle:
Principle #1Segmentation

3Productivity

If multiple quantization matrices are used for different regions, then coding efficiency is improved, but ease of operation decreases

Engineering Contradiction:
Improvecoding efficiencyVSAvoidencoding process simplicity
Core Design Contradiction:
ProductivityVSEase of operation

Solution Approach 1:

The patent determines the division of regions and selection of quantization matrices in advance based on transform coefficient characteristics, and encodes this information in the bitstream. This preliminary action allows the decoder to reconstruct the same region division and quantization matrix selection without complex real-time decision-making, thereby maintaining coding efficiency while simplifying the encoding process operation.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentUS20250071279A1Encoder, decoder, encoding method, and decoding method
Publication Date: 2025.02.27 PANASONIC INTELLECTUAL PROPERTY CORP OF AMERICA
  • US20250071279A1 patent drawing
  • US20250071279A1 patent drawing
  • US20250071279A1 patent drawing

AI summary

An encoder includes circuitry and memory coupled to the circuitry. In operation, the circuitry: performs quantization on a plurality of transform coefficients of a current block to be encoded, using a quantization matrix when orthogonal transform is performed on the current block and secondary transform is not performed on the current block; and performs quantization on the plurality of transform coefficients of the current block without using the quantization matrix when orthogonal transform is not performed on the current block and when both orthogonal transform and secondary transform are performed on the current block.