Adaptive Loop Filtering Coefficient Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current image/video compression technologies face challenges in efficiently compressing, transmitting, and storing high-resolution, high-quality images/videos, especially with the increasing demand for immersive media, leading to increased transmission and storage costs, and require improved methods for adaptive loop filtering to enhance compression efficiency and visual quality.

Innovation Solution

A method and apparatus for efficiently applying adaptive loop filtering (ALF) using a 0-th order exponential Golomb scheme for parsing ALF filter coefficients, which reduces operational overhead and complexity, and fixing the range of values for ALF coefficients to enhance image/video coding efficiency.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If adaptive loop filtering is applied to enhance compression efficiency and visual quality, then compression efficiency and visual quality are improved, but computational overhead and complexity increase

Engineering Contradiction:
Improvecompression efficiencyVSAvoidcomputational overhead
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent applies parameter changes by using exponential Golomb coding to represent ALF filter coefficients, transforming the representation method of coefficient values. This coding scheme efficiently represents a wide range of coefficient values with variable bit-length, reducing the bitrate required to signal filter parameters while maintaining filtering effectiveness, thus improving compression efficiency without proportionally increasing computational complexity

Inventive Principle:
Principle #35Parameter changes

Solution Approach 2:

The patent segments the ALF parameter signaling by dividing coefficients into different groups (e.g., luma and chroma components, different coefficient positions) and applying differential coding or constrained ranges to each segment. This segmentation allows the encoder to signal only necessary parameter variations rather than full parameter sets, reducing computational overhead while maintaining visual quality

Inventive Principle:
Principle #1Segmentation

2Measurement precision

If ALF-related information is signaled with high precision, then filtering accuracy is improved, but transmission overhead increases

Engineering Contradiction:
Improvefiltering accuracyVSAvoidtransmission overhead
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The patent uses copying by deriving certain ALF coefficients from previously coded coefficients or from template filter sets. Instead of signaling all coefficients independently, the encoder copies reference coefficient values and signals only differences or selection indices, maintaining filtering accuracy while significantly reducing transmission overhead through redundant information elimination

Inventive Principle:
Principle #26Copying

Solution Approach 2:

The patent applies partial action by signaling ALF parameters at selective locations (e.g., only at block boundaries or for significant coefficients) rather than for every coefficient. This partial signaling approach maintains sufficient filtering accuracy for visual quality while reducing the total amount of parameter information that must be transmitted

Inventive Principle:
Principle #16Partial or excessive action

Data Source

PatentUS20240430418A1Adaptive loop filtering-based image coding apparatus and method
Publication Date: 2024.12.26 LG ELECTRONICS INC
  • US20240430418A1 patent drawing
  • US20240430418A1 patent drawing
  • US20240430418A1 patent drawing

AI summary

According to embodiment(s) of the present document, computational overhead and complexity can be reduced by using a 0th-order exponential Golomb coding scheme (ue(v)) for procedures of parsing syntax elements/information about absolute values of luma/chroma ALF filter coefficients. In addition, coding using ue(v) can be efficiently performed by fixing the range of values of the information about the absolute values of the luma/chroma ALF filter coefficients.