Usage-Aware Video Encoding for Human and AI Perception

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image compression technologies are not optimized for artificial intelligence services, lacking efficiency and suitability for machine tasks, and there is a need for improved encoding/decoding methods that consider image usage and machine perception.

Innovation Solution

An encoding/decoding method and apparatus that optimize image data based on its usage, determining data optimization information for human or machine tasks, and encoding this information into a bitstream for efficient transmission and decoding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If existing image compression technologies are used for high-resolution and high-quality image processing for human vision, then image quality is improved, but they are not suitable for artificial intelligence services

Engineering Contradiction:
Improveimage qualityVSAvoidsuitability for AI services
Core Design Contradiction:
Manufacturing precisionVSAdaptability or versatility

Solution Approach 1:

The patent introduces dynamic optimization by allowing the image compression system to adapt its encoding parameters based on the intended usage. The decoder receives usage information (human perception or machine task) and dynamically adjusts decoding parameters to optimize for the specific application, making the system versatile for both high-quality human viewing and AI processing tasks

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The patent changes key parameters of the image compression system based on usage type. For human perception, it optimizes for visual quality metrics, while for machine tasks, it optimizes for task performance metrics. This parameter adaptation allows the same compression framework to achieve both high image quality for humans and high effectiveness for AI services

Inventive Principle:
Principle #35Parameter changes

2Manufacturing precision

If image compression is optimized for human vision, then human perception quality is improved, but encoding/decoding efficiency for machine tasks deteriorates

Engineering Contradiction:
Improvehuman perception qualityVSAvoidencoding/decoding efficiency for machine tasks
Core Design Contradiction:
Manufacturing precisionVSProductivity

Solution Approach 1:

The patent segments the image processing pipeline into distinct optimization paths: one for human perception and another for machine tasks. The system receives usage information that determines which path to follow, allowing specialized optimization for each segment without compromising the other. This segmentation enables efficient machine task processing when that usage is detected

Inventive Principle:
Principle #1Segmentation

3Adaptability or versatility

If general-purpose image compression is used, then broad applicability is improved, but specialized optimization for specific usage deteriorates

Engineering Contradiction:
Improvebroad applicabilityVSAvoidspecialized optimization performance
Core Design Contradiction:
Adaptability or versatilityVSManufacturing precision

Solution Approach 1:

The patent creates a universal image compression framework that can serve multiple functions by incorporating usage information handling. The same compression core processes images for both human perception and machine tasks, with the optimization parameters adapting based on the detected usage type. This multi-functionality maintains broad applicability while achieving specialized optimization performance

Inventive Principle:
Principle #6Universality (Multi-functionality)

Data Source

PatentEP4723628A1Encoding/decoding method and apparatus based on video optimization according to video usage, and method for transmitting bitstream
Publication Date: 2026.04.08 LG ELECTRONICS INC
  • EP4723628A1 patent drawingFigure 1
  • EP4723628A1 patent drawingFigure 2
  • EP4723628A1 patent drawingFigure 3

AI summary

The present disclosure relates to an encoding/decoding method and apparatus for optimizing quantization according to video usage, a method for transmitting a bitstream, and a bitstream storage medium. A decoding method according to an embodiment of the present disclosure may comprise the steps of: acquiring data optimization information related to a video data optimization method according to the purpose of video data; and reconstructing the video data on the basis of the data optimization information, wherein the data optimization information includes data optimization application information indicating whether optimization is performed for each usage of the video data.