Model-Based Image Prediction for Lower-Bit Video Encoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Conventional image encoding/decoding methods face limitations in improving encoding efficiency due to redundancy of information between luma and chroma component samples and neighbor samples within a target block.

Innovation Solution

A method and device that utilize model-based prediction to reduce signaling bits by configuring a sample set and applying a mapping model for entropy-encoding/decoding, leveraging redundancy between samples.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If conventional image encoding/decoding methods are used, then image data can be transmitted and stored, but encoding efficiency is limited due to redundancy of information between luma and chroma component samples and neighbor samples within a target block

Engineering Contradiction:
Improveencoding efficiencyVSAvoidredundancy of information
Core Design Contradiction:
ProductivityVSLoss of information

Solution Approach 1:

The patent extracts and removes redundant information from the encoding process by applying model-based prediction to predict sample values. By identifying and eliminating redundant information between luma/chroma components and neighbor samples, the encoding efficiency is improved without losing essential image information.

Inventive Principle:
Principle #2Taking out (Extraction)

Solution Approach 2:

The patent changes the encoding parameters by introducing model-based prediction methods that transform how sample information is processed. Instead of directly encoding raw sample values, the system uses prediction models to generate predicted values and encodes only the differences, thereby reducing redundancy and improving encoding efficiency.

Inventive Principle:
Principle #35Parameter changes

2Productivity

If model-based prediction is applied to reduce signaling bits, then encoding efficiency is improved, but device complexity increases due to configuring sample sets and deriving mapping models

Engineering Contradiction:
Improveencoding efficiencyVSAvoidcomplexity of configuring sample set and deriving mapping model
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent segments the encoding process into distinct stages: configuring sample sets, deriving mapping models, applying predictions, and entropy encoding. By dividing the complex task into manageable segments, the system can implement model-based prediction efficiently while maintaining clarity in the encoding workflow.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent performs preliminary actions by pre-configuring sample sets and pre-deriving mapping models before the actual encoding of image data. This preparation work is done once and reused, reducing the overall complexity during the main encoding process and improving efficiency through reuse of pre-computed elements.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4697701A1Image encoding/decoding method and device using model-based prediction method, and bitstream storage medium
Publication Date: 2026.02.18 ELECTRONICS & TELECOMM RES INST
  • EP4697701A1 patent drawingFigure 1
  • EP4697701A1 patent drawingFigure 2
  • EP4697701A1 patent drawingFigure 3

AI summary

The present invention relates to an image encoding/decoding method and device, and a recording medium, the method comprising the steps of: obtaining a prediction block of a target block by performing prediction of the target block on the basis of a mapping model; and reconstructing the target block on the basis of the prediction block, wherein the mapping model may be determined by means of any one of a plurality of mapping model candidates.