Layered Reference Picture Control for Lower-Load Image Coding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing image coding and decoding methods, particularly those conforming to the H.264 standard, suffer from high processing loads due to the flexibility in reference structures that allow references to multiple pictures in various directions, leading to increased computational demands.

Innovation Solution

Implement a layer-based restriction on reference images, where reference images are limited to a specific range defined by the layer to which the current image belongs, and include flags to indicate such restrictions, thereby controlling the selection of reference images during coding and decoding processes.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If motion compensation is performed for all pictures in a video sequence, then image quality is improved, but processing load increases significantly

Engineering Contradiction:
Improveimage qualityVSAvoidprocessing load
Core Design Contradiction:
Manufacturing precisionVSDevice complexity

Solution Approach 1:

The video sequence is segmented into different layers based on motion characteristics. High-motion pictures and low-motion pictures are separated into different layers, allowing motion compensation to be selectively applied only where necessary. This segmentation resolves the contradiction by maintaining image quality for high-motion regions while reducing processing load for low-motion regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different quality levels and processing approaches are applied to different parts of the video sequence based on local motion characteristics. Pictures with high motion receive full motion compensation processing, while pictures with low motion receive reduced processing. This local differentiation maintains overall image quality while reducing total processing load.

Inventive Principle:
Principle #3Local quality

2Manufacturing precision

If motion compensation is performed for all pictures, then image quality is improved, but the amount of calculation increases

Engineering Contradiction:
Improveimage qualityVSAvoidamount of calculation
Core Design Contradiction:
Manufacturing precisionVSPower

Solution Approach 1:

The video sequence is segmented into different layers based on motion characteristics. High-motion pictures and low-motion pictures are separated into different layers, allowing motion compensation to be selectively applied only where necessary. This segmentation resolves the contradiction by maintaining image quality for high-motion regions while reducing processing load for low-motion regions.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different quality levels and processing approaches are applied to different parts of the video sequence based on local motion characteristics. Pictures with high motion receive full motion compensation processing, while pictures with low motion receive reduced processing. This local differentiation maintains overall image quality while reducing total processing load.

Inventive Principle:
Principle #3Local quality

3Productivity

If pictures are classified into layers and processing is optimized, then processing load is reduced, but device complexity increases

Engineering Contradiction:
Improveprocessing load reductionVSAvoiddevice complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

Pictures are pre-classified into different layers based on motion characteristics before the main encoding process. This preliminary classification allows subsequent processing to be optimized for each layer, reducing overall processing load. The classification structure is established in advance, enabling efficient processing without requiring complex real-time decisions during encoding.

Inventive Principle:
Principle #10Preliminary action

Solution Approach 2:

The processing approach is made dynamic by adapting the encoding strategy to each picture's motion characteristics. Different encoding parameters and motion compensation strategies are applied dynamically based on the layer classification and motion magnitude, allowing optimal processing load reduction while maintaining quality where necessary.

Inventive Principle:
Principle #15Dynamics

Data Source

PatentEP4465629B1Image coding and decoding
Publication Date: 2026.05.13 SUN PATENT TRUST
  • EP4465629B1 patent drawingFigure 1
  • EP4465629B1 patent drawingFigure 2
  • EP4465629B1 patent drawingFigure 3

AI summary

With an image coding method which allows reduction of processing loads, when pictures included in a video signal are classified into layers so as to belong to their respective layers, a picture that belongs to the second layer that is located in a range restricted according to the first layer to which a current picture to be coded that is one of the pictures belongs, is referred to as a reference picture for the current picture (S10), and the current picture is coded based on the reference picture (511).