Video Encoding Reference Picture List Segmentation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

State-of-the-art video coding schemes, such as MPEG-4 AVC/H.264 and HEVC, always prioritize reference pictures closest to the current picture for bi-predictive inter-picture prediction, which may not always be the best choice for optimal coding efficiency and picture quality.

Innovation Solution

The method involves creating two reference lists for inter-picture prediction, where one list is ordered based on the temporal level of reference pictures and the other on their proximity to the current picture, allowing for hierarchical coding structure that improves coding efficiency and picture quality by prioritizing pictures at lower temporal levels for better quality encoding.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Reliability

If reference pictures are sorted by temporal distance to the current picture, then the most recent reference pictures are prioritized for prediction, but coding efficiency and picture quality deteriorate when the closest reference pictures are not the best for prediction

Engineering Contradiction:

Improveprediction accuracyVSAvoidcoding efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The reference picture list is segmented into multiple lists (first reference picture list and second reference picture list) with different sorting criteria. The first list is sorted by temporal level while the second list is sorted by temporal distance, allowing the encoder to select the most appropriate list for each prediction operation, thus resolving the contradiction between prediction accuracy and coding efficiency.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The reference picture list ordering is made dynamic by allowing switching between different sorting methods (temporal level-based and temporal distance-based) depending on the prediction requirements. This dynamic adaptation enables the system to optimize for either prediction accuracy or coding efficiency based on the specific coding context.

Inventive Principle:

Principle #15Dynamics

2Device complexity

If reference pictures are sorted by temporal level, then hierarchical coding structure is improved and bits for reordering signals are minimized, but access to nearest reference pictures becomes less direct

Engineering Contradiction:

Improvereference list structureVSAvoidreference picture access

Core Design Contradiction:

Device complexityVSEase of operation

Solution Approach 1:

The reference picture management is segmented into multiple organized lists with different sorting criteria. The first list is organized by temporal level for hierarchical coding, while the second list is organized by temporal distance for direct access. This segmentation allows both hierarchical structure and easy access to be maintained simultaneously through appropriate list selection.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Multiple reference picture lists act as intermediaries between the raw reference pictures and the prediction process. These intermediary lists with different sorting methods enable flexible access patterns, allowing the system to balance hierarchical organization with operational ease depending on the specific coding needs.

Inventive Principle:

Principle #24Intermediary (Mediator)

3Adaptability or versatility

If multiple reference picture lists with different sorting methods are created, then flexibility in picture selection is improved, but system complexity and processing overhead increase

Engineering Contradiction:

Improvereference picture selection flexibilityVSAvoidreference list management

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The reference picture management system is segmented into multiple pre-organized lists with different sorting criteria. This segmentation provides flexibility in picture selection while controlling complexity by maintaining fixed, pre-defined sorting methods for each list, avoiding the need for complex dynamic reordering operations.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system changes the parameter of reference picture ordering from a single sorting method to multiple sorting methods organized in separate lists. This parameter change increases adaptability by allowing selection based on different criteria while managing complexity through structured organization of the multiple ordering approaches.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS10841573B2Methods and apparatuses for encoding and decoding video using multiple reference pictures

Publication Date: 2020.11.17 SUN PATENT TRUST

AI summary

A video encoding method using reference pictures includes: writing a parameter indicating a temporal level of a reference picture or a period of a type of a reference picture into the reference picture for each of the reference pictures, creating a first list of reference pictures comprising the reference pictures sorted based on the parameter; and encoding a current picture of the video using at least the first list of reference pictures. A video decoding method using a reference pictures includes: parsing a parameter indicating a temporal level of a reference picture or a period of a type of a reference picture from each of the reference pictures; creating a first list of reference pictures comprising the reference pictures sorted based on the parameter; and decoding a current picture of the video using at least the first list of reference pictures. Corresponding video encoding and decoding apparatuses are provided.