Video Decoding Intra Prediction Accuracy via SDIP

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Current video decoding methods face accuracy degradation in intra prediction as the distance between reference and prediction samples increases, particularly with square-shaped partitions, necessitating improved methods for error reduction.

Innovation Solution

The method introduces rectangular or line-shaped prediction units for intra blocks, utilizing Short Distance Intra Prediction (SDIP) to decrease prediction error by deriving intra prediction modes based on block partition information and directionality, with offset values determined by direction information to enhance prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of manufacture

If square-shaped partitions are used for intra prediction, then device complexity is reduced and ease of manufacture is improved, but prediction accuracy deteriorates as distance between reference sample and prediction sample increases

Engineering Contradiction:
Improveease of implementationVSAvoidprediction accuracy
Core Design Contradiction:
Ease of manufactureVSMeasurement precision

Solution Approach 1:

The current coding block is divided into multiple prediction units with different shapes (square, rectangular, line-shaped) based on the distance from reference samples. This segmentation allows each region to use the most appropriate prediction shape, improving overall prediction accuracy while maintaining implementation feasibility through standardized shape categories.

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

Different prediction unit shapes are applied to different regions within the coding block based on local characteristics - specifically the distance from reference samples. Regions closer to reference samples use one shape while distant regions use another, optimizing prediction accuracy for each local area rather than applying a uniform square partition throughout.

Inventive Principle:
Principle #3Local quality

2Measurement precision

If rectangular or line-shaped prediction units are introduced to improve prediction accuracy, then prediction error decreases, but device complexity and processing power requirements increase

Engineering Contradiction:
Improveprediction accuracyVSAvoidprocessing complexity
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The prediction unit shape is dynamically selected based on the distance from reference samples and the specific coding block characteristics. The system adapts between square, rectangular, and line-shaped prediction units according to local requirements, achieving high prediction accuracy without requiring all complex shapes to be processed uniformly across the entire image.

Inventive Principle:
Principle #15Dynamics

Solution Approach 2:

The invention changes the shape parameter of prediction units from fixed square to variable shapes (rectangular, line-shaped) based on distance parameters from reference samples. This parameter adaptation allows the system to achieve better prediction accuracy while controlling complexity through conditional shape selection rather than universal complex processing.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If intra prediction mode is derived based on block partition information and directionality, then prediction accuracy improves, but loss of information increases due to additional flag information transmission

Engineering Contradiction:
Improveprediction accuracyVSAvoidbitrate increase
Core Design Contradiction:
Measurement precisionVSLoss of information

Solution Approach 1:

The invention applies partial action by transmitting flag information only when rectangular or line-shaped prediction units are used, rather than encoding all prediction modes with full directionality information. This selective transmission reduces the overall bitrate penalty while still providing accurate prediction where non-square shapes are beneficial.

Inventive Principle:
Principle #16Partial or excessive action

Solution Approach 2:

The prediction unit shape and basic direction are determined preliminarily based on block partition information before detailed mode encoding. This preliminary determination allows the system to establish a baseline prediction accuracy without requiring extensive additional information, reducing the overall information loss.

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP3282700B1Method decoding video signals
Publication Date: 2021.07.07 LG ELECTRONICS INC
  • EP3282700B1 patent drawingFigure 1
  • EP3282700B1 patent drawingFigure 2
  • EP3282700B1 patent drawingFigure 3(a)~3(b)

AI summary

The present invention provides a method for decoding a video signal including obtaining block partition information of a current coding unit; deriving an intra prediction mode of a partition block based on the block partition information; and generating a prediction signal of the partition block based on the derived intra prediction mode, and the intra prediction mode of the partition block is defined by a limited number based on a prediction direction.