Intra Prediction Using Multiple Reference Lines for Video Decoding

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video encoding/decoding technologies face challenges in efficiently performing intra-prediction for high-resolution and stereographic image content, particularly in selecting and utilizing multiple reference lines for improved compression efficiency.

Innovation Solution

A method and apparatus that derive and select multiple reference sample lines for intra-prediction, applying different reference lines to different regions of a current block based on its shape and size, and adaptively determining the intra prediction mode and number of modes used.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Manufacturing precision

If conventional image compression technology is used for high-resolution images, then transmission and storage costs increase, but image quality and resolution requirements cannot be met

Engineering Contradiction:
Improveimage qualityVSAvoidtransmission and storage costs
Core Design Contradiction:
Manufacturing precisionVSLoss of energy

Solution Approach 1:

The patent changes the parameter of reference line selection by deriving multiple reference sample lines (first, second, and third reference lines) from different neighboring blocks instead of using a single reference line. This allows adaptive selection of the most appropriate reference line based on the current block's characteristics, improving prediction accuracy and compression efficiency for high-resolution images without proportionally increasing transmission costs

Inventive Principle:
Principle #35Parameter changes

2Measurement precision

If multiple reference lines are used for intra-prediction, then prediction accuracy improves, but processing complexity increases

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

Solution Approach 1:

The patent segments the reference sampling process by dividing it into multiple distinct reference lines (first reference line from top neighboring block, second reference line from left neighboring block, third reference line from diagonal neighboring block). Each reference line is derived independently from specific neighboring blocks, allowing selective use of multiple references without requiring complete processing of all possible references, thus managing complexity while improving accuracy

Inventive Principle:
Principle #1Segmentation

Solution Approach 2:

The patent applies local quality by selecting different reference lines based on the specific characteristics and position of the current block. The encoder/decoder determines which reference line (first, second, or third) is most appropriate for each current block based on its neighboring blocks' availability and quality, rather than uniformly applying the same reference selection strategy to all blocks

Inventive Principle:
Principle #3Local quality

3Productivity

If adaptive reference line selection is implemented, then compression efficiency improves, but encoding/decoding complexity increases

Engineering Contradiction:
Improvecompression efficiencyVSAvoidencoding/decoding complexity
Core Design Contradiction:
ProductivityVSDevice complexity

Solution Approach 1:

The patent performs preliminary action by pre-defining three specific reference line derivation methods before the actual encoding/decoding process. The first reference line is derived from the top neighboring block, the second from the left neighboring block, and the third from the diagonal neighboring block. These reference lines are prepared in advance, allowing the encoder/decoder to simply select among them based on block characteristics without performing complex real-time analysis

Inventive Principle:
Principle #10Preliminary action

Data Source

PatentEP4407990B1Video signal processing method and device
Publication Date: 2025.12.24 KT CORP
  • EP4407990B1 patent drawingFigure 1
  • EP4407990B1 patent drawingFigure 2
  • EP4407990B1 patent drawingFigure 3~4

AI summary

A method for decoding a video according to the present invention may comprise: deriving a plurality of reference sample lines for a current block, selecting at least two among the plurality of reference sample lines, and performing intra prediction for the current block based on the selected reference sample lines.