Video Block Prediction Using Guided Reference Vectors for MPM Refinement

Resolve Bottlenecks,
Find Innovative Solutions
Generate Solutions

Solution Overview

Problem

Existing video coding technologies face inefficiencies in predicting current blocks using most probable mode (MPM) due to inadequate intra prediction modes derived from neighboring blocks, leading to imprecise and inefficient block reconstruction.

Innovation Solution

The use of a guiding reference vector to determine a guided reference block and relocate intra prediction modes, forming an MPM list for precise block unit prediction and reconstruction, involving determining guiding and relocated intra prediction modes based on specific positions and vectors.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Ease of operation

If intra prediction modes are derived from neighboring blocks for MPM candidates, then the prediction process is simplified, but the precision of block prediction deteriorates

Engineering Contradiction:
Improvesimplicity of prediction processVSAvoidprecision of block prediction
Core Design Contradiction:
Ease of operationVSMeasurement precision

Solution Approach 1:

The patent introduces a guiding reference block as an intermediary between neighboring blocks and the current block. Instead of directly using neighboring block modes, the system first determines a guiding reference block based on a guiding reference vector, then derives MPM candidates from this intermediary block. This mediator provides more accurate prediction information while maintaining a relatively simple process structure.

Inventive Principle:
Principle #24Intermediary (Mediator)

Solution Approach 2:

The system performs preliminary determination of a guiding reference block and its intra prediction mode before generating MPM candidates for the current block. This preliminary action ensures that the most probable modes are derived from a more relevant reference, improving prediction precision before the actual candidate selection process occurs.

Inventive Principle:
Principle #10Preliminary action

2Adaptability or versatility

If multiple MPM candidates are generated from neighboring blocks, then the coverage of potential modes increases, but the efficiency of block reconstruction decreases

Engineering Contradiction:
Improvecoverage of prediction modesVSAvoidefficiency of block reconstruction
Core Design Contradiction:
Adaptability or versatilityVSProductivity

Solution Approach 1:

The patent changes the reference parameter from neighboring block modes to guiding reference block modes. By modifying which block's intra prediction mode is used as the basis for MPM candidates, the system achieves better adaptability without requiring exhaustive search, thus maintaining reconstruction efficiency.

Inventive Principle:
Principle #35Parameter changes

3Measurement precision

If MPM refinement mode is implemented to improve prediction accuracy, then the precision of block reconstruction improves, but the complexity of the coding system increases

Engineering Contradiction:
Improveprecision of block reconstructionVSAvoidcomplexity of coding system
Core Design Contradiction:
Measurement precisionVSDevice complexity

Solution Approach 1:

The patent extracts the essential function of MPM refinement by directly using the guiding reference block's intra prediction mode to determine MPM candidates, rather than implementing a complex iterative refinement process. This extraction approach achieves high precision by focusing on the most important reference block while avoiding unnecessary computational complexity.

Inventive Principle:
Principle #2Taking out (Extraction)

Data Source

PatentUS20260012576A1Electronic device and non-transitory machine-readable medium for decoding and/or encoding video data
Publication Date: 2026.01.08 SHARP KK
  • US20260012576A1 patent drawing
  • US20260012576A1 patent drawing
  • US20260012576A1 patent drawing

AI summary

A method of decoding and/or encoding video data is provided. The method includes: receiving the video data; determining a block unit from an image frame of the video data; determining a guiding reference vector of the block unit; determining a guided reference block that is indicated by the guiding reference vector of the block unit, the guiding reference vector starting from the block unit; determining a first relocated intra prediction mode based on a first relocated position of the guided reference block, where the first relocated intra prediction mode is one of several most probable mode (MPM) candidates in an MPM list of the block unit; and reconstructing the block unit based on the MPM list of the block unit. Furthermore, an electronic device and a non-transitory machine-readable medium of an electronic device using the aforementioned method are also provided.