Subblock Intra Prediction Mode Adaptation for Video Coding Efficiency

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing video encoding techniques face inefficiencies in intra prediction, particularly when blocks are subpartitioned, as the optimal prediction mode for the current block may not be suitable for subpartitioned blocks, leading to increased overhead and reduced coding efficiency.

Innovation Solution

A method and apparatus that modify the intra prediction mode of a current block based on the shape, direction of subpartitioning, and prediction direction of subpartitioned blocks, allowing for efficient intra prediction per subblock basis by considering subblock information and adapting prediction modes accordingly.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Loss of information

If a single intra prediction mode is applied to the entire current block, then the overhead for transmitting prediction mode information is reduced, but the coding efficiency deteriorates when the block is large or has significant variation in pixel-to-reference-pixel distance

Engineering Contradiction:

Improveoverhead for transmitting intra prediction modeVSAvoidcoding efficiency

Core Design Contradiction:

Loss of informationVSManufacturing precision

Solution Approach 1:

The current block is divided into multiple subblocks, and each subblock is assigned a separate intra prediction mode. This segmentation allows different prediction modes to be applied to different regions of the block, improving coding efficiency by adapting to local characteristics while maintaining manageable overhead through shared mode information among subblocks.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different intra prediction modes are applied to different subblocks based on their local characteristics. Each subblock can have its prediction mode selected or modified according to its specific needs, such as its position, shape, and the distance to reference pixels, thereby optimizing the local prediction quality without requiring full independence for each subblock.

Inventive Principle:

Principle #3Local quality

2Manufacturing precision

If the current block is split into smaller subblocks to reduce residual energy, then the intra prediction efficiency improves, but the overhead for transmitting intra prediction modes for each subdivided block increases

Engineering Contradiction:

Improveintra prediction efficiencyVSAvoidoverhead for transmitting intra prediction mode

Core Design Contradiction:

Manufacturing precisionVSLoss of information

Solution Approach 1:

Multiple subblocks share common intra prediction mode information. Instead of independently encoding the prediction mode for each subblock, the patent merges the mode information across subblocks, allowing them to inherit or derive their prediction modes from a common source, thereby reducing the total overhead while maintaining the benefits of subblock-level prediction.

Inventive Principle:

Principle #5Merging (Combining)

Solution Approach 2:

The prediction mode information is prepared and made available at the current block level before subblock processing. This preliminary action allows subblocks to utilize pre-determined mode information, reducing the need for additional mode transmission and enabling efficient derivation of subblock-specific modes without increasing overall overhead.

Inventive Principle:

Principle #10Preliminary action

3Loss of information

If ISP technique is applied to subpartition blocks, then the overhead is reduced and intra prediction efficiency improves, but the intra prediction mode applied to the current block may not be optimal for the subpartitioned blocks

Engineering Contradiction:

ImproveoverheadVSAvoidoptimal prediction mode for subpartitioned blocks

Core Design Contradiction:

Loss of informationVSAdaptability or versatility

Solution Approach 1:

The intra prediction mode is made dynamic and adaptable to the subpartitioned blocks. Instead of applying a static mode to the entire current block, the patent enables the prediction mode to be modified or adjusted for each subblock based on its specific characteristics, such as its position, shape, and orientation, thereby achieving both low overhead and high adaptability.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The prediction mode parameters are changed or adjusted for each subblock. The patent modifies the intra prediction mode parameters (such as the prediction direction or mode index) based on subblock-specific information, allowing each subblock to use an optimal prediction mode while sharing the same base mode information, thus resolving the contradiction between overhead reduction and mode optimality.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20230319307A1Video encoding and decoding method and apparatus using subblock based intra prediction

Publication Date: 2023.10.05 HYUNDAI MOTOR CO LTD

AI summary

A video encoding/decoding method and an apparatus using intra prediction per subblock basis are provided. The video encoding/decoding method and apparatus modify the intra prediction mode of a current block in a direction suitable for a subpartitioned block by considering the shape of the subpartitioned block, the direction of subpartitioning, and the prediction direction of the current block. Based on the modified intra prediction mode of the current block, the video encoding/decoding method and apparatus generate an intra prediction mode of the subpartitioned block to perform intra prediction per subblock basis effectively.