Intra Sub-Partition Prediction for VPDU-Aligned Image Decoding

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

The increasing demand for high-resolution and high-quality images leads to increased transmission and storage costs due to the higher amount of information required, necessitating a more efficient image compression technique.

Innovation Solution

The implementation of Intra Sub-Partitions (ISP) in image coding systems, considering the decoding pipeline for hardware, to enhance image coding efficiency by applying ISP to blocks smaller than or equal to the size of a Virtual Pipeline Data Unit (VPDU) and addressing issues of miss-aligned VPDU processing.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If Intra Sub-Partitions (ISP) are applied to blocks larger than VPDU size, then image compression efficiency is improved, but miss-aligned VPDU processing issues occur and hardware implementation becomes difficult

Engineering Contradiction:

Improveimage compression efficiencyVSAvoidVPDU processing alignment

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The patent divides the current block into multiple sub-partitions (first sub-partition and second sub-partition) when the block size exceeds VPDU dimensions. This segmentation allows each sub-partition to be processed independently within the VPDU framework, preventing misalignment issues while maintaining the compression benefits of fine-grained intra prediction.

Inventive Principle:

Principle #1Segmentation

2Measurement precision

If ISP is applied without considering VPDU size constraints, then prediction accuracy improves, but hardware decoding pipeline implementation becomes complex

Engineering Contradiction:

Improveprediction accuracyVSAvoiddecoding pipeline complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent applies different processing strategies based on the local characteristics of the current block relative to VPDU boundaries. When a block exceeds VPDU size, it is segmented into sub-partitions that align with VPDU dimensions, allowing the decoding pipeline to process each region with appropriate complexity while maintaining overall prediction accuracy.

Inventive Principle:

Principle #3Local quality

3Quantity of substance

If conventional intra prediction is used for high-resolution images, then transmission cost increases due to larger data volume, but implementation simplicity is maintained

Engineering Contradiction:

Improvedata transmission volumeVSAvoidcoding technique complexity

Core Design Contradiction:

Quantity of substanceVSDevice complexity

Solution Approach 1:

The patent dynamically selects between conventional intra prediction and ISP-based prediction based on block size relative to VPDU dimensions. For blocks within VPDU size, conventional methods are used; for larger blocks, dynamic segmentation into sub-partitions is applied. This dynamic approach reduces transmission volume for high-resolution images while maintaining implementation feasibility through conditional logic.

Inventive Principle:

Principle #15Dynamics

Data Source

PatentUS12388996B2Intra prediction method and device based on intra sub-partitions in image coding system

Publication Date: 2025.08.12 LG ELECTRONICS INC

AI summary

A method by which a decoding device decodes an image, according to the present document, comprises the steps of: receiving prediction mode information about a current block; receiving intra sub-partitions (ISP)-related information about the current block on the basis of the size and the maximum conversion block size of the current block; deriving an intra prediction mode of the current block on the basis of the prediction mode information; and generating a prediction sample of the current block on the basis of the intra prediction mode and the ISP-related information.