Linear Model Prediction for Video Coding Blocks

Solution Overview

Problem

Current linear prediction methods in video coding face challenges in accuracy, especially for larger coding blocks where pixels farther from adjacent reconstructed pixels result in lower prediction accuracy.

Innovation Solution

The method involves dividing a current coding block into at least two sub-blocks, constructing a linear model for each sub-block using adjacent reconstructed pixels, and predicting each sub-block to improve prediction accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a single linear model is constructed for the entire current coding block using adjacent reconstructed pixels, then the coding complexity is low, but the prediction accuracy deteriorates for larger coding blocks where pixels are far from the reference pixels

Engineering Contradiction:

Improveprediction accuracyVSAvoidcoding complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The current coding block is divided into multiple sub-blocks, and a separate linear model is constructed for each sub-block using its own adjacent reconstructed pixels. This segmentation allows each sub-block to have a dedicated prediction model, improving accuracy for pixels that are far from the reference block while maintaining manageable complexity through localized processing.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

Different linear models are constructed for different sub-blocks based on their local characteristics and positions within the current coding block. Each sub-block uses its own adjacent reconstructed pixels to build a locally optimized linear model, ensuring that the prediction accuracy is adapted to local variations in the image data.

Inventive Principle:

Principle #3Local quality

2Measurement precision

If the current coding block is divided into multiple sub-blocks with separate linear models, then the prediction accuracy improves for each sub-block, but the coding complexity increases

Engineering Contradiction:

Improveprediction accuracyVSAvoidcoding complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The current coding block is divided into multiple sub-blocks, and a separate linear model is constructed for each sub-block using its own adjacent reconstructed pixels. This segmentation allows each sub-block to have a dedicated prediction model, improving accuracy for pixels that are far from the reference block while maintaining manageable complexity through localized processing.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The linear model parameters (α and β) are calculated separately for each sub-block based on the specific characteristics of that sub-block's adjacent reconstructed pixels. This parameter differentiation allows each sub-block to have optimized prediction parameters, improving overall accuracy while the systematic approach to parameter calculation keeps the complexity increase controlled.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS12301831B2Linear model prediction method and coder

Publication Date: 2025.05.13 ZHEJIANG DAHUA TECH CO LTD

AI summary

A linear model (LM) prediction method and a coder are disclosed. The method includes dividing a current coding block into at least two sub-blocks; for each of the sub-blocks, constructing a linear model, wherein parameters of the linear model are calculated by using pixel values of adjacent reconstructed pixels of the sub-block and adjacent reconstructed pixels of a reference sub-block of the sub-block; and predicting the sub-block by using the linear model, and obtaining a LM prediction value of the sub-block.