Cross-Color Intra Prediction Using Linear Model Parameters
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing chroma Intra prediction methods in video coding, such as the Linear Model (LM) mode, face challenges in improving performance while managing computational complexity and buffer requirements, particularly in handling the spatial correlation between luminance and chrominance components.
Innovation Solution
The proposed method employs a cross-color Intra prediction technique using a linear model that determines LM parameters based on multiple rows or columns of neighboring pixels, allowing for reduced computational complexity by reusing existing buffers and incorporating distance-weighted predictors, thereby generating accurate predictors without increasing buffer requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If multiple rows or columns of neighboring pixels are used to determine LM parameters, then prediction accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent divides the prediction process into multiple LM modes (e.g., LM mode 0, LM mode 1, LM mode 2) that use different numbers of neighboring pixel rows/columns. This segmentation allows the system to select appropriate complexity levels based on prediction needs, resolving the contradiction between accuracy and computational load.
Solution Approach 2:
The patent implements distance-weighted predictors that use partial information from neighboring pixels, weighted by their distance from the current block. This partial action approach achieves good prediction accuracy without requiring all available neighboring pixels, thus reducing computational complexity while maintaining precision.
2Measurement precision
If multiple rows or columns of neighboring pixels are used to determine LM parameters, then prediction accuracy is improved, but buffer requirements increase
Solution Approach 1:
The patent designs buffer structures that serve multiple functions: they store neighboring pixels for LM parameter derivation, support multiple LM modes, and can be reused across different prediction operations. This multi-functionality allows accurate prediction using multiple neighboring pixel rows/columns without proportionally increasing buffer requirements.
Solution Approach 2:
The patent implements buffer management that discards less critical neighboring pixel data when memory is constrained, while recovering and retaining critical pixels needed for accurate LM parameter derivation. This selective discarding and recovering maintains prediction accuracy while managing buffer limitations.
3Loss of substance
If cross-color Intra prediction is used, then bitrate is reduced, but encoding complexity increases
Solution Approach 1:
The patent performs preliminary derivation of LM parameters using neighboring pixels before the actual prediction process. By pre-calculating these parameters and storing them in buffers, the encoding complexity during the main prediction phase is reduced, while still achieving bitrate savings through accurate cross-color prediction.
Solution Approach 2:
The patent uses already-reconstructed pixels from neighboring blocks (which are already available in the buffer from previous encoding operations) to derive LM parameters for the current block. This self-service approach uses existing data resources to reduce encoding complexity while maintaining prediction accuracy and bitrate efficiency.
Data Source
AI summary
In one implementation, a method operates by receiving neighboring reconstructed first-color pixels and current reconstructed first-color pixels of a current first-color block and receiving neighboring reconstructed second-color pixels of a current second-color block collocated with the current first-color block. The method then determines linear model (LM) parameters according to a linear model for one or more LM Intra modes. The method then receives input data associated with current second-color pixels of the current second-color block and generates a cross-color Intra predictor from the current reconstructed first-color pixels of the current first-color block using the LM parameters associated with a LM Intra mode selected from said one or more LM Intra modes. Finally, the method applies cross-color Intra prediction encoding or decoding to the current second-color pixels of the current second-color block using the cross-color Intra predictor for the selected LM Intra mode.