VVC Colour Component Prediction with Padded Reference Samples
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The computational complexity and non-uniformity in model parameter derivation for colour component prediction in video coding, particularly in VVC, due to varying availability of neighbouring reference samples, leading to additional processing and inefficiencies.
Innovation Solution
A method for colour component prediction that determines a first reference sample set, uses a preset value when samples are insufficient, screens for a second set when sufficient, and derives model parameters only when the second set meets a threshold, thereby standardizing and reducing computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If a prediction model is constructed using neighbouring reference samples, then prediction accuracy is improved, but computational complexity increases due to additional processing requirements
Solution Approach 1:
The patent changes the parameter of reference sample quantity from variable (0-4 samples) to fixed (exactly 4 samples). When fewer than 4 neighbouring reference samples are available, the method fills them with padding values to construct a complete set of 4 samples. This parameter transformation unifies the processing flow and eliminates the need for multiple prediction models, thereby reducing computational complexity while maintaining prediction accuracy.
Solution Approach 2:
The patent uses padding values to copy/create virtual reference samples when the actual number of neighbouring reference samples is insufficient. This copying mechanism ensures that the prediction model always receives exactly 4 reference samples, standardizing the input and simplifying the computational process without sacrificing prediction quality.
2Reliability
If additional processing is added to handle varying sample availability, then prediction reliability is improved, but processing time increases
Solution Approach 1:
The patent creates equipotential processing conditions by ensuring that every coding block receives exactly 4 reference samples through padding. This eliminates the variability in reference sample quantity that previously required complex conditional processing. The unified processing flow reduces decision-making overhead and simplifies the algorithm, thereby decreasing processing time while maintaining reliable predictions.
3Measurement precision
If multiple prediction models are constructed for different sample quantities, then prediction accuracy is improved, but device complexity increases
Solution Approach 1:
The patent creates a universal prediction model that can handle all cases (0-4 neighbouring reference samples) by standardizing the input to always contain exactly 4 samples. This single multi-functional model replaces the need for multiple specialized models, reducing device complexity while maintaining high prediction accuracy across different scenarios.
Data Source
AI summary
Colour component prediction method is provided, which includes that: first reference sample set corresponding to colour component to be predicted of coding block in video image is acquired; when available sample number in first reference sample set is less than preset number, preset component value is taken as predicted value corresponding to the colour component to be predicted; when available sample number in first reference sample set is not less than preset number, first reference sample set is screened to obtain second reference sample set; when available sample number in second reference sample set is equal to preset number, model parameter is determined through second reference sample set, and prediction model corresponding to colour component to be predicted is obtained based on model parameter, prediction model is used for prediction processing of colour component to be predicted to obtain predicted value corresponding to colour component to be predicted.