Color Component Prediction with Adaptive Reference Sample Screening
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 numbers of available neighbouring reference samples, lead to 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 the process and reducing complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the number of neighbouring reference samples used for model parameter derivation varies, then the prediction accuracy can be maintained under different conditions, but the computational complexity increases and additional processing is added
Solution Approach 1:
The patent applies parameter changes by dynamically adjusting the number of reference samples used for model parameter derivation based on availability conditions. When sufficient reference samples are available, more samples are used to improve prediction accuracy; when samples are limited, fewer samples are used to reduce computational complexity. This adaptive parameter adjustment resolves the contradiction between maintaining prediction accuracy and reducing computational complexity.
2Reliability
If additional processing is added to handle varying numbers of reference samples, then prediction performance is maintained, but the processing time and complexity increase
Solution Approach 1:
The patent implements dynamics by making the reference sample selection process adaptive rather than static. The system dynamically determines the number of reference samples to use based on the actual availability of neighbouring samples, allowing the prediction process to flexibly adjust to different scenarios. This dynamic approach maintains prediction performance while avoiding unnecessary processing when samples are limited.
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.