Color Component Prediction Using Thresholded 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 increased processing requirements.
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 a prediction model only when the second set meets a threshold, thereby reducing computational complexity and unifying model parameter derivation processes.
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 fixed to variable. When the number of available neighbouring reference samples is less than 4, the method uses a preset component value instead of attempting to construct a prediction model, thereby reducing computational complexity while maintaining acceptable prediction accuracy for edge cases
Solution Approach 2:
The patent applies partial action by only constructing prediction models when sufficient reference samples (≥4) are available. For cases with insufficient samples, it uses simplified preset values, avoiding the excessive computational burden of attempting model construction with inadequate data
2Reliability
If additional processing is added for model parameter derivation, then prediction performance is improved, but processing time increases
Solution Approach 1:
The patent introduces dynamic processing based on the availability of reference samples. The system adapts its behavior according to the number of available samples: using preset values when samples are insufficient (fast path) and constructing prediction models only when sufficient samples are available (accurate path), thereby optimizing the balance between processing time and prediction performance
3Ease of manufacture
If uniform processing is applied for all reference sample configurations, then implementation simplicity is maintained, but computational efficiency decreases
Solution Approach 1:
The patent applies local quality by treating different reference sample configurations differently. Instead of uniform processing, it uses preset component values for cases with fewer than 4 samples and prediction models for cases with 4 or more samples, optimizing computational efficiency for each local scenario while maintaining overall implementation simplicity through clear conditional logic
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.