Selective Weighted Prediction for Video Brightness Change
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional video compression techniques, such as H.264, face inefficiencies in coding sequences with temporal brightness variations, leading to quality losses when weighted prediction is applied universally, even in frames without global brightness changes.
Innovation Solution
A method is introduced to detect global brightness changes during encoding by dividing frames into regions, computing brightness measures, and selectively applying weighted prediction only when a predetermined threshold is exceeded, thereby optimizing the use of weighted prediction factors.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If weighted prediction is applied universally to all frames, then coding efficiency is improved for sequences with global brightness changes, but quality loss increases for sequences without global brightness changes
Solution Approach 1:
The patent changes the parameter of weighted prediction application from universal to selective based on detected brightness change conditions. By computing brightness measures and comparing against thresholds, the system dynamically adjusts whether weighted prediction factors are applied, thereby improving coding efficiency for brightness-changing sequences while preserving quality for stable sequences.
Solution Approach 2:
The system performs self-detection of brightness changes and automatically determines whether weighted prediction should be applied. The encoder computes brightness measures, compares them to thresholds, and selectively enables weighted prediction without external control, allowing the system to serve itself in optimizing both efficiency and quality.
2Reliability
If weighted prediction factors are computed and applied, then temporal brightness variation is compensated, but computational complexity and processing overhead increase
Solution Approach 1:
The patent introduces conditional parameters (brightness change thresholds) that control whether the complex weighted prediction computation is performed. By changing the parameter of computation application based on simple brightness measure comparisons, the system achieves accurate brightness compensation only when necessary, reducing overall computational complexity.
Solution Approach 2:
Instead of applying weighted prediction computation to all frames (excessive action), the patent applies it only to frames where brightness change exceeds the threshold (partial action). This selective application reduces computational overhead while maintaining reliability for frames that actually require brightness compensation.
3Measurement precision
If frames are divided into regions for brightness measurement, then detection accuracy is improved, but processing time and computational load increase
Solution Approach 1:
The patent divides frames into multiple regions and computes brightness measures for each region independently. This segmentation improves detection accuracy by capturing local brightness variations that might be missed in a global average, while still allowing efficient processing through region-level parallel computation.
Solution Approach 2:
The system performs region-level brightness measurement only when necessary, determined by the overall brightness change detection. By applying detailed region analysis selectively rather than universally, the patent improves measurement precision for frames requiring it while minimizing processing time for frames that don't require detailed analysis.
Data Source
AI summary
A method of encoding a coding unit of a digital video sequence is provided that includes computing a brightness measure for each region in a plurality of regions in the coding unit, wherein the brightness measure indicates a relative brightness between the region in the coding unit and a corresponding region in a reference coding unit, determining a number of regions in the plurality of regions with significant brightness change by comparing each brightness measure to a region brightness change threshold, detecting whether there is global brightness change in the coding unit by comparing the number of regions to a global brightness threshold, and when global brightness change is detected, computing weighted prediction factors to be used for weighted prediction in encoding the coding unit.