Adaptive Intra-Prediction Padding for Reference Frame Borders
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current video coding standards face challenges in handling referenced pixel areas that extend beyond the boundaries of the reference frame, leading to suboptimal prediction quality, particularly in omnidirectional video content with varying viewing angles.
Innovation Solution
A video codec that selects the most appropriate angular intra-prediction mode for padding based on the characteristics of the current predicted block, using a probe area outside the reference picture to evaluate and compare different modes for improved approximation.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If boundary padding is performed using a single fixed intra-prediction mode (e.g., perpendicular to frame border), then the device complexity is reduced and processing is simplified, but the prediction quality deteriorates when the referenced pixel area extends beyond the reference frame boundaries
Solution Approach 1:
The patent applies dynamics by making the intra-prediction mode adaptive rather than fixed. The system dynamically selects from multiple intra-prediction modes (including perpendicular, parallel, and angular modes) based on the characteristics of the current block and the extent to which the referenced pixel area extends beyond the reference frame. This dynamic adaptation resolves the contradiction by maintaining simplicity for cases where fixed modes work well while enabling quality improvement when boundary extension is needed.
Solution Approach 2:
The patent changes the parameter of intra-prediction mode selection from a fixed value to a variable that depends on block characteristics and boundary extension requirements. By evaluating multiple modes and selecting the most appropriate one based on the specific situation, the system improves prediction quality without requiring a complete redesign of the padding process.
2Manufacturing precision
If multiple intra-prediction modes are evaluated and selected based on block characteristics, then the prediction quality is improved, but the computational complexity and processing time increase
Solution Approach 1:
The patent applies partial action by evaluating multiple intra-prediction modes only when necessary - specifically when the referenced pixel area extends beyond the reference frame boundaries. For cases where the reference area is fully contained within the frame, the system uses a default mode without extensive evaluation. This selective approach improves quality when needed while minimizing unnecessary processing overhead.
Solution Approach 2:
The system performs preliminary evaluation of block characteristics and boundary extension requirements before committing to a specific intra-prediction mode. By assessing the situation in advance and selecting the most appropriate mode beforehand, the system avoids iterative optimization and reduces processing time during the actual padding operation.
3Manufacturing precision
If the referenced pixel area is extended beyond the reference frame using simple boundary copying, then the processing is simple and fast, but the approximation of the current predicted block becomes inaccurate
Solution Approach 1:
The patent applies local quality by using different intra-prediction modes for different regions and situations. Instead of applying a single uniform padding method throughout, the system selects specific modes (perpendicular, parallel, or angular) based on the local characteristics of each block and its boundary extension requirements. This localized adaptation improves approximation accuracy while maintaining implementation feasibility through clear mode selection criteria.
Data Source
Figure 1
Figure 2
Figure 3~4
AI summary
Video codec for supporting temporal inter-prediction, configured to perform padding of an area of a referenced portion of a reference picture which extends beyond a border of the reference picture, which referenced portion is referenced by an inter predicted block of a current picture by selecting one of a plurality of intra-prediction modes, and padding the area using the selected intra-prediction mode.