Iterative Multi-View Video Synthesis via Adaptive Refinement
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current image synthesis algorithms for immersive videos face challenges in achieving high visual quality due to errors in depth maps and computational complexity, as they use all pixels and points from multiple views, including unnecessary ones, which decreases performance and increases complexity.
Innovation Solution
An iterative image synthesis method that generates synthesis data from selected texture data, analyzes the synthesized image against a performance criterion, and iteratively refines the image by modifying synthesis data based on previous iterations, allowing for adaptive correction and optimization of visual quality and computational complexity.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If all pixels and points from multiple views are systematically used in synthesis, then synthesis completeness is improved, but computational complexity increases
Solution Approach 1:
The patent segments the synthesis process into multiple iterations, where in each iteration only a subset of pixels and points from multiple views are processed. This divides the complete synthesis task into manageable portions, reducing computational complexity per iteration while maintaining overall synthesis quality through progressive refinement.
Solution Approach 2:
The patent applies partial action by processing only a subset of pixels and points in each iteration rather than all data at once. This allows the system to achieve sufficient synthesis quality without the full computational burden of processing every pixel from all views simultaneously.
2Manufacturing precision
If all pixels from multiple views are used in synthesis, then synthesis completeness is improved, but unnecessary computations increase
Solution Approach 1:
The patent segments the synthesis process into iterations where only necessary portions of the data are processed in each step. This segmentation allows the system to avoid unnecessary computations by focusing computational resources only on areas that need refinement at each iteration stage.
Solution Approach 2:
The patent implements partial action by selectively processing only the necessary subset of pixels and points in each iteration, avoiding the excessive computation of processing all data from all views simultaneously, thus reducing energy consumption while maintaining synthesis quality.
3Reliability
If depth map errors are present in all views, then synthesis reliability is reduced, but error propagation increases
Solution Approach 1:
The patent segments the synthesis into iterations where depth map errors are gradually refined rather than propagated through all data at once. This segmentation limits error propagation to specific iteration steps, allowing correction of errors in later iterations without affecting the entire synthesis result.
Solution Approach 2:
The patent implements feedback mechanisms where the synthesis results from each iteration are used to guide subsequent iterations. This feedback loop allows the system to detect and correct depth map errors progressively, improving reliability by preventing error propagation to final output.
Data Source
AI summary
Synthesis of an image of a view from data of a multi-view video. The synthesis includes an image processing phase as follows: generating image synthesis data from texture data of at least one image of a view of the multi-view video; calculating an image of a synthesised view from the generated synthesis data and at least one image of a view of the multi-view video; analysing the image of the synthesised view relative to a synthesis performance criterion; if the criterion is met, delivering the image of the synthesised view; and if not, iterating the processing phase. The calculation of an image of a synthesised view at a current iteration includes modifying, based on synthesis data generated in the current iteration, an image of the synthesised view calculated during a processing phase preceding the current iteration.