Multi-view Decoding via Predictive Path Selection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current multi-view image decoding methods are computationally complex and costly, with high latency, especially when navigating through large sequences of views, due to the need to decode all preceding views, which is not suitable for free navigation and is resource-intensive.
Innovation Solution
The proposed solution involves predicting the current view using previously decoded views and storing only necessary residues, allowing the decoder to choose the optimal path based on user navigation, reducing computational complexity and storage costs by selectively decoding views relevant to the navigation path and potentially future movements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If all preceding views are decoded systematically to enable free navigation, then navigation flexibility is improved, but decoder computational complexity increases significantly
Solution Approach 1:
The patent applies preliminary action by predicting and preparing multiple potential future views before they are actually requested by the user. The decoder anticipates which views will be needed next and decodes them in advance, storing them in a buffer. This eliminates the need to systematically decode all preceding views while maintaining free navigation capability, as the needed views are already prepared.
Solution Approach 2:
The patent implements dynamics by making the decoding process adaptive to user navigation patterns. Instead of a static systematic decoding approach, the decoder dynamically adjusts which views to decode based on the user's actual navigation behavior. The system continuously monitors navigation patterns and updates its prediction of future views accordingly, optimizing computational resources in real-time.
2Reliability
If all preceding views are decoded systematically, then complete view availability is improved, but decoding latency increases
Solution Approach 1:
The patent applies skipping by selectively bypassing the decoding of views that are not needed for future navigation. Instead of decoding all preceding views systematically, the system identifies and skips views that will not be requested, rushing through only the necessary decoding operations. This significantly reduces decoding latency while maintaining complete availability of needed views through predictive pre-decoding.
3Productivity
If multiple prediction residues are stored for different navigation paths, then navigation efficiency is improved, but storage requirements increase
Solution Approach 1:
The patent applies local quality by storing prediction residues selectively based on local navigation patterns rather than uniformly for all possible paths. The system analyzes the user's specific navigation behavior and stores residues only for views that are locally relevant to the current navigation context. This optimizes storage usage by focusing resources on the most probable navigation paths while maintaining high navigation efficiency.
4Speed
If the decoder predicts future views and pre-decodes them, then navigation responsiveness is improved, but computational complexity increases
Solution Approach 1:
The patent applies partial action by predicting and pre-decoding only a subset of future views rather than all possible views. The system uses navigation pattern analysis to identify the most likely next views and pre-decodes only those, performing partial pre-decoding rather than exhaustive pre-decoding. This reduces computational complexity while still improving navigation responsiveness for the actually requested views.
Data Source
Figure 1A
Figure 1B
Figure 2A
AI summary
The invention relates to the decoding of a current view belonging to a previously coded multi-view image, in the course of which the following is implemented: - determination (D24a)), in a set of views belonging to said multi-view image or to another multi-view image, of a pathway of views necessary for the decoding of said current view, said pathway comprising at least one as yet undecoded view, - decoding (D25a)) of said at least one as yet undecoded view, - decoding (D26a)) of said current view on the basis of a set of coded data representative of at least one difference between said current view and another view of said multi-view image or of another multi-view image, and of said at least one decoded view.