Approximate Pyramidal Search for Displacement Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing methods for displacement matching, such as stereoscopic imaging and video compression, face inefficiencies and noise issues due to exhaustive searches at high resolutions, which can lead to local minima and increased computational time.
Innovation Solution
An accelerated pyramid search method that performs a complete window-based search at the coarsest resolution level, counts displacement matches, selects displacements based on an energy threshold, and refines the search at higher resolutions with reduced windows around selected displacements, thereby reducing noise and computational load.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If exhaustive search at high resolution is performed, then matching accuracy is improved, but computational time increases significantly
Solution Approach 1:
The search space is segmented into multiple resolution levels (coarse to fine). At each level, only relevant displacement candidates are selected and refined, rather than performing exhaustive search at all resolutions. This segmentation reduces the total search complexity while maintaining accuracy through progressive refinement.
Solution Approach 2:
A complete search is performed at the coarsest resolution level first to identify candidate displacements before refining at higher resolutions. This preliminary action at low resolution filters out unlikely candidates, reducing the search burden at computationally expensive high resolution levels.
2Measurement precision
If complete window-based search is performed at all resolution levels, then matching precision is improved, but computational load increases
Solution Approach 1:
Instead of performing complete window-based search at all resolution levels, the method performs partial search by selecting only the most promising displacement candidates based on coarse-level results. This partial action approach maintains precision where needed while reducing computational load in less critical areas.
Solution Approach 2:
The search strategy applies different levels of thoroughness to different displacement candidates. High-resolution complete search is applied only to selected candidates that show promise at coarse resolution, while other candidates receive less intensive processing. This local quality differentiation optimizes the balance between precision and computational load.
3Productivity
If coarse-to-fine pyramid approach is used, then search speed is improved, but noise and errors may be introduced
Solution Approach 1:
The method uses feedback from coarse-resolution search results to guide and constrain the search at finer resolution levels. The selected displacements from coarse levels serve as feedback that informs the search strategy at next levels, ensuring that refinements stay within plausible ranges and reducing the risk of noise-induced errors.
Solution Approach 2:
The coarse-resolution search acts as a cushioning layer before the finer search. By identifying candidate displacements at low resolution first, the method creates a buffer that protects against noise and local minima in high-resolution search, as the coarse-level results provide a broader contextual framework for interpretation.
Data Source
AI summary
A method for approximate pyramidal search for displacement matching is provided that includes performing a complete window-based search at a coarsest resolution level in a resolution pyramid to determine a best matching point in a destination image for each point in a source image, wherein a displacement from each point to the best matching point is determined, counting, for each displacement of the determined displacements, a number of points having the displacement, selecting displacements from the determined displacements based on the number of points counted for each displacement and a predetermined threshold, and performing a search at the next finer resolution level in the resolution pyramid to determine a best matching point in the destination image for each point in the source image, wherein the search for a best matching point for each point is limited to the selected displacements and a reduced search window around each displacement.