Particle Vector Weighting for Stable Motion Vector Association
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing technologies face challenges in accurately obtaining motion vectors between frames in a motion picture, particularly in regions with less texture, leading to unstable associations and high calculation costs.
Innovation Solution
An image processing device and method that calculates corresponding vectors by arranging particle vectors around candidate blocks, assigning weights based on correlation, and statistically processing these weights to determine corresponding vectors, thereby stabilizing the association even in regions with less texture.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If conventional corresponding point detecting methods are used, then the search process can be simplified, but the association becomes unstable in regions with less texture
Solution Approach 1:
The patent divides the search process into multiple stages: first searching using simplified conventional methods to obtain candidate vectors, then refining the search by arranging particle vectors around candidate blocks and statistically processing weights. This segmentation allows the system to benefit from both the simplicity of conventional methods and the reliability of refined statistical processing.
Solution Approach 2:
The patent introduces particle vectors as an intermediary element between the candidate blocks and the final corresponding vectors. These particle vectors, arranged around candidate blocks with associated weights, serve as a mediator that refines the association process and improves stability in regions with less texture while maintaining the overall search simplicity.
2Reliability
If comprehensive search methods are used to improve association accuracy, then the reliability improves, but the calculation cost increases
Solution Approach 1:
The patent applies partial action by first performing a simplified search to obtain candidate vectors, then applying more comprehensive statistical processing only to the particle vectors arranged around candidate blocks. This partial application of comprehensive methods to specific regions (candidate blocks) improves association accuracy while avoiding the high calculation cost of comprehensive search across the entire image.
Solution Approach 2:
The patent performs preliminary action by first obtaining candidate vectors through simplified search methods before applying the more computationally intensive statistical processing. This preliminary step identifies promising regions, allowing the comprehensive method to be applied only where needed, thus improving accuracy without proportionally increasing calculation cost.
3Reliability
If recursive processing is applied to the whole image, then the association stability improves, but the calculation time increases
Solution Approach 1:
The patent segments the recursive processing by first obtaining candidate vectors through an initial search, then applying particle vector arrangement and statistical processing only to the candidate blocks. This segmentation allows recursive processing to be applied locally to promising regions rather than the entire image, improving association stability while reducing calculation time.
Solution Approach 2:
The patent applies local quality by concentrating the computationally intensive statistical processing on particle vectors arranged around candidate blocks, rather than uniformly applying it across the entire image. This localized application improves association stability in critical regions while minimizing overall calculation time.
Data Source
AI summary
A device according to embodiments may comprise an acquisition unit, an arrangement unit, a calculating unit, and a processing unit. The acquisition unit may acquire candidate vectors from among corresponding vectors which have been calculated for each candidate block around the target block. The arrangement unit may arrange particles around each candidate block indicated by each of the acquired candidate vectors while using the target block as an origination, and arrange particle vectors while using the target block as an origination. The calculating unit may calculate a correlation between a pixel value of the target block and a pixel value of each block defined by each of the particles, and give a weight depending on the calculated correlation to each of the particle vectors. The processing unit may obtain the corresponding vector for the target block based on the weight calculated for each of the particle vectors.