Edge-Based Image Approximation via Weighted Intensity Projection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing image processing techniques struggle to accurately distinguish objects of interest from their surroundings due to difficulties in forming complete boundaries from one-dimensional edges, leading to challenges in identifying objects from edge detection methods.
Innovation Solution
The method involves projecting scaled intensity values from edges into positive and negative regions, generating and weighting maps to redistribute intensity values, and revising these maps to enhance object detection by 'rich get richer' schemes, which improve the system's response to objects and reduce artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If edge detection techniques are used to reduce data amount, then processing complexity is reduced, but object identification accuracy deteriorates
Solution Approach 1:
The patent transforms one-dimensional edge detection results into two-dimensional intensity maps by projecting edge information along scanlines. This dimensional transformation allows the system to work with simplified edge data while reconstructing spatial relationships that enable accurate object identification, thus resolving the contradiction between data reduction and accuracy maintenance.
Solution Approach 2:
The patent introduces positive and negative intensity maps as intermediary representations between raw edge detection and final object identification. These maps serve as mediators that preserve object characteristics while reducing computational complexity, allowing accurate object detection without processing the full original image data.
2Ease of manufacture
If traditional edge detection methods are used, then data processing is simplified, but ability to distinguish objects from surroundings deteriorates
Solution Approach 1:
The patent applies different processing treatments to different regions of the image by creating separate positive and negative intensity maps. Each map enhances specific local features (either bright or dark regions), allowing the system to maintain processing simplicity while improving object distinction through localized enhancement of relevant features.
3Loss of information
If edges are used to represent objects, then information quantity is reduced, but completeness of object boundaries deteriorates
Solution Approach 1:
The patent performs preliminary projection of edge information into intensity maps before object identification. By pre-processing edge data into comprehensive intensity representations, the system maintains boundary completeness in the intermediate maps while still working with reduced edge input data, thus preventing information loss from compromising boundary integrity.
Data Source
AI summary
Processes and systems for use in generating an approximation of an image from its edges project a respective scaled intensity value from each edge into regions abutting positive and negative sides of the edge. Positive and negative composite edge projection maps are generated, each including a respective combination of the projected scaled intensity values. A respective ratio of combined local intensity values is determined for each edge in each of the positive and negative composite maps. Respective intensity values weighted by the respective ratios are projected for each edge into regions abutting positive and negative sides of the edge. Revised positive and negative composite, weighted edge projection maps are generated, each including a respective combination of the projected, weighted intensity values.