Image Region Segmentation Using Distance-Based Connectivity
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current gray-level segmentation methods in medical imaging struggle to accurately isolate specific anatomical structures, such as large blood vessels, due to their reliance on density-based membership values and connectedness to seed points, which can lead to partial segmentation and inclusion of similar density structures like bone.
Innovation Solution
A method that calculates connectivity values for image voxels based on their distance from a seed point, using a modified fuzzy logic approach that incorporates path distance, allowing for adjustable weighting of connectivity to improve segmentation by excluding remote structures with similar densities.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If gray-level segmentation methods use density-based membership values and connectedness to seed points, then segmentation can be performed without a priori knowledge of threshold values, but the method includes bone structures and other similar density structures along with the target structure
Solution Approach 1:
The patent divides the segmentation process into two independent components: density-based membership values and distance-based connectivity values. By separating these factors, the system can independently control which structures are included based on both density similarity and spatial proximity to the seed point, preventing inclusion of distant bone structures while maintaining adaptability to different tissue types.
Solution Approach 2:
The patent applies different weighting factors to different spatial regions relative to the seed point. Structures closer to the seed point receive higher connectivity weights, while distant structures receive lower weights. This local differentiation allows the system to prioritize nearby tissues over distant similar-density structures, improving segmentation precision without sacrificing adaptability.
2Ease of manufacture
If gray-level segmentation algorithms base membership values purely on density and connectedness to seed point, then the algorithm is simple to implement, but structures with identical density but different locations (e.g., Arterio-Venous Malformation) cannot be properly segmented
Solution Approach 1:
The patent introduces asymmetric treatment of spatial relationships by using directed distance measurements from the seed point rather than symmetric density comparisons. The connectivity value decreases with increasing distance from the seed point, creating an asymmetric weighting scheme that naturally distinguishes between nearby and distant structures even when they have identical densities, while adding minimal computational complexity.
Solution Approach 2:
The patent modifies the membership function by introducing a distance parameter alongside the density parameter. Instead of using density alone, the system computes a composite membership value that incorporates both density similarity and inverse distance weighting. This parameter expansion resolves the ambiguity of identical-density structures at different locations while maintaining algorithmic simplicity through straightforward mathematical operations.
3Quantity of substance
If the segmentation includes all voxels with similar density to the seed point, then complete structures are captured, but remote structures with similar densities are incorrectly included
Solution Approach 1:
The patent implements dynamic weighting of connectivity values based on distance from the seed point. Rather than using a static density threshold, the system applies a dynamic connectivity function that decreases with distance. This allows nearby voxels with similar density to be included while automatically excluding remote voxels, even if they have identical densities, thereby improving boundary accuracy while maintaining reasonable completeness.
Data Source
AI summary
The invention relates to a method of segmenting an image of a structure stored as a set of spatially related data points representing variations in a predetermined parameter, said method comprising the steps of selecting a seed point within the structure to be segmented, assigning to each of the data points a value of connectivity indicative of the confidence that respective areas of the data points are part of the same structure as said seed point, said value of connectivity including a function of the distance of the respective point from said seed point, establishing a threshold value for said level of connectivity and selecting for display data points meeting said threshold value.