Weighted EMST for Pulmonary Artery and Vein Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods fail to accurately distinguish between pulmonary arteries and veins in 3D CT images and quantify small blood vessels due to complex morphological structures and overlapping vessels, making it difficult to assess lung diseases effectively.
Innovation Solution
A method using a weight-based Euclidean Minimum Spanning Tree (EMST) to classify and quantify blood vessels by assigning intensity and local shape weights to points, separating the tree into regions to distinguish between arteries and veins, and extracting voxels to quantify length information such as diameters.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If 3D CT images are used to observe small sub-millimeter structures, then spatial resolution is improved, but automated classification and quantification of blood vessels becomes difficult due to complicated morphological structures
Solution Approach 1:
The patent segments the complex 3D vascular structure into a tree graph representation, where blood vessels are divided into hierarchical branches with defined parent-child relationships. This segmentation transforms the complicated morphological data into a structured format that can be systematically classified and quantified, resolving the contradiction between high spatial resolution and algorithmic complexity.
Solution Approach 2:
The patent introduces a minimum spanning tree (MST) as an intermediary computational structure to bridge the gap between raw 3D CT image data and the desired vascular classification. The MST algorithm processes the complex spatial relationships and generates a simplified tree structure that facilitates automated artery-vein classification, reducing the computational complexity while preserving essential morphological information.
2Ease of manufacture
If simple thresholding is applied to 2D sectional CT images to quantify pulmonary vessels, then processing is simplified, but arteries and veins cannot be distinguished from each other
Solution Approach 1:
The patent transitions from 2D sectional analysis to 3D volumetric analysis by constructing a minimum spanning tree from 3D CT data. This dimensional upgrade enables the algorithm to capture the spatial connectivity and topological relationships between vessels, providing the necessary information to distinguish arteries from veins while maintaining computational efficiency.
Solution Approach 2:
The patent changes the analytical parameters from simple intensity thresholding to a multi-parameter approach that includes spatial coordinates, connectivity relationships, and topological features in the minimum spanning tree structure. These parameter changes enable vessel type differentiation while keeping the processing framework relatively simple and automated.
3Ease of operation
If 2D slice images are used to measure blood vessel diameter, then measurement is simplified, but accurate diameter measurement in direction orthogonal to vessel axis becomes difficult
Solution Approach 1:
The patent uses 3D CT volumetric data instead of 2D slices, enabling measurements in three dimensions. This allows the algorithm to determine the vessel axis orientation and measure diameter in the direction orthogonal to the axis by analyzing cross-sectional properties in 3D space, significantly improving measurement accuracy while maintaining automated processing capability.
Solution Approach 2:
The patent performs preliminary actions by first constructing the minimum spanning tree and identifying vessel centers and orientations before conducting diameter measurements. This preliminary structuring of the data enables subsequent orthogonal diameter measurements to be performed accurately and automatically, resolving the contradiction between ease of operation and measurement precision.
Data Source
AI summary
A method for distinguishing between pulmonary arteries and pulmonary veins and a method for quantifying blood vessels are disclosed. The method for distinguishing between pulmonary arteries and pulmonary veins includes: forming a set of pulmonary vessels for points corresponding to pulmonary vessels, wherein each of the points of the set of pulmonary vessels has weight information; forming a tree from the points of the set of pulmonary vessels by using the weight information; and distinguishing between the pulmonary arteries and the pulmonary veins by separating the tree into a plurality of regions. The method for quantifying blood vessels includes: extracting blood vessels as a three-dimensional set of voxels based on medical images of an organ; finding the voxels of blood vessels included in a region of interest of the organ; and quantifying length information of the blood vessels by using the found voxels.