Vascular Tree Graph Registration for Accurate CTA Image Comparison
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing automated landmark-based registration methods for comparing CT angiography images, particularly in multiphase CTA, are prone to errors and inaccuracies, leading to false vessel route representations that hinder effective diagnosis.
Innovation Solution
A computer-implemented method using directed acyclical graphs of vascular trees to register and compare 3D images, establishing correspondences based on spatial proximity and link structure, enabling automatic matching and visualization of vascular changes over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If automated landmark-based registration methods are used to compare CT angiography images, then the comparison process can be automated, but the accuracy and reliability of the registration deteriorates due to false routes and inaccuracies
Solution Approach 1:
The patent segments the vascular tree into multiple levels of hierarchical structures (e.g., main vessels, branches, sub-branches) and processes correspondences level by level. This segmentation allows the system to handle complex vascular structures systematically, avoiding false routes by processing simpler segments before complex ones, thereby maintaining reliability while achieving automation.
Solution Approach 2:
The patent introduces a hierarchical dimension to the registration process, organizing correspondences into multiple levels rather than treating all points equally. By adding this structural dimension and using it to guide the correspondence establishment, the system achieves more reliable automated registration that avoids the pitfalls of traditional landmark-based methods.
2Ease of manufacture
If traditional landmark-based registration is used, then the process is simple to implement, but the measurement precision of vascular correspondences deteriorates due to false routes
Solution Approach 1:
The patent divides the vascular tree into hierarchical segments and processes correspondences segment by segment. This segmentation maintains implementation simplicity while significantly improving measurement precision by avoiding false routes through structured processing of vascular correspondences at multiple levels.
Solution Approach 2:
The patent introduces hierarchical intermediate structures that mediate between the simple input images and the precise correspondence output. These intermediate hierarchical levels act as a bridge, allowing the system to maintain ease of implementation while achieving high measurement precision through structured intermediate processing steps.
3Reliability
If graph-based registration with hierarchical structures is used, then the reliability and precision of correspondence improve, but the device complexity increases
Solution Approach 1:
The patent segments the complex graph processing into hierarchical levels, where simpler processing steps are applied at lower levels and progressively more complex processing occurs at higher levels. This segmentation reduces the perceived complexity by breaking down the overall system into manageable, structured components while maintaining high reliability.
Solution Approach 2:
The patent performs preliminary processing at lower hierarchical levels before proceeding to more complex processing at higher levels. By completing simpler correspondence establishment tasks first, the system prepares the data structure in advance, reducing the complexity burden on subsequent processing steps while ensuring reliable correspondences.
Data Source
AI summary
A method for comparing images, comprises: receiving images with the same subject matter that have been recorded at different times; establishing structures in the images and generating directed acyclical graphs based on the structures in the images, wherein each graph has specified points; registering a graph of at least one second image to the graph of a first image; establishing a correspondence between the points of the registered graphs, based on the spatial proximity of the points in conjunction with a link structure of the graphs; registering at least regions of the images that are specified by corresponding points, according to the registered graphs; and outputting at least the registered regions of the images.