Vascular Image Segmentation for Overlapping Vessel Structure Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional vascular image processing methods struggle with accurately analyzing vascular tree structures, especially when blood vessels overlap, leading to incorrect identification of lesions and errors in syntax score application.
Innovation Solution
A vascular image processing method that extracts three distinct vascular regions from an image - a first region for the entire blood vessel, a second region for a target vessel and its branch vessels, and a third region for the target vessel - and predicts the vascular structure by classifying points into target, branch, and separate vessels, with error correction using additional images and lesion information.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If binary segmentation into vascular and non-vascular regions is used, then the processing is simple, but the connectivity of overlapping vascular regions becomes unclear and vascular tree structure identification becomes difficult
Solution Approach 1:
The patent divides the vascular region into multiple hierarchical levels: first vascular region (entire blood vessel), second vascular region (target vessel and branch vessels), and third vascular region (target vessel only). This multi-level segmentation resolves the connectivity ambiguity of overlapping vessels while maintaining structured analysis capability.
2Measurement precision
If deep learning-based blood vessel segmentation is used, then segmentation performance is improved, but the accuracy of vascular tree structure analysis deteriorates when vessels overlap
Solution Approach 1:
The patent applies multi-level segmentation to separate overlapping vessels into distinct hierarchical regions, enabling accurate vascular tree structure analysis while preserving the segmentation performance benefits of deep learning methods.
Solution Approach 2:
The patent assigns different properties and analysis methods to different vascular regions based on their hierarchical level and connectivity characteristics, allowing tailored processing that maintains both segmentation accuracy and structural analysis reliability.
3Ease of operation
If vascular regions are not separately classified, then the analysis process is simplified, but lesion identification accuracy and syntax score application deteriorate
Solution Approach 1:
The patent classifies vascular regions into three distinct types with specific processing rules for each, enabling accurate lesion identification and syntax score application while maintaining a systematic and manageable analysis framework.
Data Source
AI summary
A vascular image processing method performed by a processor includes extracting, from a vascular image, a first vascular region corresponding to an entire blood vessel included in the vascular image, a second vascular region corresponding to a target vessel and one or more branch vessels connected to the target vessel, and a third vascular region corresponding to the target vessel, and predicting a vascular structure in the vascular image, based on the first vascular region, the second vascular region, and the third vascular region, which are extracted from the vascular image.