Aneurysm Nidus Segmentation Using Vessel Centerline Image Sections
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for intracranial aneurysm segmentation in medical images are time-consuming, labor-intensive, and prone to inaccurate measurements due to the small size and varying shapes of aneurysms, which are often confused with parent arteries, leading to low accuracy in segmentation.
Innovation Solution
A method involving vessel segmentation, centerline extraction, and image cutting to obtain object medical and vessel segmentation results, followed by inputting these results into a nidus segmentation model for accurate segmentation, utilizing trained models like 3D convolutional neural networks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual delineation of aneurysm nidus region is performed, then diagnostic accuracy can be maintained through doctor expertise, but time consumption and labor intensity increase significantly
Solution Approach 1:
The system performs preliminary vessel segmentation and centerline extraction to prepare processed data before the actual nidus segmentation. This preliminary processing organizes the medical image data and vessel information in advance, enabling the deep learning model to perform accurate segmentation more efficiently without requiring manual intervention for each step.
Solution Approach 2:
The patent introduces an intermediary processing system that includes centerline extraction and image cutting modules. These intermediary components process the raw medical images and vessel segmentation results into formatted inputs suitable for the deep learning model, bridging the gap between raw data and final segmentation output while automating the workflow.
2Productivity
If deep learning methods are applied for automated nidus segmentation, then time efficiency and productivity improve, but segmentation accuracy deteriorates due to small size and shape variability of aneurysms
Solution Approach 1:
The patent divides the medical image processing into multiple segments: vessel segmentation, centerline extraction, image cutting into sections, and final nidus segmentation. This multi-stage segmentation approach allows each module to focus on specific tasks, improving overall accuracy while maintaining automated efficiency. The deep learning model processes segmented sections rather than entire images, enhancing precision for small aneurysms.
Solution Approach 2:
The system applies local quality processing by extracting centerlines and cutting images into sections based on vessel geometry. This allows the deep learning model to focus computational resources on local regions containing potential aneurysms rather than processing entire images uniformly. The local processing enhances detection accuracy for small, variable-shaped aneurysms while maintaining high productivity.
3Extent of automation
If standard deep learning models are used for nidus segmentation, then automation extent increases, but measurement precision decreases due to confusion between aneurysms and parent arteries
Solution Approach 1:
The patent extracts the vessel centerline and surrounding sections from the full medical image, isolating the region of interest containing the potential aneurysm. This extraction removes irrelevant background information and focuses the deep learning model on the critical area where aneurysms occur, reducing confusion with parent arteries and improving segmentation accuracy while maintaining full automation.
Solution Approach 2:
The system transitions from processing entire 3D medical volumes to processing 2D cross-sectional slices centered on extracted vessel centerlines. This dimensional change allows the deep learning model to analyze thin sections with higher resolution and focus on local anatomical features, improving the distinction between aneurysms and parent arteries while maintaining automated processing.
4Measurement precision
If manual segmentation is performed to ensure accuracy, then measurement precision is maintained, but device complexity increases due to need for expert intervention
Solution Approach 1:
The system performs self-service through automated vessel segmentation, centerline extraction, and nidus segmentation without requiring manual expert intervention. The deep learning models automatically process medical images, extract relevant features, and generate segmentation results, eliminating the need for doctor involvement in the segmentation process while maintaining high accuracy through algorithmic precision.
Data Source
AI summary
A method of nidus segmentation includes: acquiring a medical image to be processed; performing a vessel segmentation on the medical image to be processed to obtain a first vessel segmentation result; extracting a first centerline of a first vessel according to the first vessel segmentation result; cutting the medical image to be processed for several times to obtain a plurality of first sections and further obtain a first object medical image by synthesizing the plurality of sections, and cutting the first vessel segmentation result for several times to obtain a plurality of second sections and further obtain a first object vessel segmentation result by synthesizing the plurality of second sections; and obtaining a nidus segmentation result based on the first object medical image and the first object vessel segmentation result.