Machine-Learning Vascular Tree Segmentation for Low-Contrast Angiography
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current vascular segmentation and feature identification in angiographic images are prone to errors due to low contrast and complex environments, leading to inaccurate three-dimensional models of the heart.
Innovation Solution
A system utilizing machine learning models, such as convolutional neural networks, to select an optimal angiographic image from a sequence based on contrast and quality scores, combined with classical computer vision techniques to enhance image quality and reduce manual adjustments.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If automated analysis is used for vascular segmentation and feature identification, then time and effort are reduced, but accuracy decreases due to low contrast and complex environments
Solution Approach 1:
The patent introduces an intermediary system that combines automated image processing with machine learning-based classification. The automated analysis first performs initial segmentation, and then a trained classifier (intermediary) evaluates the segmented features to correct errors and improve accuracy, resolving the contradiction between automation speed and precision
Solution Approach 2:
The system implements feedback mechanisms where the classification results are used to refine and re-evaluate the automated segmentation. The classifier provides feedback on which automated results are reliable and which need manual review, creating a closed-loop system that improves accuracy while maintaining automation benefits
2Measurement precision
If manual confirmation and correction of vascular tree is performed, then accuracy is improved, but time and skill requirements increase
Solution Approach 1:
Instead of requiring complete manual review of all vascular features, the system applies partial manual intervention only where needed. The classifier identifies specific regions or features that require human expertise, allowing operators to focus their time and skill on problematic areas rather than performing exhaustive manual correction
Solution Approach 2:
The system performs self-correction through the automated classification process. The trained classifier automatically identifies and corrects many errors in the automated segmentation without human intervention, making the system self-sufficient for routine corrections and reducing the time burden on operators
Data Source
AI summary
Methods for automated identification of vascular features are described. In some embodiments, one or more machine learning (ML)-based vascular classifiers are used, with their results being combined to with results of at least one other vascular classifier in order to produce the final results. Potentially advantages of this approach include the ability to combine certain strengths of ML classifiers with segmentation approaches based on more classical (“formula-based”) methods. These strengths may include particularly the identification of anatomically identified targets mixed within an image also showing similar looking but anatomically distinct targets.