Cardiovascular Image Patch Segmentation for Lesion Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for processing cardiovascular images struggle to accurately detect cardiovascular lesions, particularly in detailing blood vessels and calculating stenosis rates with high accuracy.
Innovation Solution
A method involving the extraction of image patches from cardiovascular images and centerline image masks, followed by local refinement using a trained model to generate refined image masks and improve blood vessel contour accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the entire cardiovascular image is used as input data to detect main blood vessels, then the performance in distinguishing main blood vessels is excellent, but the ability to process blood vessels in more detail is insufficient
Solution Approach 1:
The patent segments the cardiovascular image into multiple image patches of predetermined sizes, processing each patch independently through the neural network model. This segmentation allows the system to maintain the ability to detect main blood vessels across the entire image while simultaneously processing detailed local structures within each patch, resolving the contradiction between overall detection accuracy and local detail processing capability.
Solution Approach 2:
The patent introduces a multi-scale processing dimension by extracting image patches of different sizes and feeding them into the neural network model. This dimensional approach enables the system to analyze both the global structure of blood vessels across the entire image and the local detailed structures within each patch, achieving both excellent main blood vessel distinction and detailed blood vessel processing.
2Manufacturing precision
If image patches are extracted and processed through local refinement, then the segmentation of lesion candidate regions is improved, but the computational complexity increases
Solution Approach 1:
The patent applies local refinement only to specific image patches that contain lesion candidate regions, rather than processing the entire image uniformly. By identifying and focusing computational resources on local areas with potential lesions, the system improves segmentation accuracy for these critical regions while reducing overall computational complexity compared to uniform processing of the entire image.
Solution Approach 2:
The patent performs preliminary actions by extracting image patches and identifying lesion candidate regions before applying the full local refinement process. This preliminary segmentation and classification allows the system to prepare data in advance and apply computational resources more efficiently, improving final segmentation accuracy while managing computational complexity through staged processing.
Data Source
AI summary
A method for processing a cardiovascular image for detection of cardiovascular lesions is provided, which is performed by one or more processors of a computing device. The method includes receiving a cardiovascular image, acquiring a first image mask corresponding to at least a part of blood vessels included in the cardiovascular image, acquiring a centerline image mask corresponding to centerlines of at least the part of blood vessels included in the cardiovascular image, extracting a first image patch from the cardiovascular image, extracting a second image patch from the centerline image mask, generating a refined third image patch by performing, based on the first image patch and the second image patch, a local refinement, and generating, based on the refined third image patch and the first image mask, a refined second image mask.