Sub-resolution Vessel Lumen Segmentation via Deconvolution
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current medical imaging techniques, such as cardiac computed tomography angiography, struggle to accurately detect and quantify sub-resolution narrowing of coronary artery lumens due to limitations in visible resolution and artifacts caused by calcium deposits, leading to overestimation of lumen diameter and blooming artifacts.
Innovation Solution
A system and method utilizing a graph min-cut algorithm for sub-resolution luminal narrowing detection and segmentation, which includes a sub-resolution luminal narrowing detector and a graph min-cut variation segmenter to improve the precision of lumen segmentation by analyzing centerline intensity profiles and voxel-wise likelihoods, and applying a smoothness penalty for optimization.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If visible imaging resolution is used (CCTA with ~1.5mm resolution), then imaging is non-invasive and feasible, but sub-resolution narrowing detection accuracy deteriorates
Solution Approach 1:
The patent applies preliminary action by performing deconvolution processing on the blurred CT images before segmentation. The deconvolution step removes the low-pass filter blur effects in advance, allowing subsequent segmentation algorithms to work with sharpened images that better represent the true lumen boundaries, thereby enabling sub-resolution narrowing detection while maintaining non-invasive imaging
Solution Approach 2:
The patent replaces traditional visible-resolution-based segmentation methods with a computational approach using deconvolution algorithms and enhanced segmentation techniques. This substitution of mechanical/imaging resolution limitations with computational processing enables precision beyond the physical resolution limits of the CT scanner
2Stability of the object's composition
If low-pass filter smoothing is applied during reconstruction, then image noise is reduced, but lumen boundary accuracy deteriorates due to blurring
Solution Approach 1:
The patent applies preliminary anti-action by using deconvolution to counteract the blurring effect of the low-pass filter. The deconvolution process reverses the smoothing operation, sharpening the lumen boundaries before segmentation occurs. This allows the system to benefit from noise reduction while compensating for the boundary blurring through computational correction
Solution Approach 2:
The patent introduces deconvolution processing as an intermediary step between image reconstruction and segmentation. This intermediary process acts as a bridge that reverses the blurring effects and restores boundary sharpness, allowing the segmentation algorithm to achieve precise lumen boundary definition even though the original reconstructed images were smoothed
3Device complexity
If typical segmentation algorithms are used on blurred images, then segmentation is computationally simple, but lumen diameter is overestimated due to blooming artifacts
Solution Approach 1:
The patent applies preliminary action by performing deconvolution to sharpen images before segmentation. This pre-processing step removes the blooming artifacts and boundary blurring in advance, allowing subsequent segmentation algorithms to accurately identify true lumen boundaries without overestimation, thereby maintaining measurement precision while keeping the segmentation approach practical
Data Source
Figure 1
Figure 2
Figure 3
AI summary
An imaging system (100) includes a sub-resolution luminal narrowing detector (112) which detects sub-resolution narrowing of a vessel lumen in an image volume by a centerline profile analysis and computes a sub-resolution determined diameter by modifying an approximated visible lumen diameter with the detected sub-resolution narrowing.