Segmented Volume Leakage Removal via Morphological Erosion and Dilation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Higher resolution diagnostic imaging systems generate noisy data, leading to inaccurate segmentation of structures like airways and vessels due to inclusion of leakage regions in the segmented volume, as traditional algorithms struggle to differentiate between accurate data and noise.
Innovation Solution
A method involving the generation of a reduced-resolution segmentation mask, followed by morphological erosion and dilation to identify and remove leakage regions, resulting in an updated segmented volume that accurately represents the anatomical structure.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional segmentation algorithms are applied to high-resolution diagnostic imaging data, then the segmented volume includes detailed anatomical structures, but the segmentation accuracy deteriorates due to noise-induced leakage regions
Solution Approach 1:
The patent applies morphological segmentation by dividing the segmentation correction process into distinct stages: erosion to separate leakage regions, dilation to restore boundaries, and selective removal of identified leakage areas. This multi-stage morphological segmentation enables precise differentiation between true anatomical structures and noise-induced leakage regions while maintaining detailed structural information from high-resolution data
Solution Approach 2:
The patent introduces an intermediary processing step using morphological operations (erosion and dilation) as a mediator between the raw segmentation result and the final corrected segmentation. These intermediary operations temporarily modify the segmentation mask to reveal leakage regions without permanently altering the underlying anatomical structure representation, enabling accurate identification and removal of errors
2Measurement precision
If morphological erosion is applied to the segmented region, then leakage regions can be identified, but the segmented volume temporarily loses boundary detail
Solution Approach 1:
The patent performs morphological erosion as a preliminary action before final segmentation output to temporarily shrink the segmented region and expose leakage regions. This preliminary modification allows identification of erroneous inclusions without permanently altering the boundary detail, as the erosion is followed by dilation to restore the original boundaries while removing identified leakage areas
Data Source
AI summary
A method of modifying a segmented volume is disclosed herein. The method includes generating a reduced-resolution segmentation mask including a segmented region. The method includes performing a morphological erosion on the segmented region to form an eroded region and performing a morphological dilation on the eroded region to form a dilated region. The method also included identifying a leakage region in the segmented volume based on the dilated region and removing the leakage region from the segmented region to form an updated segmented volume.