Deformable Image Registration for Organ Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing medical image segmentation methods fail to effectively distinguish target and background areas with the same or similar image intensities, which is a significant challenge in organ and anatomical structure segmentation.
Innovation Solution
The method employs image registration to generate a deformation field between nearby 2-D image slices, allowing for the distinction of target and background areas by mapping morphological structures from one slice to another, even when they have the same or similar image intensities, using deformable image registration procedures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If intensity-based region growing methods are used for segmentation, then the segmentation process is simple and fast, but the method fails to distinguish target and background areas with the same or similar image intensities
Solution Approach 1:
The patent transitions from 2-D intensity-based segmentation to 3-D deformable model-based segmentation. By incorporating spatial information from multiple image slices and using deformable models that evolve through 3-D space, the method distinguishes targets from backgrounds with similar intensities by considering morphological continuity across dimensions rather than relying solely on intensity thresholds.
Solution Approach 2:
The patent changes the segmentation parameters from simple intensity thresholds to complex deformable model parameters including boundary forces, internal forces, and external forces. The deformable model parameters such as elasticity, rigidity, and damping coefficients allow the segmentation to adapt to morphological variations while maintaining continuity across image slices, enabling distinction of similar-intensity regions.
2Measurement precision
If deformable image registration procedures are used to build deformation fields between image slices, then target and background areas with same or similar intensities can be distinguished, but the device complexity and computational requirements increase
Solution Approach 1:
The patent divides the 3-D image volume into multiple 2-D image slices and performs deformable registration between neighboring slices independently. This segmentation of the registration process reduces computational complexity compared to performing global 3-D registration, while still maintaining the ability to distinguish targets from backgrounds through local deformation fields.
Solution Approach 2:
The patent performs preliminary deformable registration between neighboring image slices to build deformation fields before conducting the final segmentation. This preliminary action of establishing spatial relationships and morphological correspondences between slices simplifies the subsequent segmentation process by pre-computing the deformation information needed for accurate target-background distinction.
Data Source
AI summary
A system and method to conduct image segmentation by imaging target morphological shapes evolving from one 2-dimension (2-D) image slice to one or more nearby neighboring 2-D images taken from a 3-dimension (3-D) image. One area defined by a user as a target on an image slice can be found in a corresponding area on a nearby neighboring image slice by using a deformation field generated with deformable image registration procedure between these two image slices. It allows the user to distinguish target and background areas with the same or similar image intensities.