Layer Boundary Evolution for Macular OCT Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current automatic segmentation techniques for identifying layer boundaries in macular optical coherence tomography (OCT) images lack sub-voxel precision, leading to inaccurate test results and inefficient resource utilization.
Innovation Solution
The implementation of a layer boundary evolution (LBE) technique using a data model trained with machine learning to process 2D images, generating probability maps and directional vectors to refine initial boundary positions, achieving voxel or sub-voxel precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If current automatic segmentation techniques are used to identify layer boundaries in macular OCT images, then the segmentation process can be automated, but the precision of boundary identification fails to achieve sub-voxel accuracy
Solution Approach 1:
The patent replaces traditional mechanical/image processing segmentation methods with a physics-inspired evolution system. Boundary identification is achieved through simulated physical processes (energy minimization, gradient descent) rather than conventional image thresholding or edge detection algorithms, enabling sub-voxel precision while maintaining full automation.
Solution Approach 2:
The patent transforms the segmentation problem by changing parameters from discrete pixel/voxel values to continuous probability distributions. By representing boundaries as probabilistic surfaces that evolve through energy minimization, the system achieves sub-voxel precision by operating in a continuous parameter space rather than discrete image grid space.
2Measurement precision
If higher precision boundary identification is achieved through more sophisticated segmentation algorithms, then measurement accuracy improves, but computational time and resource utilization increase
Solution Approach 1:
The patent extracts and separates the boundary identification problem from the entire image processing task. By focusing computational resources only on determining boundary locations through evolution of probability surfaces rather than processing the entire image volume, the system achieves high precision with reduced computational time compared to full-volume sophisticated segmentation algorithms.
Solution Approach 2:
The patent performs preliminary actions by pre-computing probability maps and energy functions before actual boundary evolution. This preprocessing step prepares the computational landscape in advance, allowing the evolution process to converge faster to precise boundaries without requiring iterative refinement of the entire image data during the segmentation phase.
Data Source
AI summary
A device receives a two-dimensional (2-D) image that depicts a cross-sectional view of a retina that includes a macula comprised of layers and boundaries used to segment the layers. The device converts the 2-D image to a standardized format, determines features for voxels included in the 2-D image, and generates, by using a data model to process the features, probability maps that indicate likelihoods of the voxels being in positions within particular boundaries. The device analyzes the probability maps to determine an initial set of boundary positions and to generate directional vectors that point in directions based on values included in the set of probability maps, determines a final set of boundary positions by performing a layer boundary evolution technique using the directional vectors to refine the initial set of boundary positions, and provides data that identifies the final set of boundary positions for display via an interface.