3D Retinal Disruption Detection via Volumetric OCT Analysis
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for detecting and measuring retinal disruptions using 3D OCT data are error-prone, fail to fully utilize 3D data, and often miss disruptions below referenced layers, due to dependence on 2D surface segmentations and assumptions of constant elevations, leading to inaccurate and incomplete results.
Innovation Solution
A system and method that uses 3D region growing techniques with constraints like shape, size, and texture to detect retinal disruptions, generating 3D seeds and performing image processing to obtain characteristics and measurements, while providing an interactive GUI for error correction and displaying results, enabling detection and measurement of disruptions above and below referenced layers.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If 2D surface segmentation is used to detect retinal disruptions, then the detection process is simplified, but the accuracy and reliability of disruption detection deteriorates
Solution Approach 1:
The patent transitions from 2D surface segmentation to 3D volumetric analysis of retinal disruptions. By utilizing the third dimension (depth/z-axis) in OCT data, the system performs region growing and texture analysis in 3D space, enabling more accurate detection of disruptions while maintaining clinical workflow efficiency.
2Object-affected harmful factors
If smoothing operations are applied to reduce noise in OCT data, then noise reduction is improved, but the loss of fine detail information in disruptions increases
Solution Approach 1:
The patent applies local quality analysis by examining texture and intensity variations within local 3D regions rather than applying global smoothing. The region growing algorithm operates on locally defined seed points and evaluates local texture characteristics, preserving fine details while reducing noise through localized analysis.
Solution Approach 2:
The system creates multiple representations of the OCT data including intensity images, texture images, and segmented layers. By working with copied and processed versions of the original data through different filtering and analysis methods, the system can reduce noise in certain representations while preserving details in others.
3Ease of manufacture
If constant elevation assumptions are made for segmented layers, then the detection method is easier to implement, but the clinical meaningfulness and accuracy deteriorates
Solution Approach 1:
The patent replaces static constant elevation assumptions with dynamic, data-driven 3D region characteristics. The system allows disruptions to have variable shapes, sizes, and orientations by performing 3D region growing without rigid geometric constraints, enabling the detection of realistically varied disruption morphologies that reflect actual clinical pathology.
4Device complexity
If only disruptions above referenced layers are detected, then the detection algorithm is simpler, but the completeness of disruption detection deteriorates
Solution Approach 1:
The patent creates a universal 3D disruption detection framework that can identify disruptions regardless of their position relative to segmented retinal layers. The region growing algorithm operates independently of layer boundaries, enabling detection of disruptions above, below, or involving multiple layers, making the system universally applicable to various disruption types and locations.
Data Source
AI summary
System and method for 3D retinal disruption/elevation detection, measurement and presentation using Optical Coherence Tomography (OCT) are provided. The present invention is capable of detecting and measuring the abnormal changes of retinal layers (retinal disruptions), caused by retinal diseases, such as hard drusen, soft drusen, Pigment Epithelium Detachment (PED), Choroidal Neovascularization (CNV), Geographic Atrophy (GA), intra retinal fluid space, and exudates etc. The presentations of the results are provided with quantitative measurements of disruptions in retina and can be used for diagnosis and treatment of retinal diseases.