Neural Network OCT Volume Segmentation for Choroidal Neovascularization Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying choroidal neovascularization (CNV) in the human eye, such as fluorescein angiography, are invasive and carry risks of adverse reactions, while non-invasive techniques like optical coherence tomography (OCT) lack the specificity and sensitivity to accurately differentiate between classic and occult CNV types.
Innovation Solution
A computer-implemented method using optical coherence tomography (OCT) data and a neural network to generate volume segments of the eye, which allows for the identification of classic and occult choroidal neovascularization by detecting tissue layer and fluidic segments, and weighting specific fluidic segments for accurate classification.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If fluorescein angiography is used to identify choroidal neovascularization type, then diagnostic accuracy is improved, but patient safety deteriorates due to invasive procedures and adverse reactions
Solution Approach 1:
The patent replaces the invasive mechanical/chemical system of fluorescein injection and angiography with a non-invasive optical imaging system using OCT technology. The neural network analyzes optical coherence tomography images to classify CNV types, eliminating the need for contrast dye injection and associated adverse reactions while maintaining diagnostic accuracy through advanced image processing and AI analysis
Solution Approach 2:
The patent introduces a neural network as an intermediary between the OCT imaging system and the diagnostic decision-making process. This AI intermediary processes the optical images, extracts relevant features, and classifies CNV types, serving as a bridge that translates non-invasive optical data into accurate diagnostic information without requiring invasive procedures
2Object-affected harmful factors
If optical coherence tomography is used for choroidal neovascularization identification, then patient safety is improved by eliminating invasive procedures, but measurement precision deteriorates due to inability to accurately differentiate CNV types
Solution Approach 1:
The patent applies segmentation by dividing the complex task of CNV classification into distinct processing stages: OCT image acquisition, pre-processing, feature extraction, neural network analysis, and classification. The neural network itself segments the image data into relevant features and patterns that distinguish classic from occult CNV, enabling accurate differentiation without invasive procedures
Solution Approach 2:
The patent transforms the OCT imaging parameters and data representation to enhance diagnostic capability. By adjusting imaging depths, processing multiple B-scans, and transforming image data into features suitable for neural network analysis, the system extracts sufficient information from non-invasive OCT images to accurately classify CNV types, overcoming the traditional limitation of OCT precision
3Ease of operation
If traditional OCT imaging is used to evaluate choroidal neovascularization, then non-invasive imaging is achieved, but information completeness deteriorates due to insufficient detail for accurate CNV type differentiation
Solution Approach 1:
The patent enhances information completeness by transitioning from standard two-dimensional OCT B-scans to three-dimensional volumetric analysis. By processing multiple B-scans and utilizing depth information across multiple layers, the neural network accesses additional dimensional data that reveals subtle structural details of CNV lesions, enabling accurate type differentiation while maintaining non-invasive imaging
Applied Scientific Principles
This section explains which scientific principles are used to turn an abstract innovation direction into a practical engineering solution.
Function Achieved in This Case
The method provides a non-invasive and accurate means to identify the type of choroidal neovascularization, improving specificity and sensitivity over existing techniques, and reducing the risks associated with invasive procedures.
Implementation Method 1
Optical coherence tomography (OCT) offers a method for the creating an optical cross-section of the eye which is non-invasive and high-resolution
Data Source
AI summary
A method for identifying a type of choroidal neovascularization in a retina of a human eye is disclosed, the method comprising receiving, in a processor, optical coherence tomography data of the eye; generating, in the processor, volume segments of the eye using the optical coherence tomography data and a neural network; and identifying, in the processor, the type of choroidal neovascularization in the eye, using the volume segments, the type of choroidal neovascularization comprising one or more of the following: classic choroidal neovascularization and occult choroidal neovascularization.