CNN-Based CNV Membrane and Vasculature Segmentation in OCTA
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for identifying and monitoring choroidal neovascularization (CNV) using optical coherence tomographic angiography (OCTA) face challenges such as two-dimensional visualization limitations, invasive dye-based angiography drawbacks, and susceptibility to imaging artifacts, leading to inaccurate diagnosis and time-consuming clinical evaluation.
Innovation Solution
The development of systems and methods employing Convolutional Neural Networks (CNNs) for automated segmentation and identification of CNV membranes and vasculature from OCTA images, capable of generating membrane and vasculature masks to assist clinicians in accurate diagnosis and monitoring, even in low-quality scans with artifacts.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If dye-based angiography techniques are used for CNV identification, then diagnostic accuracy is improved, but patient safety deteriorates due to invasive procedures
Solution Approach 1:
The patent replaces invasive mechanical/dye-based angiography systems with non-invasive optical coherence tomographic angiography (OCTA) systems. OCTA uses optical interference patterns to visualize blood flow and CNV membranes without requiring contrast dyes, thereby maintaining diagnostic accuracy while eliminating the harmful effects of invasive procedures on patients
Solution Approach 2:
The patent introduces artificial intelligence algorithms as an intermediary between OCTA imaging and clinical diagnosis. These AI algorithms automatically segment and identify CNV membranes and vasculature from OCTA images, serving as a mediator that enhances diagnostic precision while reducing reliance on invasive traditional angiography methods
2Object-affected harmful factors
If traditional OCTA methods are used for CNV visualization, then noninvasive imaging is achieved, but diagnostic precision deteriorates due to projection artifacts
Solution Approach 1:
The patent extracts and removes projection artifacts from OCTA images through AI-based image processing. The system specifically identifies and eliminates artifacts that obscure CNV membranes and vasculature, thereby improving diagnostic precision while maintaining the noninvasive nature of OCTA imaging
Solution Approach 2:
The patent uses AI algorithms to create enhanced copies of OCTA images with artifacts removed. The system generates refined image representations that accurately depict CNV structures without the distortions present in original images, improving diagnostic precision without additional patient exposure
3Reliability
If manual evaluation of OCTA images is performed, then clinical assessment is conducted, but time consumption increases
Solution Approach 1:
The patent implements self-service through automated AI algorithms that independently perform CNV membrane and vasculature segmentation from OCTA images. The system automatically identifies, segments, and quantifies CNV features without requiring manual clinical evaluation, thereby maintaining assessment quality while dramatically reducing time consumption
Solution Approach 2:
The patent performs preliminary automated segmentation and identification of CNV structures before final clinical diagnosis. The AI system pre-processes OCTA images to highlight key features and generate initial assessments, allowing clinicians to review pre-analyzed results rather than performing manual evaluation from scratch
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
Enables accurate and efficient identification and segmentation of CNV, facilitating early detection and management of neovascular age-related macular degeneration, reducing reliance on invasive techniques and improving diagnostic precision.
Implementation Method 1
Optical coherence tomographic angiography (OCTA), can measure flow signal in vivo by evaluating motion contrast between subsequent OCT B-scans at the same location
Data Source
AI summary
Methods and systems for identifying CNV membranes and vasculature in images obtained using noninvasive imaging techniques are described. An example method includes generating, by a first model based on at least one image of a retina of a subject, a membrane mask indicating a location of a CNV membrane in the retina. The method further includes generating, by a second model based on the membrane mask and the at least one image, a vasculature mask of the retina of the subject, the vasculature mask indicating CNV vascularization in the retina.