Accurate identification and segmentation of choroidal neovascularization (CNV) are crucial for detecting and treating exudative age-related macular degeneration. (AMD). Cross-sectional and en-face imaging of CNV is possible with PR-OCTA or projection-resolved optical coherence tomographic i.e., OCT angiography.
Due to the persistence of residual artifacts, CNV segmentation and detection remain challenging even with PR-OCT Angiography. In this article, researchers present a convolutional neural network (CNN)-based algorithm for CNV diagnosis and segmentation.
This research used a clinical dataset that included scans of healthy and diseased eyes scanned with and without CNV. Additionally, no scans were disallowed because of poor image clarity. In testing, 95% specificity and 100% sensitivity allowed for diagnosing all CNV instances from non-CNV controls. The CNV membrane segmentation’s mean intersection over union peaked at 0.88. The suggested algorithm should be advantageous for CNV diagnosis and visualization monitoring because it enables automated categorization and segmentation.
A major factor in vision loss and permanent blindness is age-related macular degeneration (AMD). Due to choroidal neovascularization (CNV), a pathological situation in which new blood vessels develop from the choroid into the outer retina, AMD is classified as neovascular. Because CNV can cause subretinal hemorrhage, lipid exudation, subretinal fluid, intraretinal fluid, or the development of fibrotic lesions, it frequently causes vision loss.
Related Content: Low-Light Imaging In SWIR With Homodyne Detection