Optical coherence tomography (OCT) is a non-invasive imaging technology that can be used to obtain a cross-sectional view of the retina. OCT-scans are typically acquired as a stack of linear slices, providing a high-resolution three dimensional image of the interior layered structure of the retina. In various retinal diseases, changes in thickness of these retinal layers occur. Accurate quantification of these changes helps in the diagnosis, prognosis and severity grading of several eye complications, such as glaucoma, diabetic retinopathy or age-related macular degeneration.
Manual segmentation of these retinal layers is a time-consuming and potentially subjective task. Therefore, an automated system that accurately performs this task is needed. Although there are currently some methods available for automatic segmentation of retinal layers, these methods often fail in the presence of severe pathology. Deep learning has revolutionized image processing in general, and is now also making its way into medical imaging. With the help of deep learning, we could develop and automated method for retinal layer segmentation that works well even for heavily affected retinas.
The goal of this project is to develop an algorithm for the detection and segmentation of retinal layers in OCT-scans of severely affected retinas.
A large database of retinal images with different eye disease manifestations is available to develop and evaluate the system.
The work will be executed in the Diagnostic Image Analysis Group (DIAG) of the Radboud University. DIAG is a leading research group in computer-aided detection and diagnosis (CAD).