Maryam (Pouneh) Dinpajhouh





OCT in advanced glaucoma: Analysis of retinal tissue properties using Deep Learning

In October 2023, I joined the Laboratory of Experimental Ophthalmology (LEO) and the Machine Learning Laboratory (MLL) at the University Medical Center Groningen (UMCG). During my three-year PhD journey, I am supervised by Prof. Nomdo M. Jansonius and Prof. Peter van Ooijen at UMCG. My research also includes two secondments: one at the Otto von Guericke University (OVGU) under the supervision of Prof. Michael Hoffmann, and another at Heidelberg Engineering, as a cross-sector collaboration.

Click here to read more about my project!

Personal Background and Interest

I earned my Bachelor’s and Master’s degrees in Biomedical Engineering (Bioelectric branch) from Amirkabir University of Technology (Tehran Polytechnic). My research interests focus on applying Artificial Intelligence and Image Analysis to address complex challenges in the medical field, particularly in the assessment of retinal disorders and the prediction and detection of their progression over time.

Aim of the project

The primary objective of this project is to assess advanced glaucomatous retinal tissue and evaluate the utility of Optical Coherence Tomography (OCT) and OCT Angiography (OCTA) in detecting precise structural and microvascular changes associated with advanced glaucoma.
Accordingly, this project focuses on:

• Detecting structural changes in advanced glaucomatous retinal tissue using optic disc OCTs.

• Detecting structural changes in advanced glaucomatous retinal tissue using macular OCTs.

• Detecting microvascular changes in advanced glaucomatous retinal tissue using OCTA.

Current activities/Accomplishments

To date, I have:

• Extracted volumetric peripapillary (optic disc) and macular OCT scans from the Canon OCT Imaging Database available at the Ophthalmology Department of the UMCG.

• Prepared macular and optic disc datasets encompassing a wide range of disease severities, from mild to advanced stages.
(These datasets cannot be publicly released due to institutional data policy regulations.)

• Implemented a Generative Adversarial Network (GAN)-based model to reduce speckle noise in OCT images (work in progress).

• Begun drafting a research paper on the OCT denoising study.

• Developed a deep learning model for retinal layer assessment in advanced glaucoma, focusing specifically on peripapillary OCTs (work in progress).

• Started drafting a research paper on the optic disc analysis.

My further studies will concentrate on achieving the objectives outlined in the aims of the project section.