Enabling sensitive retinal biomarkers using adaptive optics imaging, simulation and AI

About the Project

Retinal diseases profoundly impact sight and quality of life. As these conditions are currently incurable and progressive, early detection is essential for preserving remaining vision. Measuring cellular-level changes offers a path toward earlier diagnosis, a better understanding of disease progression, and closer monitoring of existing and emerging treatments

Cellular resolution imaging is not yet standard in clinics, but research groups are increasingly using adaptive optics to visualise individual light-sensitive photoreceptors—rods and cones—in the living eye. Evidence suggests that precise measures of photoreceptor density can serve as early biomarkers, revealing pathological changes before they are visible through traditional clinical methods.

Reliable and automated methods for detecting cells in retinal images are needed, but current algorithms have only been developed and tested using limited datasets of images of healthy retinae. Manual correction is often needed and this is time-consuming and subject to human error. To address this, we developed ERICA, a model that generates realistic synthetic images of cone cells, as seen with adaptive optics. This project will advance ERICA to simulate biologically realistic rod mosaics and overlay vascular networks. These advanced synthetic datasets will drive new machine learning approaches for the simultaneous detection and segmentation of cones, rods, and blood vessels. The project will contribute to a generalised automated analysis tool for adaptive optics retinal imaging—accelerating discovery and translation in ophthalmic research.

This project is ideal for students with backgrounds in computer science, physics, engineering, with a passion for applying their skills in the life sciences and clinical research. It would also appeal to students with a life sciences background who have a strong interest in developing technical skills in programming and machine learning. You will gain hands-on experience in deep learning, synthetic data generation, and and ethical and explainable AI principles in medical imaging. There will also be opportunities to work with our imaging system to collect or use data from volunteers, and to collaborate with members of our team who work in instrument development and in clinical research across the university and local NHS Trust.

Funding

Students who have, or are expecting to attain, at least an upper second-class honours degree (or equivalent) in a relevant subject, are invited to apply. Funding is available for Home (UK) students to cover tuition fees, a tax-free stipend at the UKRI rate (indicative amount in year 1 in 2026-27, £21,805) and research costs, for four years. Applicants normally required to cover International fees will have to cover the difference between the Home and the International tuition fee rates. There is no additional funding available to cover NHS Immigration Health Surcharge (IHS) costs, visa costs, flights etc.

Funding for this studentship is awarded on a competitive basis and is not guaranteed; availability will depend on the outcome of the selection process and subject to final approval by the University.