WP 1.2: Predicting the optical and visual performance of intraocular lenses

The research focus is on developing computational methods for objective assessment of ocular optical quality, with direct applications to intraocular lens performance prediction. A deep learning approach has been developed combining physics-based simulation with machine learning. The methodology can retrieve wavefront aberrations from point spread function (PSF) images, even in the presence of straylight - a critical capability for evaluating real-world optical performance. The model demonstrates promising accuracy in predicting optical aberrations across varying straylight conditions, with fast processing times that could enable real-time applications. This represents a potential improvement over traditional methods, which are slower and prone to convergence issues.

The developed model directly supports the WP 1.2 objectives by providing a rapid, objective tool for assessing optical quality that could be integrated into intraocular lens evaluation workflows. The ability to separate aberration and scatter contributions from standard PSF measurements offers clinicians a more comprehensive understanding of how different lens designs will perform in real patients, particularly older adults where straylight becomes more prevalent. Currently, the approach has been developed and validated using simulated data (Ntatsis, K., Christaras, D., Artal, P. and Ginis, H., 2025. Deep learning ocular aberration retrieval from simulated retinal images under straylight. Biomedical Optics Express, 16(7), pp.2709-2718).