AI at the forefront of the eye (in person)

In 2020, this all-female team received a Seed Fund Award from the Melbourne Centre for Data Science to build a digital platform for robust, automated image selection and immune cell analysis of human corneal confocal images.

A typical confocal imaging session yields several hundred images per patient, varying in quality and often overlapping in field of view. Currently, these images are manually reviewed, in order to select a high-quality non-overlapping image set of manageable size for subsequent analysis. The density and morphology of corneal immune cells present in these images can be used as markers of ocular and/or systemic disease.

As stated, this evaluation is currently undertaken manually, which is time-consuming, poorly reproducible, and subject to bias. We have developed a methodology that builds on deep learning and signal processing techniques to minimise the time required for identifying and analysing patient image datasets.

Join us for a presentation on their work to date.

Research team:

A/Prof Laura Downie, Eye Disease, Department of Optometry and Vision Sciences, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne

Prof Karin Verspoor, Honorary, Biomedical Natural Language Processing, School of Computing and Information Systems, The University of Melbourne

Dr Holly Chinnery, Senior Lecturer, Department of Optometry and Vision Sciences, Faculty of Medicine, Dentistry and Health Sciences, The University of Melbourne

Dr Vlada Rozova, Honorary, Machine Learning, School of Computing and Information Systems, The University of Melbourne

Recording coming soon