The product is a deep learning based system to predict visual field (VF) function from spectral domain optical coherence tomography (SDOCT) images. Based on SDOCT imaging of the macula and optic nerve head, deep learning models are used to estimate global 24-2 and 10-2 VF summary metrics (mean deviation and pattern standard deviation) as well as visual function at finer scale (sectoral and individual test point pattern deviation). This product is intended to supplement time-consuming VF testing to help clinicians identify patients likely to undergo functional loss. 

For more information about our group, please go to https://shileyeye.ucsd.edu/research/computational_ophthalmology