Projects
Here's a small sampling of what we’re working on…
Improving IOL Power Selection for Cataract Surgery with ML
Selecting the appropriate IOL power in cataract surgery remains a challenge. We have developed ML techniques to predict postoperative IOL position and postoperative refraction with greater accuracy, and are working to extend these benefits to special cases of IOL power selection.
Video-Based AI-Powered Surgical Analysis
Automating the analysis of surgical actions through computer vision techniques has the potential to greatly improve surgical safety. Our team has been developing ML techniques to break new ground in the recognition and analysis of surgical actions, leveraging our BigCat database, containing over 4 million deeply annotated frames of surgical video.
Using AI to Diagnose Tumors of the Ocular Surface
High-resolution anterior segment optical coherence tomography has enabled the non-invasive diagnosis of cancers of the surface of the eye. Our team is working on methods to automate the imaging and diagnosis of ocular surface tumors using AI, as well as attempting to identify novel imaging biomarkers with diagnostic and prognostic value.
Using Bioinformatics Approaches to Identify Candidates for Treatment of TGFBI Corneal Dystrophies
TGFBI-related corneal dystrophies are dominantly inherited and cause loss of vision due to pathogenic proteinaceous deposits of misfolded TGFBIp (transforming growth factor-beta-induced protein). Prevention of TGFBIp aggregation may be an effective approach for halting the recurrent erosions and loss of corneal clarity that are associated with TGFBI-related corneal dystrophies. We are working toward identifying candidates for prevention of TGFBIp aggregation.