Various Glaucoma related machine learning and computer vision projects done with Caprioli lab at Stein Eye.

A model for generating rim segmentations from images of the optic nerve head. We strove to achieve good performance even in instances where the rim is incomplete.

Paper describing the work and authorship:

TODO(tyler): Add links to the papers

Tool to grade optic nerve head fundus images with the disc damagelihood scale. DDLSNet takes the rim to disc ratio and absent rim width values from RimNet, combines them with a disc size category generated by a classification model (DiscNet), and calculates DDLS. The grading process is done in the script Final Pathway.py.

TODO(Tyler): Add some pictures to illustrate how this works?