The advances of Artificial Intelligence (AI) methods have resulted in power algorithms that can outperform humans in performing tasks that require human-level intelligence such as classification, recognition, and detection tasks. These algorithms have been used in many medical areas and they solved many difficult tasks such as diagnosis of different diseases and detecting abnormalities in medical data (i.e., signals, images, …, etc.) with performance that can outperform expert performance. In our work, we are focusing on using AI methods in the field of ophthalmology with more focus on the anterior segment of the eye. We have developed AI systems for the diagnosis of corneal diseases such as dry eye syndrome, keratoconus, and Fuchs’ endothelial dystrophy using optical coherence tomography (OCT) images. The diagnosis of corneal diseases is done using many exams and with the help of expert specialists, but we could obtain high and accurate results using only OCT images. Also, we developed methods for the automatic detection of corneal layer boundaries in OCT images and compared them to manual operators. Our results are better that the classical methods. In conclusion, AI methods are powerful methods that can change and improve the standard-of-care methods for ophthalmology and cornea in specific.

Additional author(s): Mohamed Abou Shousha, Mohamed Abdel-Mottaleb