Seminar: Pixel Club

ECE Women Community

On Glaucoma Identification using Computer Vision

Date: July,12,2022 Start Time: 11:30 - 12:30 Add to:
Lecturer: Sahar Blank
Affiliations: The Andrew and Erna Viterbi Faculty of Electrical & Computer Engineering

 

Glaucoma or “the silent thief of sight” is an eye disease that gradually leads to peripheral vision loss and blindness. It is usually painless and advances moderately if not treated on time. Glaucoma affects about 80 million people worldwide. In this work we consider and develop various approaches to diagnose Glaucoma and find the most useful one in terms of accuracy, sensitivity, and specificity. The tested methods are classical image processing and AI-based approaches including feature segmentation with various pre-processing methods, implementation of automated ophthalmologists’ exams, and using extracted medical features as an input to machine-learning models. It has been found that JPEG2000 is a successful pre-processing step that provides efficient input to pre-trained networks. The results achieve over 90% sensitivity, which is the most important parameter for screening tests.

 

M.Sc. student, under the supervision of Moshe Porat and Zvi Friedman.

 

All Seminars
Skip to content