Glaucoma, an eye condition which leads to permanent blindness, is typically asymptomatic and therefore difficult to be diagnosed in time if a doctor is not seen regularly. However, if diagnosed in time, Glaucoma can effectively be slowed down by using adequate treatment, therefore an early diagnosis is of utmost importance.
Nonetheless, the conventional approaches to diagnose Glaucoma adopt expensive and bulky equipment that requires qualified experts, making it difficult, costly and time-consuming to diagnose large amounts of people.
Considering the aforementioned, this work proposes a solution to this problem, by developing a computeraided diagnosis system that is capable of diagnosing Glaucoma, in an autonomous way, using fundus images. These images contain the necessary morphological features to determine the presence of Glaucoma and can be obtained with considerably more affordable equipment, being therefore very relevant for the task at hand.
Several computer-aided systems for Glaucoma assessment with fundus images have already been proposed, however, they lack in several aspects.
This work is a step forward in the field, by improving existing techniques and addressing still unexplored fields.
The developed solution uses Deep Learning techniques powered by a collection of public datasets and an acquired private dataset to construct an interpretable Glaucoma assessment pipeline that runs offline in mobile and embedded devices.
This pipeline is then integrated on the EyeFundusScope android application, developed by Fraunhofer AICOS, which is part of a larger system that already provides solutions for fundus image acquisition and Diabetic Retinopathy detection.
The result demonstrates the potential that these systems can have in the contribution to an early Glaucoma diagnosis and enhances the EyeFundusScope system, turning it into a more complete screening tool.
Author: José Martins
Type: MSc thesis
Partner: Faculdade de Engenharia da Universidade do Porto