
260 Image Analysis and Modeling in Ophthalmology
and texture methods. The signicance of these features was established using Student’s
t-test before they were used as input to the SVM classication algorithm. The resulting sys-
tem could detect whether or not a particular optical eye image is normal or abnormal with
an accuracy of 90%. This result supports the premise that it is possible to build automated
disease detection systems based on optical eye images. This support is necessary in order
to proceed to the implementation phase, which aims to create a physical solution for the
disease diagnosis problem. Furthermore, the results were obtai ...