Classification of Ischemic Optic Neuropathy Using Custom Image Processing Algorithm-Statistical Based Analysis

Abstract

Ischemic optic neuropathy (ION), is one of the most well-known disease which could lead to vision loss or blindness. ION impacts the head of the optic nerve (ONH),referred to as anterior (AION), and the other parts of the optic nerve, referred to as posterior (PION).This article presents a customized image processing method based on low pass filtration and segmentation of the area around the ONH , which is the region of interest (ROI) to detect ION. The extracted RIOs features for the optic nerve were analyzed and employed by statistical analysis's such as descriptive analysis and test of hypothesis. To test the sensitivity of the proposed imaging method to discriminate images of individuals with ION from healthy ones, 76 fluorescein angiography images were analyzed. The images were classified based on participants status such as: healthy, ION over the age of 45, ION under the age of 45. The proposed method demonstrated a 95% accuracy in discriminating healthy images from those with ION, and consequently would help identify the clinical features of ION.