Correlated analysis of morphological patterns between SD-OCT and FFA imaging for diabetic maculopathy detection: Conformal mapping-based approach

Abstract

The aim of this work is to identify the morphological patterns associated with macular leaks for diabetic maculopathy from spectral-domain optical coherence tomography (SD-OCT) a noninvasive technique and fundus fluorescein angiogram (FFA) an invasive technique. Here, an attempt has been made to identify the morphological pattern of SD-OCT images, which has association with FFA images affected by diabetic maculopathy based on conformal mapping. The preprocessing step consists of removing the speckle noise using different low-pass filters; we found that wavelet filters are efficient. Out of the 60 eyes, we were able to detect pathologies like micro-cysts in around 52 eyes which resulted in an accuracy of *87%. The results also showed that when SD-OCT image look normal, the conformal mapping showed angiogram leakages as micro-cysts. This is the first attempt toward correlating the features of two different modalities in retinal imaging from an image processing perspective.