Detection of Choroidal Neovascularization through Parametric Modeling of the Intensity Variation in Fluorescein Angiograms

Abstract

: Choroidal Neovascularization (CNV) is a devastating sequela resulting from a wide range of disorders that affect the Retinal Pigment Epithelium (RPE)-Bruch's membrane-choriocapillaris complex. That is popularly known as Age Regulated Degeneration (AMD), which is the leading cause of permanent serious loss of vision among senior citizens in the developed parts of the world. A mathematical model implemented in this work is used to detect the stages of CNV using artificial intelligence technique. This model is developed based on the intensity variation caused by the size of lesions in the FA (fluorescein angiography) sequence with respect to time. A collection of features is calculated from the mathematical model, and a feature vector is constructed using those features. These feature vectors are used to categorize the severity of the disease through a classifier. The mathematical model of lesions in CNV is used to find out the pattern that CNV follows which enhances the pattern recognition capabilities of the classifier. Defining CNV lesions is typically a time-consuming and tiresome task. Therefore, this proposed method of developing a mathematical model of the intensity variation in the FAs will help to identify CNV efficiently. Irrespective of the machine learning classifiers, this model provides good accuracy, sensitivity, and specificity and F1 score