Hypertensive Retinopathy Screening through Fundus Images-A Review

Automatic segmentation of fundus images is an important task in computer-aided diagnosis (CAD) for analysis of medical images to diagnose diseases such as Hypertensive retinopathy, diabetic Retinopathy (HR). HR occurs due to hypertension for a prolonged period of time. It may lead to vision loss, if not treated at early stages. It can be observed by the changes in retinal vasculature that are caused by arterial hypertension. There are various features observed in fundus images such as arterial narrowing, bifurcation, tortuosity, etc. Computer-aided diagnosis plays a vital role in screening and grading of retinal images. There is still a need for automatic detection and grading of HR. This article presents the state-of-the-art methodologies used by researchers for predicting hypertensive retinopathy. Features that are present in retinal images have also been discussed for early detection of HR. The probable extraction techniques have been explored and evaluated based on layers used by different models. The classification technique discussed by different researchers has also been explored followed by the conclusion. This review will be beneficial for the researchers who want to focus on medical image analysis and enhancing the diagnosis of the system through CAD.