The Smart Cyber Ecosystem for Sustainable Development

11Non-Invasive Process for Analyzing Retinal Blood Vessels Using Deep Learning Techniques

Toufique A. Soomro¹*^†, Ahmed J. Afifi^2†, Pardeep Kumar³, Muhammad Usman Keerio⁴, Saleem Ahmed⁵ and Ahmed Ali⁶

¹Electronic Engineering Department QUEST, Larkana, Pakistan

²Computer Vision & Remote Sensing, Technische Universität Berlin, Berlin, Germany

³Computer Systems Engineering QUEST, Nawabshah, Pakistan

⁴Electrical Engineering Department QUEST, Nawabshah, Pakistan

⁵Electronic Engineering Department, Dawood University of Engineering and Technology, Karachi, Pakistan

⁶Electrical Engineering Department, Sukkur IBA University, Sukkur, Pakistan

Abstract

Deep learning is an active area of research, and it demonstrated a considerable effect in diagnosing many diseases through analyzing medical images. In this book chapter, a deep learning model has been designed to analyze the retinal images for eye disease diagnosing. This research work will contain the proposal of analyzes of retinal vessels to overcome the problem of low contrast by proposing an enhancement technique as well as the segmentation of retinal blood vessels. These proposed models will help ophthalmologists analyze the network of retinal vessels by computer to recommend early treatment of eye diseases, especially diabetic retinopathy, and such a procedure has an important role in reducing the injection-based method (invasive eye diseases diagnostic method).

Keywords: Retinal color fundus images, enhancement, Principal Component ...