An optimized learning based retinal image segmentation model for early detection of retinal disorders

Abstract

With rapid advancements in science and technology, image processing has become a major stake holder in almost all fields of application aimed at servicing for society concerns. Analysis of images through appropriate image processing methods, of satellite images for weather prediction and forecasting, remote monitoring using surveillance videos and images, material analysis are some of the prominent and well known applications of image processing in real time. Amongst them, medical image processing has been growing to be one of the most widely attractive and critical fields of research in recent times, due to their immense potential as well as the huge volume of underlying information that needs to be extracted through systematic methods of image processing. Medical images through various modalities like Magnetic Resonance Imaging (MRI), Computer Tomography Images (CT), Positron Emission Tomography Images (PET), X-rays etc. reveal a wide range of indepth information which is critical towards detection of abnormalities and physical health of the individual of concern. Medical image processing, which is quite different, from conventional image processing involving natural images, is a complex process which also demands zero tolerance towards any loss of data during the processing stages. Any minute loss of data may result in wrong diagnosis and consequently to increasing fatality rate. Medical image processing has been a great aiding tool for clinical physicians to speed up the process of early detection of certain critical conditions which helps them to speed up the treatment process. Moreover, continual medical image processing aided with prediction techniques also greatly aid the physicians to observe the growth rate of the tissue or infected newlineregion or any region under interest over a period of time. newline newline