Fundus Image Based DR Detection Using Image Processing

Abstract

Almost any field of application that aims to address societal issues now relies heavily on image processing, thanks to recent breakthroughs in the field. Medical image processing is among the most widely appealing and vital disciplines of research in recent times, owing to its enormous potential as well as the vast volume of underlying information that needs to be recovered using systematic image processing techniques. However, this research work, has taken processing of retinal images as its prime issue of concern to predict and aid early detection of retinal disorders. This arises from the motivation that certain symptoms appearing in a particular part of the human body may indicate onset of certain diseases in which diabetes could be quoted as the best example. Retina of the eye is an essential component from an image processing aspect, as it reveals several disorders associated with it at an earlier stage, if properly processed and analysed. Some well-known retinal disorders include diabetic retinopathy, retinal tear and detachment, cataract, retinitis pigmentosa, age related macular disorders (AMD) etc. if these conditions are left undetected, they may result in increasing disorders affecting other parts of the human body. In view of the above issues and prospects of retinal image analysis, this thesis has taken retina as the prime subject of study for early detection of retinal disorders.