Certain investigations of diabetic retinopathy in retinal images using texture features and machine learning methods

Abstract

Diabetes mellitus is a disease worldwide, which can cause serious results. The estimated prevalence of diabetes for all age groups worldwide was 2.8% in 2000 and is expected to be 4.4% in 2030, meaning that the total number of diabetes patients is forecasted to rise from 171 million in 2000 to 366 million in 2030. According to a report from World Health Organization (WHO), more than 347 million people worldwide have diabetes, and WHO proposes that diabetes will be the 7th leading cause of death in 2030. Diabetic Retinopathy (DR) is a serious complication of diabetes, and is the leading cause of blindness among people of working age in developed countries. It has long been a desire to develop a convenient and cost-effective computeraided auxiliary diagnosis system, which is able to help DR patients in under-developed areas to know timely their glaucoma conditions and obtain treatment suggestions from doctors. Accurate and timely diagnosis is essential for successful treatment of any disease. Computer Aided Diagnosis (CAD) may prove beneficial for screening of diseases over a large population and may be time saving as compared to the physical examination by medical professionals. It will augment and aid the clinical healthcare in the developing countries where there is shortage of trained professional ophthalmologists. In this thesis work, demonstrated a possible solution of a computeraided healthcare system for DR classification and grading based on fundus image analysis. Retina consists of several light-sensitive neuron layers, lining the inner surface of the eye, in which many diseases manifest themselves, such as macular degeneration, glaucoma and diabetic retinopathy. Ophthalmologists and scientists have been seeking the approach to examining the retina for a long time. In the first work, a novel DR-CAD system can be of paramount significance for efficient DR diagnosis for retinal images using various intensity, GLCM, DWT and TCM based features. The overall classification accuracy of hybrid color and structure