Machine Learning for Maximizing the Detection Rate of Diabetic Retinopathy Using Image Processing

Diabetic retinopathy (DR) is a serious diabetes condition that harms the retina and can result in blindness if not treated early. DR is diagnosed by examining the retina pictures of the eye. However, manually grading photos to determine the severity of DR disease needs a significant amount of resources and time. Automated systems give accurate results along with saving time. Ophthalmologists may find it useful in reducing their workload. The aim of the study is to present an approach to maximize the DR detection rate. The proposed work presents the method to correctly identify the lesions and classify DR images efficiently. Blood leaking out of veins forms features such as exudates, microaneurysms, and hemorrhages, on the retina. Image processing techniques assist in DR detection. Median filtering is used on gray-scale converted images to reduce noise. The features of the pre-processed images are extracted by textural feature analysis. Optic disc (OD) segmentation methodology is implemented for the removal of OD. Blood vessels are extracted using haar wavelet filters. KNN classifier is applied for classifying retinal images into diseased or healthy. The proposed algorithm is executed in MATLAB software and analyzed results with regard to certain parameters such as accuracy, sensitivity, and specificity. The outcomes prove the superiority of the new method with a sensitivity of 92.6%, specificity of 87.56%, and accuracy of 95% on the Diaretdb1 database. The study concluded that the new method is capable of identifying true lesions and rejecting false ones. The suggested technique provides better results in comparison with the state-of-the-art technique.