Diabetic retinopathy detection through image mining for type 2 diabetes

摘要

Diabetes leads to death in developing countries, detection of diabetes is a work of research. Researchers are continuously working to detect diabetes by methods of image processing, data mining and various areas. Diabetes is classified as type 1 or type 2 diabetes. Type 1 diabetes is hard to cure and leads to death but type 2 diabetes can be cured if detected at early stage. Type 2 diabetes can be detected easily through retina analysis known as Diabetic Retinopathy analysis. Image processing techniques have been applied by many researchers to detect the DR presence in retina. In our work we have proposed a technique using image processing and data mining, texture and wavelet features are extracted for detection. The Results are obtained for image of standard database DIARETDB1 and evaluated using parameters of sensitivity, specificity and accuracy. Our technique yields a high accuracy of 97.75% which can help to detect and prevent diabetes.