Diabetic retinopathy detection using ensembled transfer learning based thrice CNN with SVM classifier

The present-day highly regarded and active ophthalmic research has paved the way for the detection of several retinal problems, aiding the ophthalmologist in planning and carrying out prompt treatment. Regular screening is one of the most challenging responsibilities due to the greater number of patients with retinal defects and the smaller number of medical specialists. Diabetic Retinopathy is a retinal vascular abnormality. Most people with long-term diabetes mellitus will eventually acquire this condition, which may lead to blindness. Here in this research article work retina fundus images are taken from the both Public Messidor, EyePACs dataset and the in-house clinical dataset from Chaithanya Eye Hospital Kerala. The first stage is to remove noise from the input image and enhance the contrast of the images. For noise reduction, a Trilateral Filter is utilized first, followed by enhancement utilizing contrast-limited Adaptive Histogram Equalization with an unsharp technique. Then Thick Blood vessels are segmented from the enhanced image using the Extended Piecewise Fuzzy C-Means Clustering (EPFCMC) Method. From the segmented image, GLCM features are extracted and then features are selected using War strategy optimization. Finally, an ensemble transfer learning using thrice CNN and SVM classifier model is used which classifies the image as Diabetic Retinopathy (DR) or Normal case. Using Thrice CNN and the SVM classifier model, an accuracy of 98.94% is obtained. The results obtained through this framework for diabetic retinopathy classification prove the effectiveness and wide applicability of the proposed approaches. It is also hoped that the developed automatic detection techniques will assist clinicians in diagnosing Diabetic Retinopathy at an early stage.