摘要
Diabetic Retinopathy (DR) is a serious hazard that can result inirreversible blindness if not addressed in a timely manner. Hence, numeroustechniques have been proposed for the accurate and timely detection ofthis disease. Out of these, Deep Learning (DL) and Computer Vision (CV)methods for multiclass categorization of color fundus images diagnosed withDiabetic Retinopathy have sparked considerable attention. In this paper,we attempt to develop an extended ResNet152V2 architecture-based DeepLearning model, named ResNet2.0 to aid the timely detection of DR. TheAPTOS-2019 datasetwas used to train the model. This consists of 3662 fundusimages belonging to five different stages of DR: no DR (Class 0), mild DR(Class 1), moderate DR (Class 2), severe DR (Class 3), and proliferativeDR (Class 4). The model was gauged based on ability to detect stage-wiseDR. The images were pre-processed using negative and positive weightedGaussian-based masks as feature engineering to further enhance the qualityof the fundus images by removing the noise and normalizing the images. Upsamplingand data augmentation methods were used to address the skewnessof the original dataset. The proposed model achieved an overall accuracyof 91% and an area under the receiver-operating characteristic curve (AUC)score of 95.1%, outperforming existing Deep Learning models by around 10%.Furthermore, the class-wise F1 score for No DR was 92%, Mild DR was 82%,Moderate DR was 66%, Severe was DR 89% and Proliferative DR was 80%.
