Transfer Learning based Diabetic Retinopathy Detection with a Novel Preprocessed Layer

Abstract

one of the major reasons for impaired vision in the world nowadays is diabetic retinopathy (DR). Many people could be saved from permanent blindness with early detection. The manual diagnosis is erroneous and tedious. Hence, numerous computerized vision methods for the automatic detection of diabetic retinopathy and its distinctive stages from retinal images were proposed. Various image processing techniques have been developed besides deep learning methods. In image processing techniques, complex features are manually identified. Most of the earlier works used very small dataset which has a great chance to be over-fitting and worked with grayscale image after transforming color fundus images. In our paper, we developed a deep learning model with transfer learning from VGG16 model followed by a novel color version preprocessing technique. It reduced the training time and provided an average accuracy of 0.9132683 implemented to new Kaggle dataset “APTOS 2019 Blindness Detection”. Moreover, to avoid the over-fitting problem for long run we used Stratified K-fold cross validation.