Add to Favourites
To login click here
This article presents a review of various deep and machine learning techniques used to identify various complications in diabetic retinopathy (DR). The authors propose a hybrid convolutional neural network involving a grey wolf optimization framework to extract and select the optimal features yielding an optimal classification of DR. Additionally, a two-stage U-Net architecture is proposed for DR segmentation and classification using the CNN-SVD (singular value decomposition) model. The authors also propose an augmented bio-inspired multidomain feature extraction and selection model for diabetic retinopathy severity identification using an ensemble learning process.
