Detection and Classification of Diabetic Retinopathy using Deep Learning
* Corresponding author
Detection of Diabetic Retinopathy at the early stages could significantly reduce the need for complicated and expensive surgeries. The availability of large datasets has fuelled research in this field. In this project, diabetic retinopathy is detected and classified into five stages: no DR, severe DR, Moderate DR, Proliferative DR, and mild DR. This is made possible with the help of various deep learning techniques. A trained model (ResNet-50 architecture) is used for the extraction of various features from the images. This model gives an accuracy of 0.47% in testing. The dataset used is the Aptos 2019 dataset which is available on Kaggle.
Duraichi N., Jalaja S., C.D. Merlin, Meena Jasmine S., Kamali R.N., Keerthana Manoj Detection and Classification of Diabetic Retinopathy using Deep Learning . Cardiometry; No.26 February 2023; p.-; DOI: .; Available from: https://www.cardiometry.net/issues/no26-february-2023/detection-and-classification-of-diabetic-retinopathy-using-deep-learning