Abstract:
Despite advancements in medical imaging technology, the interpretation of medical imagery still necessitates the expertise of specialists, which can present challenges in terms of practicality and accessibility. However, emerging technologies such as deep learning offer a promising solution to address these challenges. By leveraging deep learning algorithms, the accuracy and efficiency of diagnosis can be significantly improved, enabling faster and easier identification of various medical conditions. Among these conditions, diabetic retinopathy stands out as one that critically necessitates the advancements offered by deep learning. In this work, we exploit GANs, CNNs and Transfer learning to diagnose Diabetic Retinopathy (DR), by proposing an architecture that also allows to augment the data from real images. The experimental results obtained are very promising and a part of this work has been presented in an international conference " Colloque sur les Objets et systèmes Connectés-COC 2023 ".
