Classification of Diabetic Retinopathy via Fundus Photography: Utilization of Deep Learning Approaches to Speed up Disease Detection
This work addresses faster disease detection for diabetic patients, but it is incremental as it applies standard deep learning techniques to a known medical imaging problem.
The paper tackled Diabetic Retinopathy classification from fundus photos by comparing a shallow neural network, which performed well on frequent classes but poorly on less frequent ones, with a transfer learning approach using a deep network, which improved generalization to less frequent classes.
In this paper, we propose two distinct solutions to the problem of Diabetic Retinopathy (DR) classification. In the first approach, we introduce a shallow neural network architecture. This model performs well on classification of the most frequent classes while fails at classifying the less frequent ones. In the second approach, we use transfer learning to re-train the last modified layer of a very deep neural network to improve the generalization ability of the model to the less frequent classes. Our results demonstrate superior abilities of transfer learning in DR classification of less frequent classes compared to the shallow neural network.