Uncertainty-aware deep learning methods for robust diabetic retinopathy classification
This work addresses the clinical need for reliable uncertainty estimates in medical image classification, though it is incremental as it builds on existing Bayesian deep learning methods.
The authors tackled the problem of uncertainty estimation in diabetic retinopathy classification by investigating clinical datasets and a 5-class scheme, developing a new uncertainty measure that generalizes better than existing methods, achieving improved performance on clinical data.
Automatic classification of diabetic retinopathy from retinal images has been widely studied using deep neural networks with impressive results. However, there is a clinical need for estimation of the uncertainty in the classifications, a shortcoming of modern neural networks. Recently, approximate Bayesian deep learning methods have been proposed for the task but the studies have only considered the binary referable/non-referable diabetic retinopathy classification applied to benchmark datasets. We present novel results by systematically investigating a clinical dataset and a clinically relevant 5-class classification scheme, in addition to benchmark datasets and the binary classification scheme. Moreover, we derive a connection between uncertainty measures and classifier risk, from which we develop a new uncertainty measure. We observe that the previously proposed entropy-based uncertainty measure generalizes to the clinical dataset on the binary classification scheme but not on the 5-class scheme, whereas our new uncertainty measure generalizes to the latter case.