Diabetic Retinopathy

As a weakly supervised learning technique, neural network (NN) has shown an advantage over supervised learning methods for automatic detection of diabetic retinopathy (DR): only the image-level annotation is needed to achieve both detections of DR images and DR lesions, making more graded and de-identified retinal images available for learning. However, the performance of existing studies on this technique is limited by the use of handcrafted features. We propose a NN method for DR detection, which jointly learns features and classifiers from data and achieves a significant improvement on detecting DR images and their inside lesions. Specifically, a pre-trained neural network is adapted to achieve the patch-level DR estimation, and then global aggregation is used to make the classification of DR images.