Abstract
Diabetic retinopathy (DR) is a leading cause of blindness worldwide. However, 90% of DR caused blindness can be prevented if diagnosed and intervened early. Retinal exudates can be observed at the early stage of DR and can be used as signs for early DR diagnosis. Deep convolutional neural networks (DCNNs) have been applied for exudate detection with promising results. However, there exist two main challenges when applying the DCNN based methods for exudate detection. One is the very limited number of labeled data available from medical experts, and another is the severely imbalanced distribution of data of different classes. First, there are many more images of normal eyes than those of eyes with exudates, particularly for screening datasets. Second, the number of normal pixels (non-exudates) is much greater than the number of abnormal pixels (exudates) in images containing exudates. To tackle the small sample set problem, an ensemble convolutional neural network (MU-net) based on a U-net structure is presented in this paper. To alleviate the imbalance data problem, the conditional generative adversarial network (cGAN) is adopted to generate label-preserving minority class data specifically to implement the data augmentation. The network was trained on one dataset (e_ophtha_EX) and tested on the other three public datasets (DiaReTDB1, HEI-MED and MESSIDOR). CGAN, as a data augmentation method, significantly improves network robustness and generalization properties, achieving F1-scores of 92.79%, 92.46%, 91.27%, and 94.34%, respectively, as measured at the lesion level. While without cGAN, the corresponding F1-scores were 92.66%, 91.41%, 90.72%, and 90.58%, respectively. When measured at the image level, with cGAN we achieved the accuracy of 95.45%, 92.13%, 88.76%, and 89.58%, compared with the values achieved without cGAN of 86.36%, 87.64%, 76.33%, and 86.42%, respectively.
Highlights
According to 2016 WHO report, from 1980 to 2014 the number of adults living with diabetes has risen from 108 million to 422 million with the global prevalence increasing from 4.7% to 8.5% [1]
To address the problems discussed hereinbefore, we developed an ensemble deep convolutional neural network (MU-net) based on the U-net framework that was mainly designed for medical applications with limited ground truth data
Proposed methodology we presented the modified U-net (MU-Net) structure for exudate detection and demonstrated applying conditional generative adversarial network(cGAN) to generate synthetic images as a new method of data augmentation and minority class upsampling. 3.1
Summary
According to 2016 WHO report, from 1980 to 2014 the number of adults living with diabetes has risen from 108 million to 422 million with the global prevalence increasing from 4.7% to 8.5% [1]. Among many complications that diabetes leads to, diabetic retinopathy (DR) is a significant cause of blindness. After 20 years of diabetes, most patients with Type I diabetes and >60% of patients with Type II diabetes have some degree of retinopathy [3]. At the early stage of DR, patients may not have any symptoms of vision problems. When it develops to the late stage, it may permanently cause vision loss and blindness. It is important for diabetic patients to have a comprehensive retina screening at least once a year. It is shown that blindness due to DR can be prevented in 90% of the cases by early detection through regular screening [4]
Sign-in/Register to view highlights & summary 
Published Version (
Free)
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call