Detection for diabetic retinopathy and diabetic macular oede
A Study was conducted to present and validate a deep ensemble algorithm to detect diabetic retinopathy (DR) and diabetic macular oedema (DMO) using retinal fundus images.

A total of 8739 retinal fundus images were collected from a retrospective cohort of 3285 patients. For detecting DR and DMO, a multiple improved Inception-v4 ensembling approach was developed. The algorithm’s performance was measured and a comparison was made with that of human experts on the primary dataset, while its generalization was assessed on the publicly available Messidor-2 dataset. Also, the impact of the size and number of input images used in training on model’s performance was investigated systematically , respectively.

--On our primary test dataset, the model achieved an 0.992 AUC corresponding to 0.925 sensitivity and 0.961 specificity for referable DR, while the sensitivity and specificity for ophthalmologists ranged from 0.845 to 0.936, and from 0.912 to 0.971, respectively.

--For referable DMO, our model generated an AUC of 0.994 with a 0.930 sensitivity and 0.971 specificity, whereas ophthalmologists obtained sensitivities ranging between 0.852 and 0.946, and specificities ranging between 0.926 and 0.985.

This study finally revealed that the deep ensemble model had outstanding performance in DR and DMO detection and had good strength and generalization that might help to promote and expand DR/DMO screening programs.