Table 6. Comparitive study of DMENet’s performance with recent solutions for DME screening using IDRiD dataset.

Author	Year	Technique	Results
He et al. [61]	2019	Auxiliary learning approach and XGBoost classifier	Sensitivity-95.53% Specificity-93.84% Accuracy-94.17%
Li et al. [62]	2019	Cross-disease attention network	Joint Accuracy-65.1% (classification of both DME and DR)
Proposed DMENet	2019	Hierarchical Ensemble of CNNs (HE-CNN)	Sensitivity-97.88% Specificity-94.49% Accuracy-96.76%