Fig. 4: Performance the Freeze-Missing module in secondary hospitals. | npj Digital Medicine

Fig. 4: Performance the Freeze-Missing module in secondary hospitals.

From: A three-tier AI solution for equitable glaucoma diagnosis across China’s hierarchical healthcare system

Fig. 4

Performance of the Freeze-Missing module on multiple missing datasets and its performance comparison with SOTA models. a The line graphs showing accuracy, sensitivity, specificity, and AUC for the Freeze-Missing module, including mean and standard deviation at various missing rates. b The AUC curve, c calibration curve, and d clinical benefit curve for the Freeze-Missing module versus SOTA models (MMD model, Cheerla et al.’s model).

Back to article page