Skip to main content
. 2021 Oct 14;11:20384. doi: 10.1038/s41598-021-99986-3

Figure 4.

Figure 4

Discriminative performance of Chest-Xray classification algorithms on prospectively collected testing set of 3289 images. Discriminative capacity of CovIx (blue), COVID-Net (red), and DeepCOVID-XR (black) is represented as Receiver Operating Characteristics (ROC) (A-D) or Precision-Recall (PR) (E-H) curves. 95% Confidence Intervals (CI), generated using 2000 bootstrap samples, are visualised as pale curves.