Skip to main content
. 2022 Sep 5;22:382. doi: 10.1186/s12903-022-02422-9

Table 2.

The diagnostic performance of three CNN models in auto-selection group

Accuracy (%) Sensitivity (%) Specificity (%) PPV (%)
Resnet50 91.4 92.1 90.7 90.8
Densenet169 87.1 80.6 93.5 92.6
VGG19 87.8 89.2 86.3 86.7