Skip to main content
. 2021 Mar 31;68:102588. doi: 10.1016/j.bspc.2021.102588

Table 9.

Evaluation results for each network in each fold.

Fold Network Correct classified images Wrong classified images COVID correct classified COVID not classified Wrong classified as COVID Normal correct classified Normal not classified Wrong classified as normal Overall accuracy COVID accuracy Normal accuracy COVID sensitivity Normal sensitivity COVID specificity Normal specificity COVID precision Normal precision
Our model 8214 108 436 26 82 7778 82 26 0.987 0.987 0.987 0.9437 0.9896 0.9896 0.9437 0.8417 0.9967
1 Xception 8165 157 456 6 151 7709 151 6 0.9811 0.9811 0.9811 0.987 0.9808 0.9808 0.987 0.7512 0.9992
ResNet50V2 8169 153 453 9 144 7716 144 9 0.9816 0.9816 0.9816 0.9805 0.9817 0.9817 0.9805 0.7588 0.9988



Our model 8215 128 443 22 106 7772 106 22 0.9847 0.9847 0.9847 0.9527 0.9865 0.9865 0.9527 0.8069 0.9972
2 Xception 7921 422 458 7 415 7463 415 7 0.9494 0.9494 0.9494 0.9849 0.9473 0.9473 0.9849 0.5246 0.9991
ResNet50V2 8109 234 456 9 225 7653 225 9 0.972 0.972 0.972 0.9806 0.9714 0.9714 0.9806 0.6696 0.9988



Our model 8143 186 427 19 167 7716 167 19 0.9777 0.9777 0.9777 0.9574 0.9788 0.9788 0.9574 0.7189 0.9975
3 Xception 8113 216 434 12 204 7679 204 12 0.9741 0.9741 0.9741 0.9731 0.9741 0.9741 0.9731 0.6803 0.9984
ResNet50V2 8079 250 439 7 243 7640 243 7 0.97 0.97 0.97 0.9843 0.9692 0.9692 0.9843 0.6437 0.9991



Our model 8205 110 442 17 93 7763 93 17 0.9868 0.9868 0.9868 0.963 0.9882 0.9882 0.963 0.8262 0.9978
4 Xception 7854 461 449 10 451 7405 451 10 0.9446 0.9446 0.9446 0.9782 0.9426 0.9426 0.9782 0.4989 0.9987
ResNet50V2 8166 149 439 20 129 7727 129 20 0.9821 0.9821 0.9821 0.9564 0.9836 0.9836 0.9564 0.7729 0.9974



Our model 8141 94 419 31 63 7722 63 31 0.9886 0.9886 0.9886 0.9311 0.9919 0.9919 0.9311 0.8693 0.996
5 Xception 8058 177 440 10 167 7618 167 10 0.9785 0.9785 0.9785 0.9778 0.9785 0.9785 0.9778 0.7249 0.9987
ResNet50V2 7991 244 449 1 243 7542 243 1 0.9704 0.9704 0.9704 0.9978 0.9688 0.9688 0.9978 0.6488 0.9999



Our model 8183.6 125.2 433.4 23 102.2 7750.2 102.2 23 0.9849 0.9849 0.9849 0.9496 0.987 0.987 0.9496 0.8126 0.997
Avg Xception 8022.2 286.6 447.4 9 277.6 7574.8 277.6 9 0.9655 0.9655 0.9655 0.9802 0.9647 0.9647 0.9802 0.636 0.9988
ResNet50V2 8102.8 206 447.2 9.2 196.8 7655.6 196.8 9.2 0.9752 0.9752 0.9752 0.9799 0.9749 0.9749 0.9799 0.6988 0.9988

The bold values show the maximum value in a column.