Multiclass Convolution Neural Network for Classification of COVID-19 CT Images

Table 5

The performance metrics on the test sets.

CNN architecture	TPR sensitivity	TNR specificity	Positive predictive value (PPV) precision	Negative predictive value (NPV)	Accuracy	F1 score
DenseNet201	93.809	90.00	89.14	94.323	91.78	91.41531
GoogLeNet	80.952	72.50	72.034	81.308	76.44	76.23318
ResNet 50	86.19	95.83	94.764	88.803	91.33	90.27431
ResNet 101	88.571	99.58	99.465	90.875	94.44	93.70277
ResNet 152	90	98.33	97.927	91.829	94.44	93.79653
ResNeXt101	95.714	100.00	100	96.386	98.00	97.81022