Table 4.

Performance metrics on TBX11K.

Model	Accuracy	AUC (TB)	Sensitivity	Specificity	Average precision	Average recall
ResNet18	91.01	95.69	76.82	96.08	89.48	88.75
ResNet50	94.85	98.76	87.02	96.99	93.67	93.6
DenseNet121	88.04	94.82	71.41	95.89	86.85	85.13
DenseNet201	92.64	97.23	79.91	97.18	91.63	90.6
InceptionV3	89.58	94.95	69.4	97.51	89	86.21
EfficientNet-B7	84.07	73.82	51.78	95.86	82.14	78.39
Proposed model	96.91	99.38	91.81	98.42	96.33	96.1