Table 2.

Accuracy metrics of the four classifier networks for three-class tissues

Architecture	Accuracy	Size	Precision			Recall			Dice			AUC
			Cancer	Fat	Muscle	Cancer	Fat	Muscle	Cancer	Fat	Muscle	Cancer	Fat	Muscle
GoogLeNet	46.88	0.96	69.02	86.30	11.52	62.02	25.54	63.18	60.34	37.16	18.21	91.28	89.70	84.82
Inception-ResNet-v2	52.36	1.00	69.32	82.88	12.94	67.26	29.46	59.62	67.71	40.74	20.49	91.60	89.24	86.18
ResNet-50	50.58	1.00	68.02	84.40	13.30	65.86	28.02	63.04	65.50	39.79	20.84	91.01	89.88	86.30
Siamese	54.64	1.21	67.26	83.44	12.94	76.74	28.96	46.34	70.23	39.84	18.84	40.42	28.55	33.80