TABLE 2.

Performance of binary classification on the DFU ischaemia dataset.

Network	Accuracy	Sensitivity	Precision	Specificity	F-measure	AUC score
ResNet-18	88.30 ± 1.36	82.40 ± 3.48	93.23 ± 1.73	94.16 ± 1.03	87.43 ± 2.03	95.48 ± 0.89
ResNet-50	88.13 ± 1.77	81.46 ± 3.49	93.80 ± 1.42	94.76 ± 0.59	87.16 ± 2.26	95.09 ± 1.24
ResNet-101	89.95 ± 1.29	85.17 ± 2.50	94.10 ± 1.25	94.78 ± 0.55	89.38 ± 1.37	96.29 ± 1.17
ResNet-152	88.62 ± 2.18	82.45 ± 3.16	93.92 ± 2.22	94.84 ± 1.55	87.79 ± 2.42	95.58 ± 1.15
RegNetY-4GF	89.55 ± 0.89	83.64 ± 1.53	94.66 ± 1.53	95.41 ± 1.05	88.80 ± 1.34	95.92 ± 1.33
RegNetY-8GF	89.36 ± 1.23	83.64 ± 1.61	94.45 ± 0.60	95.13 ± 0.45	88.70 ± 0.80	95.59 ± 1.07
RegNetY-16GF	90.48 ± 1.01	85.54 ± 2.00	94.79 ± 1.54	95.40 ± 1.19	89.91 ± 1.25	96.43 ± 1.03
EfficientNet-B0	87.26 ± 1.74	79.07 ± 3.89	94.38 ± 1.11	95.38 ± 0.68	85.99 ± 2.29	94.81 ± 0.90
EfficientNet-B2	88.23 ± 0.71	81.17 ± 2.42	94.37 ± 1.46	95.20 ± 1.23	87.24 ± 1.21	95.67 ± 0.92
EfficientNet-B4	87.24 ± 1.77	79.11 ± 3.75	94.27 ± 1.30	95.31 ± 0.67	85.98 ± 2.33	94.46 ± 1.22
EfficientNet-B6	87.40 ± 2.38	80.57 ± 4.96	93.41 ± 0.95	94.34 ± 0.95	86.41 ± 2.50	94.53 ± 1.21
EfficientNet-B7	86.41 ± 2.40	78.57 ± 4.95	93.06 ± 0.90	94.21 ± 0.58	85.11 ± 2.94	94.64 ± 1.62
MoCo	89.74 ± 1.29	86.01 ± 3.24	92.92 ± 1.86	93.56 ± 1.53	89.28 ± 1.41	95.70 ± 1.01
DeiT-S	88.89 ± 2.13	82.35 ± 4.19	94.58 ± 1.08	95.38 ± 0.64	87.99 ± 2.54	96.45 ± 0.95
DeiT-B	89.10 ± 2.32	81.76 ± 5.02	95.84 ± 1.13	96.50 ± 0.80	88.13 ± 2.68	96.19 ± 1.49
DeiT-S-D	89.96 ± 1.88	83.44 ± 3.65	95.96 ± 1.06	96.58 ± 0.63	89.21 ± 1.96	97.06 ± 1.07
DeiT-B-D	89.69 ± 1.93	82.51 ± 3.48	96.29 ± 0.63	96.88 ± 0.17	88.83 ± 2.12	96.61 ± 1.01
CKB-DeiT-S-D	90.27 ± 1.90	84.09 ± 4.00	95.97 ± 1.41	96.59 ± 0.86	89.57 ± 2.04	97.28 ± 0.91
CKB-DeiT-B-D	90.90 ± 1.74	86.09 ± 2.98	95.00 ± 1.29	95.59 ± 0.71	90.30 ± 1.83	96.80 ± 1.16