Table 2.

Performance of the models tested (SzNet-5, SzNet-15 and SzNet-35) using different spatial ROI evaluated with five metrics: accuracy; AUC-ROC; precision; recall; and specificity.

Model	ROI	Accuracy	AUC-ROC	Precision	Recall	Specificity
SzNet-35	-	0.71 ± 0.07	0.71 ± 0.08	0.70 ± 0.7	0.77 ± 0.09	0.65 ± 0.11
SzNet-15	Left	0.67 ± 0.09	0.67 ± 0.09	0.67 ± 0.14	0.79 ± 0.15	0.55 ± 0.26
SzNet-15	Midline	0.76 ± 0.07	0.76 ± 0.07	0.79 ± 0.13	0.77 ± 0.17	0.74 ± 0.19
SzNet-15	Right	0.73 ± 0.10	0.72 ± 0.10	0.69 ± 0.10	0.88 ± 0.15	0.56 ± 0.19
SzNet-5	Left3	0.71 ± 0.10	0.70 ± 0.10	0.70 ± 0.08	0.77 ± 0.19	0.63 ± 0.17
SzNet-5	Left2	0.72 ± 0.05	0.71 ± 0.04	0.75 ± 0.11	0.72 ± 0.20	0.71 ± 0.17
SzNet-5	Left1	0.67 ± 0.08	0.67 ± 0.08	0.71 ± 0.13	0.70 ± 0.20	0.65 ± 0.24
SzNet-5	Midline	0.78 ± 0.08	0.78 ± 0.08	0.82 ± 0.13	0.77 ± 0.15	0.79 ± 0.17
SzNet-5	Right1	0.77 ± 0.08	0.77 ± 0.09	0.81 ± 0.21	0.72 ± 0.21	0.77 ± 0.09
SzNet-5	Right2	0.72 ± 0.11	0.72 ± 0.11	0.70 ± 0.10	0.78 ± 0.19	0.65 ± 0.14
SzNet-5	Right3	0.67 ± 0.07	0.67 ± 0.07	0.70 ± 0.14	0.72 ± 0.23	0.61 ± 0.26
SzNet-5	Midline**	0.77 ± 0.09	0.77 ± 0.09	0.86 ± 0.16	0.68 ± 0.15	0.85 ± 0.16

The highest values achieved for each metric are highlighted in a different color. Similarly, the model and ROI that achieved the best results are marked.

^**randomly aligned ROI.