Table 4.

Results on the sectionwise classification task for all sections from the Test part of the data. The highest value for each metric is highlighted in bold font.

Setting	Accuracy	Sensitivity	Specificity	${\mathit{F}}_{\mathit{\beta }}$
UNet	0.916	0.972	0.842	0.945
ResNet34-UNet	0.958	0.956	0.960	0.960
ResNet34-UNet + DS	0.964	0.963	0.965	0.966
ResNet34-UNet + Linear	0.959	0.950	0.970	0.958