Table 1.

Means (bold) and standard deviations of DSC (and MRDSC) agreement of the differently trained segmenters on the two datasets. Multi-rater comparison was possible on the HAD only. The inter-rater agreement between R1 and R2 on HAD is $67.54\pm 10.03$%

Net	Healthy aging dataset				PD dataset
	R1	R2	Intersection	MRDSC	R1
NetR1	$\mathbf{70.53}\%\pm 11.12\%$	$\mathbf{69.27}\%\pm 11.19\%$	$\mathbf{69.14}\%\pm 11.94\%$	$\mathbf{69.87}\%\pm 9.84\%$	$\mathbf{57.89}\%\pm 12.90\%$
NetR2	$\mathbf{68.82}\%\pm 11.32\%$	$\mathbf{71.49}\%\pm 11.79\%$	$\mathbf{70.77}\%\pm 11.56\%$	$\mathbf{70.04}\%\pm 10.33\%$	$\mathbf{59.16}\%\pm 12.98\%$
NetInt	$\mathbf{66.06}\%\pm 13.45\%$	$\mathbf{68.41}\%\pm 13.67\%$	$\mathbf{73.92}\%\pm 11.72\%$	$\mathbf{66.81}\%\pm 12.63\%$	$\mathbf{54.13}\%\pm 15.28\%$
NetRnd	$\mathbf{71.59}\%\pm 10.08\%$	$\mathbf{71.83}\%\pm 10.51\%$	$\mathbf{70.77}\%\pm 11.02\%$	$\mathbf{71.71}\%\pm 8.81\%$	$\mathbf{60.87}\%\pm 12.83\%$