Table 1.

Dice coefficients for cross-validation of each stage. For each split, 10 patients were selected for validation and the rest was used for training

Mask	Split					$\mu \pm \sigma$
	0	1	2	3	4
Pharynx	0.892	0.885	0.875	0.903	0.8845	$0.89\pm 0.01$
Tongue	0.91	0.905	0.907	0.899	0.891	$0.9\pm 0.007$
Soft palate	0.84	0.833	0.869	0.825	0.8435	$0.842\pm 0.016$