Table 3:

Comparing the efficacy of CapsNets and UNets in segmenting images that were not represented in the training data^{^a}

Brain Structure	CapsNet Dice (95% CI)	UNet Dice (95% CI)	CapsNet vs UNet (P Value^{^b})
Thalamus	52% (46–58)	16% (11–21)	< .01
Hippocampus	43% (38–48)	10% (6–14)	< .01

↵a Both models were trained to segment the right thalamus and hippocampus. Then, they were tested on segmenting the contralateral left thalamus and hippocampus.
↵b df = 114 – 1 = 113.