. Author manuscript; available in PMC: 2020 Mar 1.

Published in final edited form as: IEEE Trans Cybern. 2018 Feb 8;49(3):1123–1136. doi: 10.1109/TCYB.2018.2797905

TABLE IV.

Segmentation Performance in Terms of Dice Ratio and Standard Deviation, Obtained by the Baseline Comparison Methods and Our Proposed CC-3-D-FCN on 50 Subjects. The Highest Performance in Each Tissue Class Is Highlighted in Bold

	WM	GM	CSF
FAST	.4358(.0903)	.4523(.1397)	.4213(3282)
MV	.6104(.0163)	.7300(.0196)	.5194(.0316)
RF	.8290(.0150)	.8774(.0059)	.9231(.0127)
LINKS	.8971(.0074)	.9241(.0043)	.9420(.0074)
DeepMedic [54]	.8943(.0088)	.9265(.0051)	.9484(.0086)
3D-UNet [55]	.8907(.0087)	.9228(.0049)	.9465(.0095)
CC-2D-FCN	.7861(.0240)	.8615(.0080)	.8846(.0177)
3D-CNN	.8336(.0177)	.8702(.0070)	.9051(.0168)
CC-3D-FCN	.9190(.0085)	.9401(.0052)	.9610(.0090)