Table 5.

A comparison of prediction accuracy on time-series CT dataset from different methods. All accuracy measures are represented by mean ± standard deviation, which are computed over different patients and time frames.

		Epi	LA	LV	RA	RV	Ao	PA	WH
Dice (↑)	Ours	0.902 ± 0.035	0.96 ± 0.018	0.956 ± 0.033	0.946 ± 0.014	0.944 ± 0.017	0.974 ± 0.006	0.798 ± 0.129	0.94 ± 0.012
	2D UNet	0.913 ± 0.028	0.958 ± 0.014	0.957 ± 0.023	0.927 ± 0.041	0.925 ± 0.041	0.971 ± 0.009	0.867 ± 0.114	0.937 ± 0.022
	3D UNet	0.884 ± 0.03	0.935 ± 0.012	0.946 ± 0.03	0.928 ± 0.019	0.92 ± 0.02	0.955 ± 0.01	0.831 ± 0.059	0.922 ± 0.014
	Voxel2Mesh	0.786 ± 0.072	0.933 ± 0.019	0.928 ± 0.037	0.92 ± 0.021	0.928 ± 0.019	0.924 ± 0.011	0.651 ± 0.123	0.894 ± 0.014
ASSD (nm) (↓)	Ours	0.697 ± 0.308	0.54 ± 0.205	0.574 ± 0.399	0.781 ± 0.21	0.756 ± 0.219	0.28 ± 0.073	2.714 ± 3.079	0.906 ± 0.5
	2D UNet	0.634 ± 0.281	0.569 ± 0.181	0.538 ± 0.25	1.097 ± 0.668	1.099 ± 0.737	0.281 ± 0.103	1.155 ± 1.019	0.767 ± 0.291
	3D UNet	0.811 ± 0.34	0.871 ± 0.277	0.711 ± 0.381	0.993 ± 0.325	1.017 ± 0.267	0.504 ± 0.19	1.598 ± 1.183	0.929 ± 0.259
	Voxel2Mesh	1.297 ± 0.451	0.916 ± 0.208	0.993 ± 0.423	1.194 ± 0.327	1.034 ± 0.275	0.844 ± 0.124	3.788 ± 2.008	1.438 ± 0.325