Multiresolution Mutual Assistance Network for Cardiac Magnetic Resonance Images Segmentation

Table 7

Comparison with state-of-the-art methods on the ACDC.

		UNet [24]	Att-UNet [35]	UNet++ [36]	UNet3+ [37]	TransUNet [38]	Swin-UNet [39]	Ours
Dice	RV	0.892	0.895	0.900	0.905	0.901	0.841	0.920
	Myo	0.861	0.851	0.858	0.854	0.845	0.775	0.881
	LV	0.939	0.940	0.936	0.947	0.942	0.909	0.960
Specificity	RV	0.998	0.999	0.999	0.999	0.999	0.997	0.999
	Myo	0.999	0.998	0.999	0.999	0.998	0.997	0.998
	LV	0.999	0.999	0.999	0.999	0.999	0.999	0.999
Sensitivity	RV	0.891	0.885	0.894	0.900	0.899	0.855	0.909
	Myo	0.852	0.849	0.840	0.843	0.868	0.800	0.883
	LV	0.929	0.928	0.931	0.939	0.950	0.894	0.953
F1	RV	0.894	0.898	0.902	0.907	0.903	0.845	0.921
	Myo	0.863	0.854	0.859	0.856	0.847	0.777	0.881
	LV	0.941	0.942	0.939	0.949	0.945	0.910	0.960

The best performance is shown in bold.