Table 2.

Comparison with state-of-the-art methods for pancreatic segmentation on NIH dataset.

Method	Dice	SEN	SPE
U-Net (14)	0.7470	0.5131	0.9964
Attention Unet (39)	0.8437	0.7764	0.9970
Nishio et al. (50)	0.7890	0.6883	0.9977
Li et al. (51)	0.8303	0.7207	0.9965
LMNS-Net (52)	0.8537	0.8115	0.9978
Fixed-point (53)	0.8257	0.6294	0.9968
RSTN (54)	0.8508	0.8466	0.9987
Yu et al. (55)	0.8453	0.8009	0.9972
Zhang et al. (56)	0.8490	0.7852	0.9975
RTUNet (57)	0.8625	0.8916	0.9981
Ours	0.8763	0.9126	0.9988

The bold value indicates the maximum value under the same indicator.