Attention-based dual-path feature fusion network for automatic skin lesion segmentation

Table 2 Performance evaluation of different segmentation methods on ISIC2018 dataset

Methods	F1-score	SE	SP	AC	PC	JS
UNet [18]	0.742	0.708	0.964	0.890	0.779	0.590
Attention UNet [47]	0.750	0.717	0.967	0.897	0.787	0.600
R2UNet [48]	0.766	0.792	0.928	0.880	0.741	0.620
FCN [17]	0.852	0.837	0.966	0.938	0.868	0.742
UNet++ [49]	0.856	0.817	0.975	0.942	0.900	0.748
BCDUNet [24]	0.851	0.785	0.982	0.937	0.928	0.740
HiFormeS [25]	0.883	0.928	0.911	0.918	0.848	0.795
Ours(VGG)	0.873	0.827	0.982	0.950	0.924	0.774
Ours(ResNet101)	0.890	0.933	0.918	0.927	0.880	0.819

ISSN: 1756-0381