TABLE 3.

Performance comparison with SOTA methods on ISIC 2017 and PH2 datasets. $\text{Red}$ , $\text{Green}$ , and $\text{Blue}$ indicate the best, second best and third best performance.

Network	ISIC 2017 dataset					PH2 dataset					#Params	Model size
	Acc	IoU	Dice	Sens	Spec	Acc	IoU	Dice	Sens	Spec
FrCN Al-Masni et al. (2018)	$\text{0.940}$	0.771	$\text{0.871}$	0.854	0.967	0.951	0.848	0.918	0.937	0.957	—	—
FocusNet Tong et al. (2015)	0.921	0.756	0.832	0.767	$\text{0.990}$	—	—	—	—	—	—	—
SegNet Kaul et al. (2019)	0.918	0.696	0.821	0.801	0.954	0.934	0.808	0.894	0.865	$\text{0.966}$	28.09M	112 MB
DSNet Gu et al. (2019)	—	$\text{0.775}$	—	$\text{0.875}$	0.967	—	$\text{0.870}$	—	0.929	$\text{0.969}$	$\text{10.00 M}$	—
DAGAN Valanarasu and Patel (2022)	0.935	0.771	$\text{0.859}$	0.835	$\text{0.976}$	—	—	—	—	—	—	—
FATNet Wu et al. (2022)	0.933	0.765	0.850	0.839	0.973	—	—	—	—	—	27.43 M	109 MB
ResGANet Cheng et al. (2022)	0.936	0.764	$\text{0.862}$	0.842	0.950	—	—	—	—	—	39.21M	—
U-Net Ronneberger et al. (2015)	0.926	0.736	0.825	0.828	0.964	0.943	0.851	0.915	0.946	0.957	29.59 M	118 MB
U-Net++ Berseth (2017)	0.929	0.753	0.840	0.848	0.965	0.948	0.853	0.917	$\text{0.973}$	0.937	34.48 M	138 MB
Double U-Net Isensee et al. (2021)	0.936	0.765	0.847	0.830	0.970	0.942	0.860	0.915	0.934	0.953	27.94 M	112 MB
FCANet Hu et al. (2018)	0.935	$\text{0.776}$	0.856	$\text{0.869}$	0.962	$\text{0.952}$	0.868	0.926	$\text{0.968}$	0.926	59.97 M	241 MB
Att U-Net Cheng et al. (2020)	$\text{0.939}$	0.757	0.840	0.859	0.957	0.951	0.868	0.926	0.953	0.955	30.42 M	121 MB
R2U-Net Cheng et al. (2022)	$\text{0.938}$	$\text{0.776}$	$\text{0.858}$	0.859	0.969	0.952	$\text{0.871}$	$\text{0.927}$	0.954	0.960	91.61 M	366 MB
Att R2U-Net Cheng et al. (2022)	$\text{0.939}$	$\text{0.775}$	0.857	0.857	0.961	$\text{0.954}$	$\text{0.873}$	$\text{0.928}$	0.949	0.962	92.11 M	368 MB
BiO-Net (t = 3) Zhou et al. (2018)	0.937	0.772	0.852	0.845	0.973	0.944	0.851	0.915	0.963	0.931	$\text{14.30 M}$	$\text{57.7 MB}$
BiO-Net (t = 3, INT) Zhou et al. (2018)	0.934	0.754	0.840	0.821	$\text{0.976}$	0.945	0.851	0.909	$\text{0.968}$	0.944	$\text{3.99 M}$	$\text{15.2 MB}$
UNeXt-L Valanarasu and Patel (2022)	0.935	0.773	0.856	0.864	0.966	0.949	0.859	0.919	0.941	0.955	$\text{14.30 M}$	$\text{57.7 MB}$
PL-Net† (Our)	0.932	0.769	0.852	$\text{0.871}$	0.953	0.945	0.852	0.915	0.955	0.957	$\text{3.77 M}$	$\text{15.6 MB}$
PL-Net (Our)	$\text{0.940}$	$\text{0.779}$	$\text{0.859}$	0.848	$\text{0.975}$	$\text{0.957}$	$\text{0.873}$	$\text{0.929}$	$\text{0.965}$	$\text{0.966}$	15.03 M	$\text{60.7 MB}$

The bold values indicates the best performance.