Table 3.

Comparison of positive predictive values according to the location of the attention modules.

Configurations	Datasets
	Montgomery	JSRT	Shenzhen
$\mathrm{U}-\mathrm{Net}+X\left(1\right)$	0.971 ± 0.001	0.966 ± 0.001	0.956 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(2\right)$	0.969 ± 0.001	0.965 ± 0.001	0.955 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(3\right)$	0.965 ± 0.001	0.964 ± 0.001	0.955 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(4\right)$	0.965 ± 0.002	0.965 ± 0.001	0.954 ± 0.001
$\mathrm{U}-\mathrm{Net}+Y\left(1\right)$	0.969 ± 0.001	0.965 ± 0.001	0.954 ± 0.002
$\mathrm{U}-\mathrm{Net}+Y\left(2\right)$	0.967 ± 0.001	0.966 ± 0.001	0.954 ± 0.001
$\mathrm{U}-\mathrm{Net}+Y\left(3\right)$	0.965 ± 0.001	0.965 ± 0.001	0.953 ± 0.001
$\mathrm{U}-\mathrm{Net}+Y\left(4\right)$	0.965 ± 0.001	0.965 ± 0.001	0.952 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(1\right)$ $+X\left(2\right)$	0.970 ± 0.001	0.968 ± 0.001	0.956 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(3\right)$ $+X\left(4\right)$	0.969 ± 0.001	0.966 ± 0.001	0.956 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(1\right)+X\left(2\right)+X\left(3\right)+X\left(4\right)$	0.968 ± 0.001	0.966 ± 0.001	0.953 ± 0.001
$\mathrm{U}-\mathrm{Net}+Y\left(1\right)$ $+Y\left(2\right)$	0.970 ± 0.001	0.966 ± 0.001	0.956 ± 0.001
$\mathrm{U}-\mathrm{Net}+Y\left(3\right)$ $+Y\left(4\right)$	0.968 ± 0.001	0.965 ± 0.001	0.958 ± 0.003
$\mathrm{U}-\mathrm{Net}+Y\left(1\right)$ $+Y\left(2\right)+Y\left(3\right)$ $+Y\left(4\right)$	0.969 ± 0.001	0.966 ± 0.001	0.956 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(1\right)$ $+X\left(2\right)+Y\left(1\right)$ $+Y\left(2\right)$	0.974 ± 0.001	0.971 ± 0.001	0.960 ± 0.001
$\mathrm{U}-\mathrm{Net}+X\left(3\right)$ $+X\left(4\right)+Y\left(3\right)$ $+Y\left(4\right)$	0.967 ± 0.001	0.967 ± 0.001	0.956 ± 0.001