Figure 13.

A segmentation example from INbreast by PeMNet with different backbones. (a) Pre-processed image. (b) $P{e}_{R18}$: 99.27% of IoU, 99.93% of GPA, 98.40% of DSC, 96.85% of Jaccard, 100% of Sensitivity, and 99.86% of Specificity. (c) $P{e}_{R50}$: 99.42% of IoU, 99.94% of GPA, 98.55% of DSC, 97.14% of Jaccard, 100% of Sensitivity, and 99.89% of Specificity. (d) $P{e}_{Mov2}$: 99.46% of IoU, 99.84% of GPA, 98.67% of DSC, 97.37% of Jaccard, 99.75% of Sensitivity, and 99.92% of Specificity. (e) $P{e}_{Xcep}$: 99.64% of IoU, 99.96% of GPA, 98.85% of DSC, 97.72% of Jaccard, 100.00% of Sensitivity, and 99.93% of Specificity. (f) $P{e}_{IRv2}$: 99.91% of IoU, 99.96% of GPA, 99.14% of DSC, 98.30% of Jaccard, 99.96% of Sensitivity, and 99.95% of Specificity.