MSeg-Net: A Melanoma Mole Segmentation Network Using CornerNet and Fuzzy -Means Clustering

Table 4

Comparative analysis of our work with performance scores from the ISIC-2018 challenge leaderboard.

Method	Jaccard index	Dice	Accuracy	Specificity	Sensitivity
Mask-RCNN2+segmentation	0.838	0.898	94.20%	96.30%	90.60%
Ensemble with CRFv3	0.837	0.904	94.50%	95.20%	93.40%
Lesion segmentation by DCNN	0.834	0.900	94.30%	96.40%	91.80%
Proposed	0.9783	0.9886	99.63%	99.39%	99.48%