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

Table 3

Comparative analysis of our work with performance values from the ISIC-2017 challenge leaderboard.

Method	Jaccard index	Dice	Accuracy	Specificity	Sensitivity
CDNN [66]	0.7650	0.8490	93.40%	97.50%	82.50%
U-Net [67]	0.7620	0.8470	93.20%	97.80%	82.00%
Deep residual network [68]	0.7600	0.8440	93.40%	98.50%	80.20%
U-Net [69]	0.7540	0.8390	93.10%	96.90%	81.70%
FCNN [70]	0.7.20	0.8370	93.00%	97.60%	81.30%
Proposed	0.9693	0.9813	99.32%	99.68%	98.76%