Table 7 Comprehensive performance comparison of various models
From: Skin lesion segmentation using deep learning algorithm with ant colony optimization
Model | Accuracy (%) | Sensitivity (%) | Specificity (%) | Dice coefficient (%) | Jaccard Index (%) |
|---|---|---|---|---|---|
ResNet | 93.0 | 90.0 | 92.7 | 90.0 | 82.0 |
U-Net | 92.5 | 89.5 | 91.9 | 89.5 | 81.5 |
DeepLabV3 +  | 94.2 | 93.1 | 95.8 | 91.5 | 84.0 |
SegNet-Based Model | 92.5 | 89.8 | 94.6 | 89.4 | 80.9 |
VGG19 with Transfer Learning | 94.0 | 92.3 | 96.2 | 91.0 | 83.5 |
ResNet-50 with Attention Mechanism | 94.5 | 93.7 | 95.9 | 92.3 | 85.1 |
Hybrid ResUNet (Proposed) | 95.8 | 94.2 | 96.7 | 93.1 | 87.5 |