Table 1 Comparison of existing research for skin cancer analysis.
References | Method | Dataset | Outcome | Hyperparameter optimization |
|---|---|---|---|---|
ResNet50, VGG-16, MobileNet | HAM10000 dataset | Precision 83.97% and 88.25% | Grid search | |
DenseNet and Xception | ISIC 2018 | Precision 84.37% and 87.17% | Random search | |
Shuffle-Net, GoogleNet, MobileNet-V2 | ISIC 2020 | Precision 89.37%, 90.24% and 89.07% | Hyper-opt | |
VGG-19, Res-Net 50, Resnet-152v2 | PH-2 | Precision 85.08 and 87.62% | Scikit optimize | |
VGG-16, VGG-19 | ISIC 2020 | Precision 83.75% and 84.92% | Optuna | |
EfficientNet-V2, VGG-19 | Melanoma Skin Cancer Dataset | Precision 86.24% and 87.91% | Search space | |
DenseNet, MobileNet-V3 | DERMIS Dataset | Precision 91.47% and 89.78% | Grid search | |
Efficient-Net, ResNet-50 | ISIC 2020 | Precision 90.77% and 89.64% | Random search | |
Inception-V3, Xception | ISIC-2019 and ISIC-2020 | Precision 89.81% and 87.54% | Hyper-opt | |
VGG-16 and VGG-19 | ISIC 2016 | Precision 86.22% and 89.38% | Scikit optimize | |
GoogleNet, Efficient-Net | ISIC 2018 | Precision 88.74% and 90.51% | Optuna | |
Proposed Model | Standard U-Net and MobileNet-V3 | HAM10000 | Precision of 97.84% | Bayesian optimization and grid search |
