Table 7 Class-wise accuracy of proposed feature extraction with DNN models for skin disease diagnosis over k-fold cross validation.
From: Skin disease diagnostics through federated transfer learning on heterogeneous data
Class | Model | K-fold cross validation | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
1 | 2 | 3 | 4 | 5 | 6 | 7 | 8 | 9 | 10 | ||
MV | DenseNet + DNN | 87.832 | 87.865 | 87.841 | 87.879 | 87.823 | 87.856 | 87.848 | 87.870 | 87.834 | 87.862 |
VGG19 + DNN | 88.695 | 88.728 | 88.704 | 88.742 | 88.687 | 88.719 | 88.711 | 88.735 | 88.698 | 88.726 | |
Xception + DNN | 89.493 | 89.526 | 89.502 | 89.540 | 89.485 | 89.517 | 89.509 | 89.533 | 89.496 | 89.524 | |
UNet + DNN | 90.320 | 90.353 | 90.329 | 90.367 | 90.312 | 90.344 | 90.336 | 90.360 | 90.323 | 90.351 | |
MEL | DenseNet + DNN | 87.859 | 87.836 | 87.872 | 87.844 | 87.867 | 87.829 | 87.855 | 87.841 | 87.863 | 87.837 |
VGG19 + DNN | 88.722 | 88.699 | 88.737 | 88.709 | 88.731 | 88.694 | 88.720 | 88.706 | 88.729 | 88.701 | |
Xception + DNN | 89.520 | 89.497 | 89.535 | 89.507 | 89.529 | 89.492 | 89.518 | 89.504 | 89.527 | 89.499 | |
UNet + DNN | 90.347 | 90.324 | 90.362 | 90.334 | 90.356 | 90.319 | 90.345 | 90.331 | 90.354 | 90.326 | |
BKL | DenseNet + DNN | 87.840 | 87.873 | 87.849 | 87.831 | 87.864 | 87.846 | 87.868 | 87.854 | 87.827 | 87.851 |
VGG19 + DNN | 88.703 | 88.736 | 88.712 | 88.694 | 88.727 | 88.709 | 88.732 | 88.718 | 88.691 | 88.715 | |
Xception + DNN | 89.501 | 89.534 | 89.510 | 89.492 | 89.525 | 89.507 | 89.530 | 89.516 | 89.489 | 89.513 | |
UNet + DNN | 90.328 | 90.361 | 90.337 | 90.319 | 90.352 | 90.334 | 90.357 | 90.343 | 90.316 | 90.340 | |
BCC | DenseNet + DNN | 87.866 | 87.843 | 87.825 | 87.858 | 87.880 | 87.852 | 87.834 | 87.869 | 87.847 | 87.829 |
VGG19 + DNN | 88.729 | 88.706 | 88.688 | 88.721 | 88.743 | 88.715 | 88.697 | 88.732 | 88.710 | 88.692 | |
Xception + DNN | 89.527 | 89.504 | 89.486 | 89.519 | 89.541 | 89.513 | 89.495 | 89.530 | 89.508 | 89.490 | |
UNet + DNN | 90.354 | 90.331 | 90.313 | 90.346 | 90.368 | 90.340 | 90.322 | 90.357 | 90.335 | 90.317 | |
AK | DenseNet + DNN | 87.847 | 87.870 | 87.842 | 87.864 | 87.836 | 87.859 | 87.841 | 87.875 | 87.853 | 87.828 |
VGG19 + DNN | 88.710 | 88.733 | 88.705 | 88.727 | 88.699 | 88.722 | 88.704 | 88.738 | 88.716 | 88.691 | |
Xception + DNN | 89.508 | 89.531 | 89.503 | 89.525 | 89.497 | 89.520 | 89.502 | 89.536 | 89.514 | 89.489 | |
UNet + DNN | 90.335 | 90.358 | 90.330 | 90.352 | 90.324 | 90.347 | 90.329 | 90.363 | 90.341 | 90.316 | |
VL | DenseNet + DNN | 87.854 | 87.831 | 87.867 | 87.839 | 87.862 | 87.844 | 87.876 | 87.848 | 87.830 | 87.855 |
VGG19 + DNN | 88.717 | 88.694 | 88.730 | 88.702 | 88.725 | 88.707 | 88.739 | 88.711 | 88.693 | 88.718 | |
Xception + DNN | 89.515 | 89.492 | 89.528 | 89.500 | 89.523 | 89.505 | 89.537 | 89.509 | 89.491 | 89.516 | |
UNet + DNN | 90.342 | 90.319 | 90.355 | 90.327 | 90.350 | 90.332 | 90.364 | 90.336 | 90.318 | 90.343 | |
DF | DenseNet + DNN | 87.861 | 87.838 | 87.874 | 87.846 | 87.868 | 87.850 | 87.832 | 87.857 | 87.879 | 87.841 |
VGG19 + DNN | 88.724 | 88.701 | 88.737 | 88.709 | 88.731 | 88.713 | 88.695 | 88.720 | 88.742 | 88.706 | |
Xception + DNN | 89.522 | 89.499 | 89.535 | 89.507 | 89.529 | 89.511 | 89.493 | 89.518 | 89.540 | 89.504 | |
UNet + DNN | 90.349 | 90.326 | 90.362 | 90.334 | 90.356 | 90.338 | 90.320 | 90.345 | 90.367 | 90.331 | |
