Table 2 Classification performance on the Fitzpatrick17k-C dataset.
From: MTAKD: multi-teacher agreement knowledge distillation for edge AI skin disease diagnosis
Model | Number of parameters | Accuracy (%) | Precision (%) | Recall (%) | F1-score (%) |
|---|---|---|---|---|---|
RegNetY32GF | 143,451,596 | 45.86 ± 0.62 | 47.52 ± 0.83 | 45.86 ± 0.62 | 45.31 ± 0.66 |
DenseNet201 | 19,364,018 | 42.33 ± 0.45 | 44.19 ± 0.42 | 42.33 ± 0.45 | 41.75 ± 0.52 |
Xception | 21,969,050 | 36.11 ± 0.67 | 37.37 ± 0.83 | 36.11 ± 0.67 | 34.80 ± 0.80 |
InceptionV3 | 22,910,354 | 38.42 ± 0.78 | 39.01 ± 0.73 | 38.42 ± 0.78 | 37.61 ± 0.52 |
NASNetMobile | 4,869,382 | 34.67 ± 0.70 | 36.28 ± 0.61 | 34.67 ± 0.70 | 33.94 ± 0.74 |
EfficientNetV2B0 | 6,633,666 | 35.24 ± 0.61 | 36.80 ± 0.96 | 35.24 ± 0.61 | 33.59 ± 0.86 |
MobileNetV2 | 2,972,338 | 35.07 ± 0.57 | 36.86 ± 0.91 | 35.07 ± 0.57 | 34.15 ± 0.73 |
