Research Article
A Computer-Aided Diagnosis System Using Deep Learning for Multiclass Skin Lesion Classification
Table 3
Classification results using SbE approach-based deep features fusion on augmented HAM10000 dataset.
| | Classifier | Recall rate (%) | Precision rate (%) | FNR (%) | AUC | Accuracy (%) | Time (sec) | F1-score (%) |
| | LSVM | 92.71 | 92.85 | 7.285 | 0.992 | 92.50 | 1303.8 | 92.78 | | QSVM | 94.85 | 94.85 | 5.142 | 0.997 | 94.80 | 2400.4 | 94.75 | | CSVM | 95.00 | 95.00 | 5.00 | 0.854 | 94.90 | 2868.5 | 95.00 | | MGSVM | 92.85 | 93.14 | 7.142 | 0.995 | 92.60 | 4501.6 | 92.99 | | CKNN | 61.71 | 73.14 | 38.28 | 0.910 | 62.20 | 562.0 | 66.94 | | CKNN | 84.14 | 83.57 | 15.85 | 0.975 | 82.60 | 550.5 | 83.85 | | WKNN | 83.42 | 84.57 | 16.27 | 0.971 | 82.10 | 530.7 | 83.99 | | ESD | 95.00 | 95.00 | 5.00 | 0.997 | 95.00 | 4118.2 | 95.00 | | FKNN | 88.14 | 87.57 | 11.85 | 0.931 | 87.00 | 543.2 | 87.85 | | EBT | 62.00 | 62.22 | 38.00 | 0.855 | 62.80 | 69886.0 | 62.11 |
|
|
The bold value represents best ones.
|
