Table 12 Classification accuracy of the fusion model (full convolutional basis)
Strategie | Classifer | ||||
|---|---|---|---|---|---|
Fusion strategy 1 | Fusion strategy 2 | Fusion strategy 3 | Fusion strategy 4 | FuNet | |
Sigmoid-SVM | 59.42/77.67 | 65.22/74.33 | 63.77/62.66 | 57.97/68.33 | 66.67/81.67 |
Rbf-SVM | 60.87/94.00 | 63.04/92.67 | 62.32/78.67 | 59.42/91.33 | 64.49/92.67 |
Poly-SVM | 62.32/96.00 | 65.94/95.67 | 55.07/89.00 | 60.15/97.67 | 64.49/96.33 |
Linear-SVM | 41.30/38.33 | 39.13/36.00 | 39.13/36.00 | 39.13/36.00 | 39.13/36.00 |
RF | 60.87/94.00 | 62.32/94.33 | 58.70/94.00 | 60.15/94.67 | 63.77/95.67 |
KNN | 50.00/96.67 | 46.38/91.67 | 42.75/93.33 | 42.75/91.63 | 42.75/93.00 |
Stacking | – | 64.49/93.67 | 57.97/92.67 | 60.15/95.00 | 72.46/94.33 |
Bagging | – | 73.19/72.67 | 43.48/44.67 | 56.52/76.00 | 71.01/80.00 |
Gradient boosting | – | 65.22/88.00 | 52.17/82.00 | 64.49/87.00 | 60.87/91.00 |