Table 3.
Comparative experimental results of different texture features and classifiers.
| Methods | AVG-Pre (%) | AVG-Re (%) | AVG-F1 (%) | Accuracy (%) | |
|---|---|---|---|---|---|
| Texture | Classifier | ||||
| LBP | SVM | 93.45 (±2.61) | 93.15 (±3.02) | 93.25 (±2.86) | 93.97 (±2.18) |
| HOG | SVM | 89.87 (±2.26) | 89.22 (±2.59) | 89.45 (±2.51) | 90.72 (±2.36) |
| LBP + HOG | SVM | 94.27 (±3.17) | 93.88 (±3.29) | 94.03 (±3.27) | 94.67 (±2.64) |
| LBP | KNN | 88.96 (±2.90) | 87.08 (±3.57) | 87.66 (±3.47) | 89.33 (±3.30) |
| HOG | KNN | 89.31 (±3.57) | 88.07 (±4.19) | 88.42 (±4.05) | 89.78 (±3.30) |
| LBP + HOG | KNN | 90.32 (±0.88) | 89.77 (±1.07) | 89.95 (±0.75) | 90.87 (±0.67) |
| LBP | NB | 70.29 (±6.41) | 70.65 (±6.07) | 69.91 (±6.32) | 72.68 (±6.14) |
| HOG | NB | 73.73 (±3.55) | 73.17 (±3.05) | 73.25 (±3.33) | 76.33 (±2.90) |
| LBP + HOG | NB | 78.08 (±2.95) | 77.34 (±3.64) | 77.24 (±3.32) | 79.81 (±2.88) |