Table 4.
The obtained results for the second experiment by using local feature descriptors and other classifiers
| Linear discriminant | Tree | k-NN | Naïve bayes | Average accuracy (%) | |
|---|---|---|---|---|---|
| LBP | 97.4% | 92.1% | 93.4% | 94.4% | 94.32 |
| BGP | 99.4% | 96.5% | 98.7% | 98.0% | 98.15 |
| BSIF | 98.2% | 98.2% | 97.3% | 97.0% | 97.67 |
| CENTRIST | 97.3% | 96.2% | 83.6% | 92.1% | 92.30 |
| PHOG | 98.7% | 97.6% | 99.5% | 93.4% | 97.30 |
| LPQ | 99.1% | 97.7% | 98.4% | 98.7% | 98.47 |
| FDLBP | 96.2% | 95.8% | 81.9% | 96.5% | 92.60 |
| QLRBP | 97.3% | 95.4% | 93.0% | 93.4% | 94.77 |
| Average accuracy (%) | 97.95 | 96.18 | 93.22 | 95.43 |