Table 5 Average RR for the proposed algorithm when various features elements are removed from the proposed feature vector
From: Content-based obscene video recognition by combining 3D spatiotemporal and motion-based features
Removed features | SVM core | Average RR (%) |
|---|---|---|
Features based on the information of single frames | RBF | 75.8 |
Linear | 84.5 | |
Quadratic | 80.6 | |
Polynomial | 83.2 | |
Features based on 3D STV (first group) | RBF | 77.3 |
Linear | 92.1 | |
Quadratic | 84.8 | |
Polynomial | 83.5 | |
Features based on 3D STV (second group) | RBF | 77.1 |
Linear | 93.7 | |
Quadratic | 85.3 | |
Polynomial | 80.6 | |
Features based on motion and periodicity characteristics | RBF | 74.5 |
Linear | 94.7 | |
Quadratic | 89 | |
Polynomial | 86.7 |