Table 7 Proposed results of human fall direction dataset deep learning models, (a) 3-RBNet, (b) 5-RBNet, (c) 7-RBNet and (d) 9-RBNet.
Models | Accuracy (%) | Time (seconds) | Sensitivity | Precision | AUC |
|---|---|---|---|---|---|
(a) Proposed results of 3-RBNet on human direction dataset | |||||
 Linear SVM | 87.6 | 2 | 87.5 | 87.6 | 0.97 |
 Quadratic SVM | 88 | 1.5 | 88 | 88.05 | 0.97 |
 Cubic SVM | 86.5 | 5 | 86.5 | 86.5 | 0.96 |
 Fine Gaussian SVM | 80.1 | 6 | 80.12 | 80.5 | 0.93 |
 Medium Gaussian SVM | 85.5 | 3 | 85.5 | 86.5 | 0.96 |
 Coarse Gaussian SVM | 86.2 | 2 | 86.2 | 86.4 | 0.97 |
(b) Proposed results of 5-RBNet on human direction dataset | |||||
 Linear SVM | 91.8 | 2.5 | 91.7 | 92 | 0.98 |
 Quadratic SVM | 91 | 1.7 | 91 | 91.05 | 0.98 |
 Cubic SVM | 92 | 1.2 | 91.7 | 91.75 | 0.98 |
 Fine Gaussian SVM | 81.1 | 4 | 81.1 | 84.6 | 0.94 |
 Medium Gaussian SVM | 90.9 | 7 | 90.8 | 91 | 0.98 |
 Coarse Gaussian SVM | 87.5 | 1.3 | 87.5 | 88.4 | 0.97 |
(c) Proposed results of 7-RBNet on human direction dataset | |||||
 Linear SVM | 91 | 4.1 | 91 | 91.02 | 0.98 |
 Quadratic SVM | 90.9 | 3 | 91 | 91.2 | 0.98 |
 Cubic SVM | 92.2 | 2.5 | 92.1 | 92.1 | 0.98 |
 Fine Gaussian SVM | 80.8 | 6 | 80.7 | 80.7 | 0.94 |
 Medium Gaussian SVM | 90.8 | 5.3 | 90.7 | 90.7 | 0.98 |
 Coarse Gaussian SVM | 88.9 | 4 | 88.8 | 88.9 | 0.98 |
(d) Proposed results of 9-RBNet on human direction dataset | |||||
 Linear SVM | 91.9 | 3.8 | 91.8 | 91.9 | 0.98 |
 Quadratic SVM | 92.6 | 3 | 92.7 | 92.6 | 0.98 |
 Cubic SVM | 91.5 | 2.6 | 92.5 | 92.5 | 0.98 |
 Fine Gaussian SVM | 71.4 | 6.3 | 71.3 | 71.7 | 0.91 |
 Medium Gaussian SVM | 91.8 | 4.6 | 91.7 | 91.7 | 0.98 |
 Coarse Gaussian SVM | 91.8 | 6.3 | 91.7 | 91.7 | 0.98 |
