Table 1.
A comparative study of some past methods related to the proposed work.
| Work Ref. | Dataset | Method/Classifier | Accuracy | Sensitivity | Specificity |
|---|---|---|---|---|---|
| [8] | Custom | SVM on rsfMRI | 86.96% | 78.95% | 92.59% |
| [9] | PPMI | SVM | 93% | 93% | 92% |
| [10] | PPMI | Sparse feature selection model | 80% | 84.70 ± 19.29% | - |
| [11] | PPMI | PCA followed by SVM | >90% | >90% | >90% |
| [12] | PPMI | SVM with striatal binding ratio | 96.14% | 95.74% | 77.35% |
| [13] | PPMI | SVM | 92.6% | 91.2% | 93.1% |
| [14] | PPMI | ANN | 94% | 100% | 88% |
| [15] | PPMI | AlexNet | 88.9% | - | - |
| [18] | Custom | CNN | 85% | - | - |
| [19] | PPMI | VGG-16 | 95.2% | - | 90.9% |
| [20] | PPMI | InceptionV3 | 98.4% | 98.8% | 97.6% |
| [21] | PPMI | AlexNet and LeNet | 95±0.3% | - | - |
| [31] | PD dataset | KNN | 98.46% | - | - |
| [32] | PPMI | SVM with linear kernel classifiers | 96% | - | - |
| [33] | Custom | Modified Grey Wolf Optimization | 94.83% | - | - |
| [34] | Custom | Optimized cuttlefish algorithm | 94% | - | - |
| [35] | PPMI | PCA and ANN | 97% | - | - |
| [36] | Custom | ROI based diagnosis | 86.67% | - | - |