Research Article
Feature Extraction of Surface Electromyography Using Wavelet Weighted Permutation Entropy for Hand Movement Recognition
Table 2
Recognition accuracies of d3 and undecomposed sEMG for the seven hand movements (×100%).
| Hand movement | d3 (PE/WPE) | Undecomposed sEMG(PE/WPE) | SVM | BPNN | SVM | BPNN |
| Open | 0.64/0.71 | 0.77/1.00 | 0.87/0.92 | 1.00/0.62 | Close | 0.62/0.87 | 0.69/1.00 | 1.00/0.83 | 0.92/0.62 | Point | 0.71/0.82 | 0.85/0.92 | 0.82/1.00 | 0.85/0.92 | Ok | 0.78/0.79 | 0.38/0.85 | 0.90/0.80 | 0.54/0.77 | Yeah | 0.76/0.92 | 0.92/0.77 | 0.69/0.92 | 0.92/1.00 | Tripod | 0.92/0.89 | 0.92/0.92 | 0.92/0.93 | 0.62/1.00 | Grip | 0.88/1.00 | 0.54/0.69 | 0.80/0.92 | 0.38/0.85 |
|
|