Table 2.
Classification accuracy (unit: %) in stroke subjects when both TD and WPT features were extracted from the EMG data of 46 high-density channels and 10 predefined channels, respectively.
Subject # | 46 high-density channels |
10 predefined channels |
||
---|---|---|---|---|
TD | WPT | TD | WPT | |
1 | 94.36 | 98.74 | 82.89 | 86.57 |
2 | 91.15 | 95.75 | 80.61 | 82.46 |
3 | 94.07 | 98.56 | 93.47 | 89.34 |
4 | 87.36 | 98.00 | 82.93 | 87.69 |
5 | 96.81 | 94.22 | 96.73 | 98.49 |
6 | 95.02 | 98.61 | 86.56 | 86.65 |
7 | 99.65 | 100.0 | 94.67 | 96.56 |
8 | 99.58 | 100.0 | 96.39 | 99.47 |
9 | 93.63 | 98.96 | 95.94 | 94.90 |
10 | 97.84 | 99.78 | 86.80 | 96.26 |
11 | 99.32 | 99.78 | 98.60 | 97.95 |
12 | 100.0 | 100.0 | 98.20 | 98.60 |
Average | 95.73 ± 3.90 | 98.53 ± 1.82 | 91.15 ± 6.68 | 92.92 ± 5.95 |
TD, the time domain feature set; WPT, the proposed feature set using wavelet packet transform.