Research Article
Fusion of Deep Features from 2D-DOST of fNIRS Signals for Subject-Independent Classification of Motor Execution Tasks
Table 4
The accuracy of LeNet in different scenarios.
| | Scenario | Number of channels | Fusion scheme | | Early | Late | Joint |
| | (RHT, LHT, FT) | 4 | 75.27 (77.39 ± 4.16) | 71.64 (73.03 ± 4.53) | 77.57 (79.49 ± 4.02) | | 8 | 80.26 (81.95 ± 3.41) | 77.82 (80.68 ± 3.75) | 83.85 (86.14 ± 3.24) | | 16 | 84.74 (86.83 ± 2.26) | 83.29 (85.59 ± 2.37) | 87.95 (90.71 ± 2.08) |
| | (RHT, LHT) | 4 | 82.49 (84.72 ± 3.03) | 77.79 (80.08 ± 3.22) | 83.57 (85.36 ± 2.96) | | 8 | 85.09 (88.11 ± 2.51) | 85.14 (86.63 ± 2.78) | 90.67 (92.09 ± 2.29) | | 16 | 92.16 (93.64 ± 2.08) | 89.07 (91.31 ± 2.43) | 93.54 (95.72 ± 1.91) |
| | (RHT, FT) | 4 | 80.18 (82.90 ± 3.19) | 77.26 (79.77 ± 3.35) | 82.86 (85.17 ± 3.11) | | 8 | 85.37 (87.15 ± 2.77) | 83.87 (85.35 ± 2.96) | 89.93 (91.38 ± 2.47) | | 16 | 90.48 (92.68 ± 2.28) | 88.03 (90.10 ± 2.65) | 92.26 (94.88 ± 2.06) |
| | (LHT, FT) | 4 | 79.85 (81.82 ± 3.29) | 76.58 (78.19 ± 3.48) | 82.08 (84.65 ± 3.13) | | 8 | 84.97 (86.32 ± 2.91) | 81.17 (83.64 ± 3.15) | 87.61 (90.81 ± 2.52) | | 16 | 88.13 (90.18 ± 2.59) | 86.58 (88.90 ± 2.85) | 91.05 (93.19 ± 2.23) |
|
|
The accuracy of modified LeNet is given in parentheses.
|
